Factorial Design – Introduction
- Introduction to the factorial design
o A research study involving two or more factors. This is a beneficial research strategy as it increases external validity by making the experimental conditions more akin to real life.
- Main effects and interactions
o In the factorial design, each factor produces a main effect. The factorial design is used to analyze if there is an interaction between factors
o This occurs when two factors, acting together, produce mean differences that are not explained by the main effects of the two factors
- Identifying interactions
o Main effect in one factor bot no main effect in the other with no interaction o Main effects in all factors’ bot no interaction
o No main effects for any factor with an interaction
- Types of factorial design
o Between-subject design
o Within-subject design
o Mixed design
- How to display data in the factorial design
o Tables and graphs
o The factorial matrix
How to analyze the matrix to determine the interaction
- Hypothesis tests and analyzing the results
o The P-value and what it means
o Determining if the interaction is due to chance
- Factor: in an experiment, an independent variable is often called a factor, especially in an experiment that includes two or more independent variables • Independent variables of an experiment
• Handling a mouse
• Administering a drug
• Values of the independent variable
• Handling for 0, 10, 20min, etc
• Drug dose at 10, 20, 30mg etc If you want to learn more check out What Does Government Do?
Don't forget about the age old question of Storytelling performances consisted of both what?
- Condition (treatment)
• How is the group treated in the experiment
• What is the subject experiencing by each of those groups
• A combination of the factor and levels
• Handled for 10min and given 10mg drug
• Not handled and given 30mg drug
- Factorial design: a research design that includes two or more factors - Main effect: the mean differences among the levels of one factor are the main effect of that factor
o A two factor study has two main effects; one for each of the two factors
- Interaction between factors: when the effects of one factor depend on the different levels of a second factor, then there is an interaction between factors o Occurs whenever two factors, acting together, produce mean differences that are not explained by the main effects of the two factors.
o If the main effect for either factor applies equally across all levels of the second factor, then the two factors are independent and there is no interaction.
- Mixed design: a factorial study that combines two different research designs o A common example is a factorial study with one between-subjects factor and one within-subjects factor If you want to learn more check out What is the code of event “Pre-Admit a Patient”?
- Combined strategy study uses two different research strategies in the same factorial design. One factor is a true independent variable (experimental strategy) and one factor is a quasi-independent variable (nonexperimental pr quasi experimental strategy)
- Analyzing factorial studies
o Main effects
- Types of factorial design
o Pure factorial design
o Between-subjects groups Don't forget about the age old question of What is Newton's Third Law of Motion (the law of action and reaction)?
o Within-subjects measures
- Mixed factorial design
- Advantages and disadvantages
• Factorial experiments are experiments where multiple independent variables are manipulated
• In the real world, effects have multiple causes and thus the factorial de • Advantage: The factorial design creates a more realistic situation that can be obtained by examining a single factor in isolation.
• Because behaviour is influences by a variety of factors usually acting together, the factorial design enhances external validity because it can examine two or more factors simultaneously.
• Manipulating multiple independent variables to determine whether or not they have an impact on your outcome. We also discuss several other topics like Theory that states Indigenous people may have came to America through a giant plane of land connecting Asia and North America.
• Having multiple independent variables allows them to interact with one another to make the effective one variable stronger or weaker.
• Examining multiple factors increases external validity
• Because you’re taking all possible scenarios into account.
• By being able to assess the multiple determinants of behaviour, you are able to better approximates what the behaviour would be like in the real world
Factorial design: terminology
- Uses one numeral for each factor/independent variable,
- The value of the numeral indicates the number of levels for that independent variable (factor)
- Each level of independent variable is paired with the level of every other independent variable
o 2 X 2 factorial design
Two independent variables, A and B
A and B (independent variables) Both have Two levels
for each independent variable
o Four conditions
4 X 2 factorial design We also discuss several other topics like What drives global atmospheric circulation?
Two independent variables
Four levels for independent variable A, two levels for
INDEPENDENT variable b
- Eight conditions
o 5 X 3 factorial design
Independent variables: two
independent variables (two numerals)
# of levels:
1st = 5 levels
2nd = 3 levels
Fifteen total conditions
o 2X3X2 factorial design
Three independent variables
1st = 2 levels
2nd = 3 levels
3rd = 2 levels
12 total conditions
Simplest 2x2 factorial design
- To identify an interaction in a factorial study, you must compare the mean differences between cells with the mean differences predicated from the main effects.
- If there is no interaction,
o The combination of the two main effects completely explains the mean differences between cells.
- The main effects from a two-factor design reveal how each of the factors affects behaviour.
- And the interaction reveals how the two factors operating together can affect behaviour.
o The existence of the interaction indicates that the combined effects of the two factors acting together is not the same as the effect of one factor acting alone.
- When the data from a two-factor study are organized in a matrix, o the mean differences between the rows describe the main effect for one factor
o and the mean differences between the columns describe the effect for the second factor.
- The extra mean differences that exist between cells in the matrix o Are differences that are not explained by the overall main effects o Describe the interaction and represent the unique information that is obtained by combining the two factors in a single study.
- Example: Effect of diet and exercise on body weight
o Independent variable: Diet (chow vs ketogenic diet)
o Independent variable: Exercise (housed with wheel vs no wheel) o Dependent variable: Body weight
How to plot factorial matrix
- No effect 2 Main
- Y axis= dependant variable x axis= independent variable o When there are two independent variables you choose one factor and you put that on the x=axis and the other independent variable becomes part of the legend
o Whatever is going across factorial matric, put that across the graph o Factor B becomes part of your legend
You would put a filled circle as no wheel and a empty circle as the with wheel condition
- In no effect graph
o All the groups weigh the same
o So you plot both at the same body weight and connect the dots. o You should have two data points with filled dot and two data points with non filled circle
There is no effect because all the mice weighed the same
- In the 2 main effect graph
o There is a change in body weight of the mice so you see that both the mice in groups with and without wheel lost weight when they did the ketogenic diet
How to approach a factorial matrix
- Identify the main effects, if any
o Calculate the average of each
o If the row/column averages are different
main effect of that factorIf column
averages are different main effect of
- Determine if there is an interaction
o On the graph, are the lines parallel or not
If not parallel, there is an
If parallel, there is no interaction
o Also a mathematical way to determine if
there is an interaction
- Discuss the main effects
o If there is an interaction, then you cant
describe the effect of A without taking
the effect of B into consideration
- The main effects reflect the results that would be obtained is each factor were examined in its own separate experiment.
• The effect of one independent variable
• Is there an overall effect of manipulating the independent variable? • Mean differences among the levels of one factor
• Calculate the average of row or column
• The main effect for one factor is obtained by averaging all the different levels of the second factor. Because each main effect is an average, it may not accurately represent any of the individual effects that were used to compute the average.
Mathematical way to observe effect of diet
1. Take the average of the rows and columns
2. If mean average values are not the same on row = effect
3. If mean average of the columns are not the same = effect
interpreting main effects and interactions
- When the effects of one factor depend on the different levels of a second factor then there is an interaction between factors.
o When the size of factor A depends on the effect of factor B such that as factor B increases/ decreases there is a change in factor A then there is an interaction.
o when the results are graphed, the existence of nonparallel lines (lines that converge or cross) is an indication of an interaction between the two factors - Is two factors are independent so that the effect of one is not influenced by the other then there is no interaction.
o When the size of the effect of one factor does not depend on the effect of the other = factor A does not depend on factor B and there is no interaction o When the data is plotted on a graph there is a noticeable pattern such that as the line on the graph for factor A increase/ decrease factor B changes proportionally with it. The effect is two parallel lines on a graph. And there is no interaction
- Determines how a combination of factors work together to affect behavior o Occurs when one factor has a direct influence on the effect of the other If Factor A and Factor B are independent, they have no interaction Factor A can exaggerate/minimize the effects of Factor B
- Compare the mean differences in any individual row with the mean differences in other rows
o No interaction: Size and direction of the differences in each row is the same as other rows
o Evidence of interaction: Differences change from one row to another - Nonparallel lines indicate an interaction between the two factors o Not parallel = interaction
o Parallel = no interaction
- Note: still need statistical test to determine if difference is significant - The mean differences are simply descriptive and you still need to evaluate by a statistical hypothesis test before they can be considered significant.
Lines are not parallel
- The interaction shows the effect of
o Body weight decreased when
fed a ketogenic diet. Body
weight further decreased with
- Because there is an interaction,
cannot describe the effect of
exercise without taking into account
- You only need one of the interactions
to be different to have an
- Lines are not parallel = interaction
- The main effect does not accuratelt
describe the results. It would be incorrect to conclude that there is no relationship between the two factors.
1 main effect, no interaction
- You compare the values row to row and column to
- The lines are parallel so there is no interaction
Discuss the main effects: interpreting results
- The effects of exercise (Factor A) on body weight (DV)
depend on the diet consumed (Factor B)
o Main effect of diet and exercise
o Regardless of exercise, body weight decreases
when eating a ketogenic diet (compared to
chow). However, when allowed to exercise,
body weight further decreased when eating a
Two main effects, no interactions
More comples: 2x3 factorial design
- What are the main effects
o To determine the number of main effects possible, look at the number of factors
Every factor can have an effect on your variable. So there is one main effect per factor
2x2 = 2 main effects
2x3 = 2 independent variables + 3 levels = 2 main effects
- How many interactions
o 2-way interaction = ninteraction between 2 factors
o 3-way interaction = interaction between 3 factors
- The number of interactions depends on the number of factors
o 2x2 design
One 2-way interaction
o 2x2x2 design (variables A, B, and C)
Three 2-way interactions
AB, AC, BC interact together
And one three way interaction
ABC interact together
o 3x3x3x2 (variables A, B, C, and D
Four possible main effects
AB, AC, AD, BC, BD, CD
ABC, ABD, BCD, ACD
Independence of Main effects and interactions
- The two factor study allows researchers to evaluate three separate sets of mean differences
o The mean differences from the main effect and factor A
o The mean differences from the main effect and factor B
o The mean differences from the interaction between the factors. - The three sets of mean differences are separate and completely independent. Thus it is possible for the results from a two-factor study to show any possible combination of main effects and interactions.
Mean differences between levels of factor A but no mean differences for factor B
- To identify the main effect for factor A, notice the overall mean for the top row is 10 points higher than the overall mean for the bottom row. This difference is the main effect for factor A.
- To evaluate the mean effect for factor B, notice that both columns have exactly the same overall mean, indicating no difference between levels of factor B, hence no B effect.
Data with an A effect and a B effect but no interaction - The A effect is indicated by the 10-point mean difference between rows - The B effect is indicated by the 20 point mean difference between columns. - The fact that the 10-point A effect is constant within each column indicates no interaction
o Because it doesn’t change as the B factor changes
Data showing no main effect for either factor, but an interaction - There is no mean difference between rows = no A effect
- And no mean difference between columns = no B effect
- However, within each row and column there are mean differences o The extra mean differences within the rows and columns cannot be explained by the overall main effects (individual effects of factor A and factor B) and therefore indicate an interaction
Types of Factorial Designs
- Disadvantages of a pure factorial design
o You need more participants/ subjects because separate subjects are needed in each of the two groups (control and experimental)
o Example: in a 2x4 design, each participant must be measured in eight different treatment conditions. The large number of different treatments can be time consuming, and increases participant attrition
o Having each participant undergo a long series of treatment conditions can increase the potential for testing effects (fatigue or practice effects) and make it more difficult to counterbalance the design to control for order effects
Types of factorial designs
• Independent groups
• Factorial experiments between different groups of people
• Repeated measures
• Factorial experiments using different manipulations on the same individuals
• A single group of individuals participates in all of the separate
• Mixed design
• Combination of between- and within-subject design
• One factor is between-subjects and another factor is within-subjects
Pure factorial within-subject design
- Between-groups design
o participants are randomly assigned to each cell of the design
o best suited for situations in which individual differences are relatively large and there is little reason to expect order effects to be large and disruptive. - Disadvantages
o Individual differences can become confounding
variables, increases variance of scores
o Avoid order and sequence effects
Within-group Pure factorial design
- Within-groups design – same individuals participate in ALL conditions - Cons
o Number of different treatment conditions can be high and time-consuming (participant attrition, fatigue/practice)
o Each participant/subject have to go through both coniditons (experimental and control)
o Fewer individuals needed
Reduces problems related to individual differences (choose this if you know the individual differences will be very large)
- Example: Wilson et al., 1994: Effect of marijuana on cognitive performance o 1 group of n = 10 10 subjects total
o Main effect of task difficulty and marijuana
o There is an interaction between marijuana and task perfromanse As reaction time increases when performing a hard task, this
performance can be greater when smoking something with THC
- Order becomes something you have to consider
Mixed factorial design (no longer pure factorial design) - Combines both between- and within-subjects design
- Effect of manipulations on two groups of individuals
- Often used in before-after situations between two groups
o Control for unwanted individual differences, while investigating specific individual differences
4 groups down to two groups
Limit time effects of within subject design
o Limits ability to make causal statements about the relationship between variables; limit internal validity
- Effect of caffeine to offset debilitating effects of alcohol
o Randomize in placebo or caffeine group
o Each individual gets water and alcohol, compare effect of placebo/caffeine in water/alcohol
Advantages of factorial design
- Increases external validity (Generalizations)
o Variation of single independent variable is unusual in real-life situations o Allows you to better generalize to a real effect
Because situaitons in real life are very rarely due to only one cause So by taking multiple variables into account you are taking into
consideration that multiple things can effect one behaviour
- Theories with two or more independent variables can only be tested via complex factorial designs
o Exercise, diet
- Greater experimental control
o Testing multiple variables in one experiment reduces random variation that arise when comparing different variables in different experiments
Example: Some mice are better wheel runners than others, if you had totally different groups of mice, will see more within-group variability
o Having less groups with larger number of participants you minimize the variablility between groups and makes the statistics of your analysis a bit more robust
Disadvantages in factorial design
- Too many variables can result in:
o Huge experiments and requirement for multiple conditions
And potentially a lot of participants/subjects
o Interactions that are not easily interpreted because combinations may not make sense.
Some conditions tested may not arise naturally
Interaction between day length and temperature on work
o Long/cold days
o Long/hot days
o Short/cold days
o Short/hot days
You don’t have short long days because hot days are
in the summer where the days are longer.
Towards the end of the research process.
- Descriptive statistics: are methods that help researchers organize,
summarize and simplify the results obtained from research studies. - Observations are inherently variable
o We want to have a set of criteria to ensure our measurements aren’t due to
o When observations are variable it is partly due to measurement error o Statistics allow accurate conclusions to be drawn from data
o Despite this variability statistics draw meaning from the data
- Allow quantification of observations
o Quantifying the observations = increase by 1% or decrease by 1% - Allow observations to be summarized
o Statistics allow you to discuss data/observations as a group
Types of Statistics
- Descriptive statistics used to describe data
o Organize data
o Summarize data
- Inferential statistics allows you to make inferences from data
o Use results from a sample to make conclusions about a population o Determine if there is statistical significance within data set
Examples of descriptive statistics
- Organize data
o Frequency distribution
- Summarize data
o Measures of central tendency
Mean, median, mode
- Measures of dispersion
o Standard deviation, variance, range
- Measures of relationship
o Pearson, Spearman
- One method of simplifying and organizing a set of scores by grouping scores into an organized display that shows the entire set of scores.
o Consists of a tabulation of the number of individuals in
each category on the scale of measurement.
- Displays two sets of information
o The set of categories that make up the scale of
o The number of individuals with scores in each category. - Can often look like a bar graph but it allows you to
quantify all your scores by grouping them into a display o Eligible for all scales of measurement
o Preliminary method of statistical analysis, rarely shown in
- Frequency distribution table
o Example: the number of times students in NEUR2001 visit CULearn before a
o Days= number of days before a midterm (x-axis/ first column)
o Frequency = number of visits (y-axis)
Scale of measurement
Or simply lists the set of categories into which individuals have been assigned
Frequency or the number of individuals in each category
- The same information presented on a frequency distribution table can be
presented on a graph.
o Frequency distribution graph
- Frequency distributions, especially graphs, can be an effective method for
presenting information about a set of scores.
- The distribution shows whether the scores are clustered together or spread out across the scale.
o Gives you a concrete representation of all the individual scores as well as the
appearance of the entire set of data
o You can see whether the scores are clusters or spread out.
o You can see is the scores are generally high or low (where the distribution is
o And if there are any extreme scores that are very different from the rest of the
- X-axis Scale of measurement
- Y-axis Frequency
- Frequency distribution plot can be misconstrued as a bar graph - Primary differences
o In histogram the bars are touching in a bar graph the bars are not touching o There is not relationship between 100 and 200 in the bargraph o In the histogram there is a relationship you are trying to see between
number of days and visits
o Shows a bar above each score so that the height of the bar indicates the frequency of occurance for that particular score
- Bar graph
o When categories on the cale of measurement are not numerical values (and are thus qualitative values) the frequency distribution.is presented as a bar graph.
o A bar graph is like a histogram except that a space is left between adjacent
o The height of each bar indicates the frequency associates with that particular category.
- How we identify the score with the greatest frequency
o Which one is the most common
o Applies to all scales of measurement
- Not commonly used
o If you wanted to get the most common score you would use this - Examples
o 5, 7, 7, 8, 8, 8, 10: Mode = 8
o Because 8 happened the most times (3 times)
- When the sample contains a few extreme (outliers) scores the mean tends to be distorted by the extreme values. This causes a large distribution.
- What is the score that divides the distribution in half
o (number right in the middle of your entire data
o Midway point of your distribution
- Cannot be used for nominal scale
- Can be useful if there are outliers
- Score that divides the distribution in half
o Data set: 5 6 4 9 1 1 3
Arrange them in order: 1 1 3 4 5 6 9
Middle number = median: 4
- Cannot be used for nominal scale
- Can be useful if there are outliers
- Score that divides the distribution in half
o If there were two numbers in the middle, take the average
o 5 6 4 9 1 1 3 2 1 1 2 3 4 5 6 9 3.5
- Add all scores and divide the sum by the number of individuals o Amount each individual receives if the scores were divided equally - Cannot be applied to nominal or ordinal scale
o In order to add values together, the values must be
o Easily understood and computed
o Useful for inferential statistics
o You’re not dealing with each individual number
o You can determine what the trend is/ what you expect to
o Does not produce a representative value when there are outliers o You lose the nuance in how the data set is changing over time
Recap: Scales of
- Nominal – used to name or
label a series of values
- Ordinal – give information on
the order of choices
o Put in rank order
- Interval – gives order of
values and difference between
- Ratio – gives order and
Measures of dispersion
- This measures the variability
- How much the scores in a distribution differ from the measure of central tendency - The “range” of scores in the dataset
- Difference between the largest and smallest
o From lowest to highest score and
everything in between but you don’t
know the individual scores in between
from just looking at the range. You
would only know that all scores fall
between the range
- Only based on two scores
- But not informative because it ignores most of
Measure of variance
- Variability describes the spread of scores in a distribution. When variability is
small, it means that scores are all clustered together
o Large variability means that there are big differences between individuals
and the scores are spread across a wide range of values.
- Better measurements of dispersion would include information from every score
- The means are identical for these two plots but the spread of data is different - If data is summarized only via the mean score, one would lose a lot of information about the differences between them
- Standard deviation uses the mean of the distribution as a reference point and
measures variability by measuring the distance between each score and the mean. o When the scores are clusters close to the mean The standard deviation
o When the scores are scattered widely arounf the mean the standard
deviation is large
- Standard deviation (SD) Indicates how much everyone within a sample deviates
from the sample mean
- Measures the average distance from the
o Value is derived from the variance
- Variance = sum of the squares of the
deviations from the mean, divided by the
number of scores minus 1
- Variance measures the variability by
computing the average squared distance from
- Standard deviation = square root of variance
In summary, the calculation of varience and standard deviation is a series of steps
1. For each score, measure the distance away from the mean. This distance is often called a deviation.
2. Square each of the distances and compute the average of the squared distances. This is variance.
3. Note: the average squared distance for a sample is computed by dividing the sum of the squared distances by n – 1, where n is the number of scores in the sample.
o The value of n – 1 is called degrees of freedom, or df.
o Finding the average by dividing by n – 1 (instead of n) produces a variance for the sample that is an accurate and unbiased representation of the population variance.
4. Because the variance measures the average squared distance from the mean, simply take the square root to obtain the standard deviation.
5. Thus, variance and standard deviation are directly related by a squaring or square root operation
Standard error of the mean (SEM)
- Standard deviation spread of values within a sample
- Provides a measure of how much difference is reasonable to expect between a
sample and its population
- Measures how accurately the sample reflects the population
o Sample mean varies from sample to sample
o Estimate how much sample means vary from the distribution
o “How certain am I that the sample mean accurately estimates the population
You need to find the standard error of the mean
o Standard deviation = variance within the sample
o Standard error= how well you are reflecting the population
- The expected value for the standard deviation of several sample means, and is
estimated from a single sample
- Depends on standard deviation and sample size
- Sample size increases, standard error decreases
- By contrast, standard deviation stays the approximately the same even when
sample size increases
- When values are deviating far from the mean then the standard deviation is
going to be large
o because standard variation describes the variability of individual values it is
a measurement of variability.
o It is not describing how accurate your sample mean is it is just measuring how variable your data set is from the mean within your sample.
o Standard variation increases as variance increases
- Standard error
o Can be altered as sample size changes because of the calculation o It depends on standard deviation and sample size.
o As your sample size increases, standard error will decrease
As youre getting more information contributing to results you are
getting more and more accurate of a representation
Because by including more information you are getting closer to
accurately representing your whole population.
To be able to get a sample as close as possible to population you have to sample more people/ animals
- Typically use standard deviation when trying to assess your sample or identify
o Outliers – falls outside of the normal distribution
o Values that fall two standard deviations away from mean it is considered to
be an outlier
o Which is usually excluded from the data set.
- Then use standard error to report final mean
o Because you’re trying to estimate if your data set can represent the
- Make sure to declare what is used in your results
o Declare whether or not you’re using the calculation for standard deviation o Use it if you’re trying to only look at your sample and not trying to reflect the
experimental results to the population
o If you are trying to do this then use standard error
- Are methods that use the results obtained (limited information) from
samples to help make generalizations about populations
- Allow researchers to infer/generalize observations to the larger population - Goal of all research, regardless of the research strategy, is to find a significant
- The goal of inferential statistics is to use limited information from samples as
the basis for reaching general conclusions.
- Want your results to be so extreme that it is unlikely the results are due
to chance alone
o Allow researchers to say how likely it is that the observed difference between groups is due to random chance
- The goal if inferential statistics involves making a generalization or an inference from limited information to a general conclusion.
- This leads to sampling error
o A sample does not provide a perfectly accurate picture of its population. There is some discrepancy, or error between the information available and the true situation that exists in the general population
Example: Is my coin biased
- Is it more likely to land on heads or tails and are the results just due to chance o Coin 1 80 heads, 20 tails more convincing results than 8:2
o Coin 2 53 heads, 47 tails
o Coin 3 8 heads, 2 tails
o Coin 4 60 heads, 40 tails
- Where do you draw the line in deciding if the coin is biased or not? - Coin 1 vs coin 3
o 80 vs. 20 is more convincing than 8 vs 2 even though the proportion of
scores that deviatr from expectations is the same
o So what is the threshold?
- Coin 2 vs coin 3
o 53:47 vs 8:2
o In both cases the difference is only 6
o The number of scores that deviate is the same but you might say that coin 3 is more biased because the numbers are lower
o But actually the difference between the two is fairly similar
- Where do you draw the line in deciding if the coin is biased or not? - The arbitrary line between sufficient and not sufficient is the threshold of
statistical significance P value
o 95 heads, 5 tails: the coin is probably biased
o 55 heads, 45 tails: the coin is probably not biased
o If the probability is small its probably not due to chance
- Inferential statistics – puts a value on that probability that the result is not
o the probability that the result is due to chance
- Get a probability that the coin is not biased this probability is the P value
- Range between 0 (impossible that it is due to chance) and 1 (certain it is due
o 70 heads and 30 tails, p = 0.001
There is very low probability (0.1%) that the coin is NOT biased Therefore, the coin is biased
- Refer to the probability that there is NO differences between the groups - By convention, this boundary is commonly p = 0.05
- If it is definitely due to chance then the effect that you’re seeing is not real
P value in Inferential statistics
- Allow researchers to infer/generalize observations to the larger population
(whether or not something was due to chance)
o Allow researchers to say how likely it is that the observed difference between
groups is due to random chance
- “Probability that the difference is due to chance is 5%”
- 5% chance that the null hypothesis is correct p value
o If you only have 5% chance that the effect is due to chance then the effect is
not due to chance.
o 5% chance that the null hypothesis to be correct
- Hence unlikely that the difference is due to chance, thus conclude the treatment
effect is real
- Low p value allows you to make inferences (generalizations) from the sample to
o The p value allows you to reject the null hypothesis
- Inferential statistics involves hypothesis testing
- A systematic/ statistical procedure that uses sample data to evaluate the credibility
of a hypothesis about a population
- Attempts to distinguish between two explanations for the sample data 1. That the patterns in the data represent systematic relationships among variables in the population
2. That the patterns int the data were produced by random variation from chance or sampling error
- determines whether the sample data provide convincing evidence to
support the original research hypothesis
o Procedure that allows you to assess whether your sample data is giving you
convincing evidence to support your original hypothesis
- State a null hypothesis – statement about the population parameter being
examined and always says there is no effect, no change, or no relationship. o Specifies what the population parameters should be if nothing
o For example: Exercise as a new therapy for treating depression
Begins by assuming that you’re wrong and then you conduct a study to
reject your null hypothesis.
Null hypothesis: Exercise does not reduce depression symptoms
- Conduct a study to see if this assumption can be rejected o If shoot down the assumption, then you have support for the alternative… o The alternative that the new therapy does reduce depression symptoms
P-test (a test statistic)
- A P-value-test is a summary value that measures the degree to which the sample
data are in accordance with the null hypothesis.
o A large value for the test = a large discrepancy between the sample statistic and the parameter specified by the null hypothesis and leads to rejecting the null hypothesis.
- P = 1.00 means the result is 100% due to chance ∴ no effect o Null hypothesis states that the results of the research study represent
nothing more than chance.
If this is true then the actual results and the chance results should be very similar and the test-statistic (P Value) will have a value near 1.00 or 100%
A P value near 100% means there is no effect, that any patterns in the
sample is just due to chance
- The goal of a hypothesis test is to rule out chance as a plausible explanation for the results. To do this researchers must determine which results are reasonable to
expect just by chance and which are unlikely to happen by chance alone. You do this by determining the alpha level (P-value less than 0.05)
o the level of significance for a hypothesis test is the maximum probability that the research result was obtained simply by chance. A P value of under 0.05 means that the test demands that there is a less than 5% probability that the results are caused only by chance
- Goal of the research study is to show that a treatment has a real effect and that this effect is not due to chance and therefore you want a small P value
o P = 0.05 means 5% chance the result is due to chance
- Examples of null hypothesis:
o When comparing two treatments
Null hypothesis = There is no difference between the treatments P=1.00 = 100% chance there is no difference
Therefore, there is no difference
P=0.0001 = 0.01% chance there is no difference
Therefore there is a difference
o If you’re P value is tiny then there is a difference between the treatments
(control versus experimental group)
- Examples of a null hypothesis
o Examining a correlation
There is no relationship between the two variables and the correlation for the relationship is zero
- Range between 0 (impossible) and 1 (certain)
o 70 heads and 30 tails, p = 0.001
There is very low probability (0.1%) that the coin is NOT biased
Therefore the coin is biased
- Refer to the probability that there is NO differences between the groups - By convention, this boundary is commonly p = 0.05
o If p < 0.05, reject the null hypothesis (reject that there is no difference) o If p > 0.05, fail to reject null hypothesis (cannot reject that there is no
If you cant reject that theres a null hypothesis then you cant reject that there is no difference
- There are many other tests out there that can be used
What is the logic behind test statistics
Null hypothesis is correct…
- Null hypothesis: There is no difference between treatments (blue is right behind
green so you cant see it and there is no difference)
- P = 1.00 ∴ 100% chance that there is no difference
o Therefore there is no difference
Reject null hypothesis
- Reject null hypothesis if there is a very low chance that the means are the same - P < 0.05 ∴ less than 5% chance that there is no difference
- there is a difference
Two ways to increase the difference between groups - Goal: We want to see if there is a difference between population mean vs sample
- You want the curves to be far apart in two ways
1. Increasing the difference between the means
a. By doing this you can shift the curves
2. Decrease the standard deviation between the means
a. The means are still the same distance apart but the standard deviations are smaller
b. When they are smaller we can segregate the two different groups of data that we have (clearly see the difference between group a and group b)
Take home message
- Increase the difference between the means
- Decrease the variance/standard deviation in each sample
- Decrease standard error of the mean to generalize sample data to population (by increasing sample size)
Literature search – topic 10
Primary and secondary sources
- Primary source
o Firsthand report of observations or research results written by the individuals who actually conducted the research and made the observations
o Used to obtain complete and accurate information.
Empirical journal articles
Theses and dissertations
Conference presentations of research results
- Secondary source
o A description or summary of another persons work
o Written by someone who did not participate in the research or observations being discussed.
o Provide concise summaries of past research, are always incomplete.The author has selected only bits and pieces of the original study – the selected parts may be taken out of context and re-shaped to fit a theme different from the original authors.
o Used to gain an overview and identify specific primary sources for more detailed reading. Are only a good starting point for a literature search
Books and textbooks in which the author describes and summarizes past research
Review articles or meta-analyses
Introductory section of research reports in which previous research is
presented as a foundation for the current study
Newspaper and magazine articles that report on previous research
- The primary distinction between the two
o First hand versus second hand research results
- A journal article may not be a primary source. Instead, the article may be a review of other work (as in a review article or meta-analysis), a theoretical article that attempts to explain or establish relationships between several previous studies, or a historical summary of the research in a specific area. None of these is a primary source because none is a firsthand
report of research results.
- A book or book chapter can be a primary source. Occasionally, an individual or a group of researchers will publish an edited volume that presents a series of interrelated research studies. Each chapter is written by the individual(s) who actually conducted the research and is, there- fore, a primary source.
- A journal article may be a firsthand report of research results, yet sections of the article actually may be secondary sources. Specifically, most research reports begin with an introductory section that reviews current research in the area and forms the foundation of the study being reported. This review of current research is secondary because the authors describe research conducted by others. Remember, to qualify as a primary source, the authors must describe their own research studies and results.
Conducting a literature search
- Start the literature search with only a general idea for your research topic o Your purpose is to narrow down your general idea to a specific research question. o And then to find all the published information necessary to document and support that question
- Start with a recently published secondary source (like a textbook)
- Use the chapter headings and subheadings to help you focus your search on a more narrowly defines area.
- Make notes about
o Subject words: a list of correct subject words used to identify and describe the variables in the study and the characteristics of the participants.
Researchers often develop a specific set of terms to describe a topic area. It is easier to locate related research articles if you use correct terms
o Author names: a small group of individual researchers are commonly responsible for the majority of the work being done in a specific area.
Write down the names of researchers you repeatedly encounter as they are typically the leading researchers in the field.
Purpose of a literature search
- The goal in conducting a literature search is to find a set of published research reports that define the current state of knowledge in an area and to identify an unanswered question. o A gap in the knowledge base that your study will attempt to fill
- The purpose of the literature review is to provide the elements needed for an introduction to your own research study.
- You need to find a set of research articles that can be organized into a logical argument supporting and justifying your research purpose.
Using online databases
- The typical database contains millions of publications all cross-referenced by subject words and author names.
- So entering a single word or author name into a data base provides you with all of the records and list of publications that are related to your subject or leading researcher. - Some data bases are full text
o Meaning that each record is complete, word-for-word copies of the original publication
o Contain fewer publications than other databases
o The value of a full-text database is that whenever you find a research article you would like to read, the entire article is immediately available
- Other databases provide only brief summaries of the publication.
o The summary includes a title, the name of the journal, a list of subject words that describe the publication, and an abstract.
o Holds more items than a full-text data base’
Using the advanced search option
- Limit the publication type to all journals or peer-reviewer journals
o This eliminates books and book chapters
o Eliminates dissertation abstracts
- Limit the methodology to empirical study
o This focuses your search on research reports and eliminate essays, discussions, and general review articles
- If your research is focused on a specific age group or population group refine these to limit to these areas
Beginning a literature search
- Reviewing titles of journal articles that come up on a data base is the first level of screening journal articles
o Most articles can be discarded at this stage
- When a title is relevant to your topic
o Obtain the detailed record, which contains additional information about the publication including;
- The second level of screening is reading the abstract
o By reading the abstract, you get an idea about whether or not the article is related to your topic
- Then check the list of key terms to determine whether there are specific terms that could be used to improve your search.
Process of conducting a literature search
Decide if a title is relevant to your topic
Read to determine if the article is interesting
Go to the
Go to the journal that the article was published in. once you find it, skim it and look specifically at the introduction and discussion.
If after skimming the article, it is still relevant then read in depth
discusses previous research that forms the foundation for the current research. This helps determine if the article eill be helpful in the development of your research idea
Information on participants and subjects – description of the sample that participated in the study.
Procedures – description of how the study was conducted, including a description of the questionnaires and equipment used
Findings, statistical analysis,
gives you ideas for future studies. You can use one of the ideas mentioned
Lists all the publications that were cited in the text. These publications can lead to a new subject term or author names for your search
Use the references from the aeticles you have found to expand your literature search. Find the relevant articles and add them to what you hace.
- each time you find an article that is relevant to your research question, read the article carefully and take notes or make a copy of the article for future reference. - If you are taking notes
o Get a compelte reference for the article.
Year of publication
Source of the article
o If it is a print journal
Name of the journal
- The DOI
o If the article is from an electronic source it will have a DOI
o Digital object identifier.
o This is important to take note of
Communicating your research
- You need to remember that a picture is worth 1000 words and indeed this is true for science and the picture can include illustrations and or graphs in science and it helps us summarize data in the very visual way usually when we see a picture we immediately capture the scene and we can immediately understand what the author is trying to say and that can be more effective than if we describe the figure in words so first we need to decide if the data should be displayed in a table or a graph
Data display mediums
- In deciding whether to use a table or a graph you want to consider multiple factors 1. Think about whether the independent variable and dependant variable are qualitative or quantitative.
2. Think about how many data points are being shown and
a. are you trying to make comparisons between data points.
b. Is the individual value more important than the over all trend?
c. Or are you trying to emphasize the overall trend in a table that can be useful if there are very few values to report.
3. You can use a table if both the independent variables are qualitative
4. In a graph it is useful is you are trying to show trends overtime and also when youre trying to visualise differences by comparing between groups.
a. This is because you get an illustrative/ visual representation of your data b. Example: if looking at whether group A versus group B has a difference. You can look at the height of the bars
- You can use the table to show comparison
- There are instances where you can use a table if theres a lot of values and both those values are qualitative.
- Example: in the table, it is only showing two values and you shouldn’t provide a graph in this case because the data can be easily interpreted from the table.
o this table shows the number of cells that are counted using this condition on these levels these two different modes types
- you can use a table to show numbers
o especially when there is a small number of values.
- You can use a table to show two qualitative pieces of information
o Example: the table below is showing the different brain regions and the expression of these two items or proteins and the plus value shows the relative expression.
- The left is typically reserved for an
- The dependant variables are in the
- In this case the independent variable
is the grey region and the dependant
variable is the expression level
- This table type is typically seen in
clinical papers to describe participant
o What are the baselin
characteristics of your
o What sample youre reporting
o At the top they’ve indicated:
Mean plus or mean plus ot minus standard deviation.
o They show the range (smallest and largest value)
o and the characteristic here would be your all of your variables like age weight BMI body fat blood pressure cholesterol levels etc
- Use graphs to illustrate trends
o Changes overtime if something is increasing or decreasing across different cohorts.
o You can use this to visualize differences between comparisons and it might contain the same information as a graph but you can cpture the relationship by a line graph.
- You can then take the graph and put the data points and put them into a table.
- but when it's in a graph form you can clearly visualize what the differences are
Choosing what graphs to use
- bar graph versus a line graph
- bar graphs and line graphs are useful when showing trends or comparisons. - Bar graph
o Best for discrete variables.
o If you only want to show at time point 15 weeks old what is the difference in body weights
o You will only get one time point
- line graph
o best for using or showing or expressing continuous variables where you look at changes overtime
o you can also use a line graph to make predictions and when you're trying to predict a value within the graph (interpolation)
o allows you to multiple time point
- interpolation when you're trying to predict value outside the data
o the body weight of the mouse at every single week so starting from when they were five weeks old
o Interpolation by looking at the values in the graph:
saying well what is the weight of the mouse when it say at 7.5 seven half weeks
so you're interpolating within data points
o extrapolating would be: youre going outside the values shown in the graph to make inferences about what would happen according to the trend shown what would be the weight of the mouse when it's at 20 weeks old so you're going outside the parameters of the graph to make predictions
- Bar graph example 2: differences between cholesterol levels in three different groups o So you can clearly see what the differences between each group is at one time point - show individual values if possible (the circles within the bars) o circles represent the individual animals that comprise the mean which showed which represented by the bar.
o The height of the bar represents the mean value.
The mean cholesterol level of the wildtype mouse
o The little white circles represent the individual mouse that comprise the mean.
- The mean is captured by taking the average of these eight data points and. o by taking the average of these eight data points and same is true for the Gray squares mouse and that can be illustrated here as well so in this case the dark line going across is the
o The dark line going across typically represents the standard deviation - Example of choosing between a bar and a line graph
o the researchers studied the effects of interrupted why on the excitability of a cell researcher applies NPY to brain slice for five minutes and records action potentials before during and after NPY application the results showed that NPY inhibits action potential firing by 90%
o and you're seeing a picture of action potential firing in the application of a drug You can see the membrane become more negative and then as the drug goes off you see the action potentials recovery or returning
you can report this in the table where the
left column is variable in the right column is dependent variable you are measuring the action potential frequency at baseline which is typically before drugs replied
and then during the NPY applications
how should the researcher show the maximal change in action potential firing over time?
- If you’re looking at things overtime you should use the line graph
How should the researcher show the maximal change in action potential firing ?
- If you’re looking at one point and looking at the maximum change then the bar graph should be used.
Anatomy of a table
- How to identify the independent and dependant variable given a table - left column you will have the independent variable and on the right column you have the dependent variable
o when you have to express a dependent variable to make sure you have units - Example: brain regions on the left column and the right columns express as much as possible the dependent variable and the dependent variable here is expressed using plus is the number of classes that the more the number of classes then the higher the expression of cells within these great regions how to label
Anatomy of a graph
- it's important to know what all the components of the axes or at least indicate the components on the graph
o Set up labels
o Use a legend
- Y-axis and sometimes X axis
o but X axis can be represented without the X axis title because it's understood that this is the condition treatments that are in.
o also include the units in brackets.
o a line graph also includes units express in brackets and it might indicate seconds minutes days
- Your x-axis should have ticks
o You should also try to indicate the number of subjects or participants that comprise the mean.
- your graph should also include statistical representation, social representation this would be like inferential statistics looking at comparisons between control and experimental
o you're making inferences as to whether or not these differences are statistically significant
- sometimes you can show subject numbers in brackets (for example inside the bar)
- but in line graphs you have to express your legend
o the legend should tell you what the circles represent and there should be numbers in brackets which represent the number of participants in each of the data points
- use Red green blue RGB format or CMYK format and wherever possible try not to use the colours there is one special consideration I want you to pay attention to and that is know that
Use of colour in graphs
- traditional papers will stick to very few colour choices some modern papers have colourful pictures and figures.
o You should try and avoid adding too much colour because then the reader has a hard time tracking what the representation is.
o Sometimes different subjects have different guidelines for colours to use in the formatting.
Colour – special consideration for illustrations.
- You should use magenta-green instead of red-green so that someone who is colour blind can easily understand what you are showing in the figure.
o 10% of the population is colour blind so there are some tricks you can use to make your science more inclusive.
- You can also use arrow heads to point to the cells you are referring to. - Example:
o Arrowhead is pointing to a red cell and it's trying to say that this cell expresses both the MCR one protein as well as dynorphin is indicated in this figure legend
Why we communicate research
- If Scientists only focus on their own solutions and research, they might not be able to advance as must as if they collaborated with other researchers. together they would better be able to come up with a more effective or more creative solution, so we use research to encourage new collaborations
o Collaborations is the process where multiple groups and individuals work together to come up with solutions and collaborations to help you gain new ideas and it also shared the burden of tackling large research projects together so you might not need to know everything.
- you want to collaborate and communicate your research because you might be able to tap into other sources of funding
- communicating your research is also good for career advancement
- poster presentations are the most common form of presentation - the society for neuroscience because there's a lot of people attend a conference and theres a team in your research area and you get to talk to them directly one-on-one about how they did something and discuss what's going on in your field and what are some of the latest findings
- highly interactive event
- in the paper poster it presents an overview of your recent research and it typically contains unpublished research
- an electronic poster which is presented on a screen
o allow you to incorporate videos and slides interactive charts
- in neuroscience there's a lot of videos and simultaneous recordings of neural activity with behavior and it's hard to capture that simultaneous coded coincident activity on a 2 dimensional paper so that's where the dynamic poster comes in
- Lecture: 1 hour
- Symposium: mini or nano symposium. You would have 20 minutes or 10 minute time limit.
- data blitz or like a lightning talk and this could be one to three minutes and you have to condense your 5-year-long research and summarize it into 3 minutes o well-developed presentation
o One slide, one minute
o Get people interested in your research
o Get to the bottom line, fast
o General tip – no more than 2 graphs
Things to consider when writing a manuscript
- This is the final part of the writing communication format. - Consider when writing a manuscript
o The purpose for writing a manuscript is to communicate research in the written form. Typically involving writing some kind of scientific paper - There are different types of papers in types of scientific article
o primary article, secondary article, and a systematic review article - Primary source
o Full research report containing all details needed to replicate study o traditional type of Journal article that has an introduction methods results discussion etc. it's a full research approach that contains all the details that you need to replicate that study and also the detailed findings
o Written by individuals who conducted the research
o what's happening within that project in depth and that primary article is written by the individuals who completed that research
- Secondary source
o Summarizes information from primary sources
o Descriptions of other people’s work- like a review paper
Third person format
Systematic review and summarizes information from different
primary sources so it might describe other primary literature.
(just gives you facts)
now these are facts right so textbooks describe facts but all those facts are summarized from hundreds of literature that you know his you know that has built up overtime to deliver that textbook knowledge Review papers
Provides recent overview of research field
Subjective upon the opinion/interpretation of the author
Is assembled by the author. Takes many different article and
summarized whats happening in the field. The review article
often summarizes (what this person thinks is going on in the
o Systematic review
Combines findings of all research articles that explored the same question to see if the result is statistically significant
Attempts to resolve differences between papers to get to the truth systematic review combines findings from all the research articles that have the same research question and sees if the result is statistically significant
o example: to take all of those 10 papers and does some kind of statistical analysis to see what the actual mean or overall effect of high fat diet is as described in these 10 papers, so the individual paper is going to look at within each authors data set what is the mean weight gain
o takes all the variability found between journal articles and does a statistical analysis to determine whether or not there is actually an overall effect based on the findings from all these primary articles.
- the systematic review is different from the review paper o because it's not just summarizing in words the trends that are appearing within the data
o the systematic review is going to take specific data sets and tries to crunch the numbers within those specific data sets and tell you what the overall effect of this diet (example) across these 10 papers is actually that have been shown so it attempts to resolve differences between papers
- neuroscience Journal has a rigorous peer review process
- a low impact factor doesn't mean that that paper is not very good it can still be extremely good except for some reason that Journal doesn't have a high impact factor
- the number of citations that an article gets will have an impact on the impact factor of the Journal
Choosing an appropriate scientific journal
- Journal should be peer reviewed
o that your paper has been reviewed or scrutinize if you will by your peers or by your supervisors peers
o helps you improve it. Critically analyze it for flaws. Looks to see whether or not your rationale leads from one point to another.
o a good peer reviewer actually helps you improve your paper that helps you improve the analysis your thought process
- the Journal you submit your paper two should be consistent with the aims and scope of the your paper
- select a journal that is appropriate for the topic of your research report. Most journals focus on a few special topics. A journal’s website describes what kinds of manuscripts are appropriate for that journal
o the journal should be consistent with the aims and scopes of your paper – found on the journals website
o the aims and scope is where the Journal tells you what kind of paper is going to accept
- but then it also tells you what it does not accept
o it says it does not accept the short communication because the expect your paper to be a full length investigation
short communications are where researchers try to publish a short article on a hot topic (new up and coming type things)
Journal of neurology is not interested in that it doesn't care about hot topics
- Consult the journal’s Instructions to Authors for specific submission requirements.
Factors that go into choosing an appropriate journal article - Impact factor – average number of citations an article in that journal received in the past two years
o impact factor is a number like it's 3.4 or 4.56 it gives you a sense of prestigious a Journal is
o in essence the impact factor is the average number of citations that an article in the Journal receives in the past two years
o If Journal of Neuroscience had an impact factor of 6 in 2015
o High impact factor may indicate important, cutting-edge work
o A high impact factor suggests that the higher number of times the papers within the journal is being cited. the more top papers being cited in that journal, the more impact it will have on that field.
o Consider: Proportion of review articles, scope of the journal
- Low impact factor may not necessarily indicate unimportant work
o Specialty journals naturally get less citations
- Citations – a reference to another scientific article
o Identifies the author(s) and year of publication of the specific fact/idea mentioned in the research report
o What counts as a citation?
- Beware of predatory journals
The peer review process
- When a manuscript is received by a journal editor, the editor informs the author of its receipt and distributes copies of the manuscript to reviewers.
- The reviewers are selected on the basis of their expertise in the research area of your manuscript.
- Reviewers provide the editor with an evaluation of the manuscript, but, ultimately, the editor makes the decision to accept it, reject it, or request its revision.
o they might criticize your analytical methods or it might request better referencing or citations of your paper etc.
- the editor reviews these comments
o and then it will send back to you the comments from the peer reviewers o then you have an opportunity to take the peer reviewers comments and revise and resubmit your article for a second round of peer review
- if they still don't like it or they still have criticism
o then the cycle continues
- Then the reviewers accept it
o and at that point there's no more edits that need to be made and then you get a published Journal article
- so this whole process is the peer review process how long does a peer review process take well it depends on the Journal some journals have a strict policy for making sure that public papers are reviewed in a timely manner
Trend of a preprint
- Research paper that is shared before peer-review
- any kind of like archive is a preprint server and the preprints can either be a primary or secondary article
o Often has a DOI
o Can be primary or secondary article
- Why a preprint?
- Benefits of a preprint
You can get credit right away
you're going to submit something at the peer review and it's taking so long and you're in competition with another group to get this paper out first then you're kind of worried that as it's going through peer review maybe that reviewer is like going to steal your ideas or something then you want you can submit this as art by archive so that you can claim your steak in the pot so to speak can get credit for doing it first
you can get feedback
You can get more visibility so more people are seeing your results early - Concerns
o Its more likely to get accepted because the peer review process might be influenced.
o it's not to review so if you have like some famous researcher puts in the preprint to archive and it's not peer review but everyone just assumes that because it's really famous person that it must be good then that's the danger
APA format of a research proposal
A research proposal
- A research proposal is a written report presenting the plan and underlying ra- tionale of a future research study. A proposal includes a review of the relevant background literature, an explanation of how the proposed study is related to other knowledge in the area, a description of how the planned research will be conducted, and a description of the possible results.
- there is usually a word limit the purpose of the abstract is to provide a one paragraph summary of your whole paper it should include background information it should tell you how you did somehow get the research it should have key results and have the major conclusions
- you have to be able to summarize your paper into just a few sentences for your abstract
- the Journal neuroscience definitely has a weight limit 650 words and here you want to provide enough background information to clarify
o why your research was completed
o you want to point out the objectives of the study
o tell me what you know about the field
o what is not known about the field
o then what did you actually test in your research study
o tell me briefly what you found (key results)
o this is also where you will find the hypothesis of the research
o methodology- inclide details such as the species studies, gender, general methods
- May have word limit
- Indicate the objectives of the study
- why you are doing this research
- Provide enough background information to clarify why the research was completed - State the hypotheses tested
• Describe how the research study is conducted in depth with many details. • There is never a word limit on the methodologies because is important for other reader or another researcher to replicate your findings.
• But you want to be as brief as possible but not leave out any details • Include animals used, strain, species, gender, etc.
• Tell the reader im using ___________
• have a statement indicating all of your protocols have been approved by your animal care committee or the research ethics board at your University
• reagents section
• software used
• equipment used
• identify where the company makes it and where is was produced • important because example:
• you can buy substance P from a wide number of companies but strangely enough sometimes in pharmacology experiments different companies produce different priorities of the compound and for whatever reason
• not all compounds for every company have the same efficacy and effectiveness
• example: don't change the source of my peptides especially within the same research project that could be a confounding factor
• saying the company and the city and location is important because you know you might have the company called back in that's in the United States but then you might also have it in Germany and if they're being produced in two different places it might also have a different quality • the last part is the statistics/ data analysis section
• indicate that the software that you're using
• any types of statistical tests you use whether it's the task I swear test • whether or not you were doing a factorial design
• indicate what your P level your P value indicates so is your statistical significance indicating that P less than 0.35
• talk about mean, errors and standard deviation
• indicate whether or not your data is presented as mean standard deviation or mean and error
- then within the results of your paper you want to clearly and succinctly described our findings and here is where you provide numerical data
- the discussion will have a word limit
1. (first paragraph) summarize your findings
2. Provide an interpretation of what your findings mean
a. Connect with the literature
b. Use a lot of citations here so that you can explain whether or not your findings make sense with what is known in the literature
c. compare and contrast
3. explain if there are any limitations with your work
4. end your discussion with a statement of why your work is important and how your work contributed to the field
- another section within the Journal of manuscript you submit and is usually attached at the end of the paper
- sometimes they want you to indicate or order it by alphabetical order - So what you need to know that is what goes into the reference the reference itself contains
o the authors
o the year
o the title of the manuscript
o the journals name
o the volume of the Journal
o and the page range so within that Journal volume where you going to find it so I want to explain to you what volume
- so if you were given a reference you should be able to identify that the first words or authors last names the middle and first name initial the year that the manuscript was published the title of the article and then the abbreviation for the Journal name
- the figure of legend is a text paragraph that describes the figure so it's a description of the figure and not an interpretation of the figure so it tells you what the figure is but not what the figure means
- figure titles are really important because they should be able to stand alone so I should be able to look at a figure without reading the paper and understand what your results are
- also have statistical representation and tell me what these stars mean so in this case these stars are in abbreviation for a P value that's less than 0.0001 cheers another example of a figure in a figure legend so it's a theoretical one
Concise summary of the paper that focuses on what was done and what was found.
Typically written last after the
paper is done
Subjects- how many and relevant characteristics Research method and procedures
A report of the results
Statement about the conclusions or implications
Begins with a general introduction to the topic of the paper.
knowledge (theory) behind the research study. Identifies the
problem and why it is
The introduction presents a logical development of the research question, including a review of the relevant background literature, a statement of the research question or hypothesis, and a brief description of the methods used to answer the question or test the hypothesis
Provides a relatively detailed
description of exactly how the
variables were defined and measured and how the research study was conducted. Other researchers should be able to read your method section and obtain enough information to determine whether your research strategy adequately addresses the question you hope to answer. It also allows other researchers to duplicate all of the essential elements of your research study.
Divided into sections: 1)
Describe the sample that participated in the study. The number of subjects,
For animals: describe their genus, species, and strain, the supplier, how they were housed and handled, and their specific characteristics including weight, age, sex.
For humans: 1)eligibility, and exclusion criteria, 2) the settings and location in which data were collected 3)any payments made, 4) ethical standards, 5) any methods used to divide or assign participants into groups or conditions, how many individuals in each condition. 6) a description of instruments given to participants 7) research design 8) any experimental manipulation or intervention 9) apparatus or materials used
Provides a complete and unbiased reporting of the findings just with facts no discussion of the findings. – provides a summary of the data and the statistical analysis
Tables or figures are included here
A statement of the primary outcome of the study. 2) the basic descriptive statistics (the means and standard deviation) 3) inferential statistics (results of the hypothesis test/ P-value) 4)the measures of effect size.
offer interpretation, evaluation, and discussion of the implications of your findings
If your results support your original hypothesis, it is now possible to test the boundaries of your findings by extending the research to new environments or different populations. If the research results do not support your hypothesis, then more research is needed to find out why.
1)restate hypothesis and relate it to existing literature 2)restate major results and indicate if they support or fail to support your primary hypothesis 3) relate your results to the work of others explaining how your outcome fits into the existing structure of knowledge 4)identify any limitations of the research especially factors that affect the generalization of the results 5) in the last paragraph you extrapolate beyond actual results and consider their implications/ applications
is a listing of complete references for all sources of information cited in the report, organized alphabetically by the last name of the first author.
Each item cited must appear in the references, and each item in the references must have been cited in the body of the report. One-author entries precede multiple-author entries beginning with the same first author.
Supplement the test. They do not duplicate information that has already been presented in text form.
Each on their individual page. Must have a title. Can add general notes which refer to the entire table. Specific notes which refer to items in the table that have been identified with superscript, lower-case letters. Probability notes which identify the level of significance for statistics reported in the table that have been identifies with asterisks
Supplement the test. They do not duplicate information that has already been presented in text form.
Each on their own page. Usually have a colour specification for types of papers.
may be included as a means of presenting detailed information that is useful but would interrupt the flow of text if it were presented in the body of the paper.
Examples of items that might be presented in an appendix are a copy of a questionnaire, a computer program, a detailed description of an unusual or complex piece of equipment, and detailed instructions to participants. Appendices each start on a new page
Responsible scientific conduct
- Increase knowledge of issues surrounding responsible conduct of research - Ability to make ethical or legal choice when faced with research conflict - Understand that there is a range of acceptable research practices
- Adopt a positive attitude involving research ethics and responsible conduct - Scientific articles are retracted infrequently
- Retracted articles continue to get cited!
- Articles are retracted infrequently
o It is difficult for an article to be retracted
- Retracted articles are still cited after being retracted
o This is bad because the fraudulent information continues to be propagated and cited as good and true information.
- Fraud: is the explicit effort of a researcher to falsify or misrepresent data
o Making up or changing data to make it support a hypothesis
o Plagiarism – the representation of someone else’s ideas or words as your own. - Safeguards against fraud
Is the repetition of a research study using the same basic procedures used in the original.
Either the replication supports the original study by duplicating the results
OR it casts doubt on the original study by demonstrating that the original
result is not easily repeated.
o Peer review
The editor of the journal and a few experts review the paper and critically
scrutinize every aspect.
- Publication bias
o Positive outcomes are more likely to be published than null findings (nonsignificant results – or the experimental treatment didn’t have any effects)
o Nonsignificant findings are more likely to be submitted to journal articles with a lower impact factor
This is because higher impact factor journals are less likely to publish them
o This is bad because researchers need to know when a treatment didn’t have an effect so they don’t continue to try the treatment and continue to get nonsignificant results
- Citation bias
o Researchers are more likely to cite papers that report significant findings Editors realize that researchers are less likely to cite articles with nonsignificant results so this leads to a feedback loop and thus
nonsignificant findings are less likely to be published
o Women authors are less likely to be published than men
Responsible conduct of research
- Tri-council agency framework: responsible conduct of research
o The tri-council makes up a framework for responsible conduct of research that researchers should always try to follow
- Tri-council made up of
o CIHR - Canadian Institutes of Health Research
o NSERC – Natural Sciences and Engineering Research Council of Canada o SSHRC – Social Science and Humanities Research Council
- a major source of research funding for Canadian institutions
o and thus researchers at those institutions who get funding from the tri council must follow their framework
- even if a researcher got private funding
o the funders may still require the researcher to follow the tri-council agency framework for responsible conduct of research
o this is because the private funder will want the research they are given to be conducted in the most ethical way
o and they want to receive a good research article which has produced good results.
Tri-council research integrity principle
o Making up data/methods/findings
o Including graphs or figures
o Manipulating or omitting data without acknowledging the omission This leads to inaccurate findings or conclusions
o Includes figures and graphs
3. Destruction of research records
o Destroying your own or another person’s research data/ records o Done to avoid detection of wrongdoing
o Presenting another persons work as your own, without appropriate referencing or permission
5. Redundant publications
o Re-publication of your own previously published work – without justification
i. Trying to get credit for the same thing twice
o When you publish an article that you have already published but now you change a couple things
6. Invalid authorship
o Inaccurate attribution of authorship – you don’t list all the authors of the article.
o Fail to recognize the contributions of other people in a manner that is consistent with how much they have contributed to the work
i. You have to give credit where credit is due.
7. Inadequate acknowledgement
o Failure to recognize the contributions/ authorship of other people in a manner that is consistent with the policies of relevant publications
i. Neglecting to add someone to the author list – they can then write to the editor saying that you didn’t acknowledge them but you contributed.
8. Mismanagement of conflict of interest
o Failure to manage any real, potential or perceived conflict of interest i. Bribery
1. Bringing timbits could be misconstrued as bribery.
ii. Perceived conflicts of interest
1. If you are a stockholder in Coca-Cola and you do research on
hydration – if you don’t declare you are a stake holder in the
Coca-Cola company who might have a conflict of interest where
you might have a personal gain from your research
- Is a form of scientific fraud
o You present someone else’s ideas or words as your own
o present another person’s ideas or words without referencing the person who said it or without properly referencing.
Any direct quotes should appear with quotation marks and should be correctly cited. Otherwise this is plagiarism
- All references to previous studies should be correctly cited
o State who made the claim, where they made the claim, and when they made the claim
- Minor cases of plagiarism can be unintentional
o If you read an article and then time passes and you think of the information you read in the article but you don’t remember you read it. You think that this idea is your own. So you don’t reference it and you present it as your own.
o To avoid this
Make a habit of paraphrasing all information you read in an article right when you read it
o These breaches in plagiarism can happen easily so it is important to take steps to prevent it from happening
o When you present information in a new article that you already published in a previous article.
- There are differing ideas/ opinions of what plagiarism is/ isn’t o Different people think different things are plagiarism/ are not plagiarism o People take misconduct into context. There are differing opinions on the same act of plagiarism
Guidelines for authorship contributions
- Design study and interpret results
o If you created an original idea, helped planned it and had an intellectual contribution = yes
o You had to have contributed to the actual subject + intellectual contribution o If you’ve been trained to do basic things = no authorship.
Basic is if anyone could do it
- Supervisory role
o Supervision of the project?
o Training, education, mentoring the first author
The person training you must produce substantive contribution in order to deserve authorship.
o Resources: money, animals/reagents
o If you are the person funding the research you would be acknowledged but not as an author.
Usually the supervisor contributes to the original idea, provides
direction to the study and sometimes does work in the experiment so they would get authorship.
- Data acquisition
o Original experimental work = authorship
o Technical experimental work = no authorship
o Data analysis
You get authorship for data analysis unless it is very basic
The person writing the entire paper is the first author. If you are writing parts of the paper then you are going to be an author.
o Editing/ reading/ commenting on the manuscript
Reading it over = no authorship
Unless in the process of editing you are providing substantial edits then that can be acknowledged
Preventing authorship disputes
- Be selective about your collaborations
o Collaborate with someone who you are intellectually engaged with and share a passion for the research and science.
- Share ideas, reagents, expertise with no strings attached
o If they regard your contributions as valuable don’t be stingy about sharing. By sharing you might end up making a significant contribution to the project. - Divide the labor at the start of the experiment
o Establish in writing the tasks that will be completed on your part of the collaboration
Know who is completing and contributing what to the research.
In a collaboration where both parties are equally engaged, there is mutual trust and such there is no need for this.
Bring it up in a conversation if you aren’t sure
- Respect and communication
o Discuss progress frequently with the group.
o Anticipate and make corrections to problems in a timely manner. o Be willing to share information and reagents
o May re-negotiate authorships as project develops
Because as the project evolves there could be warranted renegotiation of authorship.
How fraud gets through - Background information
- The 8 forms of misconduct
- Peer-review process
o Reviewer goes through your paper to try and point out flaws.
o Sometimes make a lot of edits.
- New information
o How can fraudulent data still get through
- Can be difficult to spot fraud
o This is one reason that fraudulent data gets through the peer-editing process.
- Most research papers are peer-reviewed prior to being published o Usually, the reviewer looks at specific things – they trust you and thus they don’t look so much at fraud.
o They are not paid to peer-reviewed articles so they might not be spending as much time.
o Publishing articles is oftentimes on a strict timeline and thus the editor may not have enough time to comb through
- Example: 1997: Editors at the British Medical Journal took a short paper and inserted 8 errors, then asked researchers to identify the mistakes – the researchers were also told that there were 8 errors in the paper
o 221 people took part in the study
o Most people only spotted 2 errors. No one found more than 5. And 16% didn’t find any errors.
Videos and articles posted on CULearn for use as background knowledge.
Sample size outlier’s exclusion criteria
- A person was trying to reproduce results in their experiment, but they were not able to reproduce the results.
- They were in contact with the original researcher about not being able to reproduce the results and he sent over his raw data
- The original researcher said that the control samples he used have been in the freezer for years (may have degraded over time)
o The original researcher failed to mention that they had kept the samples in the freezer for years
o The new researchers had been preparing the control samples fresh.
- The original researcher was very open to their input and findings. They offered to run the experiment again using a fresh control sample and then compare their results with the new researcher.
- Dr. Chee’s comment
o Points out that there are some factors we need to pay attention to
o It is the responsibility of the author to include all of the methodology used so that other researchers can replicate it with the same results.
The original author and the replicating author have a responsibility to complete their work transparently.
Lack of transparency
- New researcher could not reproduce the results of the original researcher. - The new researcher was following the same experimental procedure - One of the new researchers’ supervisors said to toss out anything greater than 10 because it is outside the original parameter.
- The other supervisor gave a suggestion for the data analysis
- Results were still inconclusive and were not consistent with the experimental results.
- The supervisor looked over the original researchers lab note book o Findings were that she excluded samples without a real reason. o The original researcher had only used female mice.
o Rigorous research requires considering sex as a biological variable when using any animal samples
- Dr. Chee’s comment
o Why you cannot exclude data points are you please
o You cannot change parameters in the middle of the experiment. o The supervisor was expecting to see a certain set of results because the original researcher provided results that they liked.
o The conclusion was that there was a mistake in the original paper and its findings
o Don’t be a bully
It makes it more likely for something to go wrong and increases the pressure for someone to take a risk to falsify results in the research field.
Article: “the science behind social science gets shaken up – again” Rogers A. The Science Behind Social Science Gets Shaken Up—Again. Wired. Published August 27, 2018. Accessed December 5, 2020. https://www.wired.com/story/socialsciencereproducibility/
- Research showed that “people see emotional impulsive people as inherently more honest”
- The results did not replicate. New studies showed that the original study failed an important step in science validity being replicability as you are supposed to be able to produce the same results if the experiment were to be redone.
- New research purpose was to study reproducibility – the question was whether or not replicators could rule out some of the excuses for why the results weren’t reproducible.
- 5 teams retested the results using larger groups
o 13 has a statistically significant effect in the same direction as in the original o The effect was half as large as the original paper reported.
o The new papers showed zero effect
- Using Altmetrics it is possible to analyze the effects of published scientific articles. - Comments
o The fact that the original paper wasn’t replaceable doesn’t mean the conclusions were false.
o Experiments can fail to replicate for many reasons.
Resource constraints and methodological arguments
- “The solution to the reproducibility crisis is better training, statistics and institutional practices that’ll stop these kinds of problems in research before they make it to pages of a journal.”
- Dr. Chee’s comment
o Experiments fail to replicate for a variety of reasons
o Maybe better training, statistics etc. could prevent it
o If a paper failed to replicate, we need to make sure we look at all the conditions, it is not always the case that the conclusions are false
NPR- science replication issue (podcast)
- Most experiments cant be reproduced with the same results.
- Reproducibility crisis
- The main effect was that stereotyping and priming has an effect on math outcomes
- We cant be quick to make judgments when researchers fail to replicate a finding - Because there are multiple factors that come into play
o Sample size
- Examples of stereotyping
o Diet drinks – people often think that women ordered the diet coke when it could just as likely had been the man
- Replication must closely match the conditions
- If a study failed to replicate – does this mean that there is fraud and scientific misconduct happening?
o If it does not hold in alternative studies that not necessarily a failure to replicate, but may contribute a more nuanced view on the original finding - Any individual study is not necessarily the truth
o There would always be more things to uncover
o Become more certain about our knowledge when we are able to replicate it. Data reproducibility
- Transparency – accurately and openly providing all key information on the design, execution and analysis
- Adequate reproduction is possible only if the methods, rationale, any pertinent information is accessible
o Human – gender, age, demographics, medical history
o Animals – gender, age, strain, background, N
o Reagents – where you sourced your resources (company, city, state), antibodies
- Reporting of only positive findings
o If you're only reporting positive findings and ignoring the null findings then it will make it hard for someone else to reproduce them because they might be seeing that the findings are not correct.
o This is the responsibility of both the original and replicating researchers
Principles to facilitate scientific rigor
- https://www.nih.gov/research-training/rigor-reproducibility/principles-guidelines reporting-preclinical-research
- Report the number of replicates and whether results substantiated by repetition - Statistics must be fully reported
o Test used, N, definition of center/dispersion measures
o What were the statistical tests used
- State whether samples were randomized and how
o Indicate if your samples were randomized
o If you were blinded to the scoring process
State if experimenters were blinded to group assignment
o Did you state the inclusion/exclusion criteria
- Clearly state inclusion and exclusion criteria, including any results omitted o Did you omit anything – if you did indicate what your criteria for exclusion was
o How sample size is determined
o Power analysis
- Especially if results did not support main findings
How can you facilitate this? Lab notebook
- Your lab notebook is a legal document
- the scientific legacy of the lab –
o after you leave the lab your lab notebook remains and it is the legacy you leave behind
- Complete record of procedures, reagents, data, thoughts to pass on to others
o Explain why experiments were created
o How the experiments were performed
o And what the results of the experiment were
o Your lab notebook can be used to validate findings and defend your claims o It is the evidence for the validity of your work
- The lab notebook is NOT:
o Record of communications with PI/colleagues – not where you make notes about conversations with the supervisor. Only record your findings from each day
o Yours to take home
Belongs to the institution.
You have the right to make photocopies or take pictures of your work - Notebooks can come in different forms
Type of Notebook
Advantages - pros
Disadvantages - cons
No lost pages, Numbered pages
Difficult to photocopy, organized by date (not logically organized)
Organize by experiment, data stored together
Sheets fall out, difficult to authenticate
Easy to search, read, store
Requires electronic security, corrupted files, software compatibility
Contents of a lab notebook
- Name, year, lab mailing address
o So it can be returned to the lab if lost
- Table of contents
o Page number, date, subject/experiment
o A record of all the work that was done organized by date and experiment type
- Body of notebook
o Indicate the experimental entries
o Observations Everything that happened, planned or unplanned, raw data, taped into page or data location
Date, title, hypothesis/goal
reagents, solutions, time course, concentrations
Reagents you were using only if you are changing it for that day
New bottles of the reagent
What reagent you are using
Otherwise you can just give reference to the lab procedure in the lab notebook
o Data analysis
- Dr. Melissa Chee lectures posted to CULearn between November 1st and December 5th
Gravetter FJ, Forzano LAB. Research Methods for the Behavioral Sciences. Wadsworth; 2012. ?
What is the purpose of a cause and effect relationship? The purpose of an experimental research strategy is to establish the existence of a cause-and effect relationship
The research process
1. Develop a research idea
2. Formulate a hypothesis and prediction
3. Determine how to define and measure variables
4. Identify subjects for the study
5. Select a research strategy
What is a variable
- Independent variable
o What you manipulate
o Manipulation: of the independent variable creates two or more conditions. Helps determine the cause
- Dependant variable
o What you measure
o Measurement of the dependant variable produces scores from each condition
o Comparison of the sores in one condition to those from another condition help indicate the effect
- Confound variable
o Variables that are not held constant and affects some groups only
o Control of other conditions ensure that the other variables do not influence the outcome on the dependant variable
- Obscuring variable
o Increases variation in the data and affects all groups
Research scenario 1
A researcher is interested in determining whether a new drug, Drug A decreases aggregation of AB plaque deposition in the hippocampus. 5XFAD transgenic mice overexpressing mutant human amyloid precursor protein were purchased from Jackson Laboratory. The 10 mice were divided into two equal groups, Group A and Group B. Both groups were housed in the animal care facility at a temperature of 23 degrees. The day before testing, the fire alarm went off in the facility causing prolonged distress among both groups of mice. On the day of testing, Group A mice received an injection of the
placebo cocktail. Unfortunately, the research felt sick and had to go home, so another researcher stepped in to inject Drug A into Group B mice. Two weeks later, the mice were sacrificed, and the researchers conducted a histological
analysis to examine AB plaque deposition in the hippocampus. The results showed no significant difference in amyloid plaques aggregation in the hippocampus among mice that received Drug A.
Was there manipulation?
- The researchers manipulated Drug administration by changing the value to create a set of two treatment conditions (Drug A vs Placebo) Was there a control group?
- The group of mice that received a placebo drug were used as a control group in this experiment.
Did the change in researchers administering the drug act as a confound?
- Yes because the change in researcher affected one group Did the fire alarm act as an obscuring variable?
- Yes because the fire alarm affected both groups
Would this study have high internal or external validity - Internal
- The researchers controlled all variables not being manipulated thus increasing the internal validity of the study.
Is this an experimental research strategy
Research scenario 2
A researcher is interested in determining the link between Instagram usage and the development of depressive symptoms among adolescent females. Participants were recruited via a paid advertisement on Instagram, Facebook, and Twitter from 2019 to 2020. The researchers collected survey data measuring the average number of hours spent on Instagram a day and depressive symptoms from 700 participants across Ontario. The average age of participants was 15 ± 2.3 years. Of the 700 participants, 40 were excluded from the study due to incomplete surveys. In the middle of the study, a global pandemic shut down the country, requiring public spaces to close and people to social distance. The mandatory safety measures enforced varied across cities. Following the analysis, the researchers determined there was a significant relationship between Instagram usage and depression scores, with higher Instagram use being associated with greater depression symptoms.
Was there manipulation
- The researchers administered a survey and did not manipulate any variable
Was there a control group
- There was no control group that acted as a control, scores for depressive symptoms and Instagram usage was taken from all participants
Did the variation in mandatory safety measures across different cities act as a confound
- Yes because the mandatory safety measures affected the participants differently
- The mandatory saftey measures affected the participants differently depending on where they lived. Individuals living in a city with strict saftey measures like social distancing and park closure bylaws may report greater depressive symptoms than individuals living in a city with less restrictive measures. Did the pandemic act as an obscuring variable
- Yes because the pandemic affected all participants
- Both groups were affected by COVID19 and this could have influenced the scores of all participants. Would this study have high internal validity or high external validity - High external
- This study does not have many controls and mimics the real world, it is generalizable to the population in which the study was conducted on.
Is this an experimental research strategy
- The researchers were not interested in a causeandeffect relationship, they simply wanted to understand the relationship between two variables.
Learning module #2 – compare, contrast and describe the five research strategies
- Recognize research questions
- Compare and contrast the five research strategies
- Apply our knowledge to real world examples
After the researcher has identified the research area, formed a hypothesis, decided how to define and measure the vairables, determined which individuals should participate in the study and how to treat them ethically, the next step is to select the research strategy
- Defined as the general approach and goals of the research strategy
- Seleting the research strategy is typically determined by the question the researcher wants to address and the kind of answer the researcher hopes to obtain
What is the relationship between the average number of hours spent
This question is asking about
the relationship between two variables (average
studying and test performance among second year neuroscience students?
number of hours spent studying and test performance). The type of research strategy selected should provide a description of the relationship.
Is the average number of hours second year neuroscience students spend studying more or less than the average number of hours second year political science students spend studying?
This question is asking about a single variable (the number of hours spent studying). The type of research strategy selected should provide a description of the variable.
Can the type of media (electronic vs paper) used to study for an exam by second year
neursoscience students cause higher or lower grades?
This question is also asking about a
relationship between two variables (type of media and grades), however, this question asks for an cause for the relationship. The type of research strategy selected should provide an explanation for the relationship
The five research strategies
- Establish unambiguous cause-and-effect relationship
- Compare pre-existing groups to obtains results. Attempt
to control treats to internal validity
- Compare pre-existing groups to obtain results. No
attempts to control treats to external validity
- Describe the relationship between two variables and
measure the strength of the relationship
- Describe individual variables as they exist
- Psychiatric symptoms and disorders among Yazidi children and
adolescents immediately after forced migration following ISIS attacks
Identify the research question
- What is the occurrence of psychiatric problems and disorders
experienced by Yazidi Kurd refugee children and adolescents
- This information can be found at the beginning of the summary section. The researchers were interested in describing a single variable (psychiatric problems and disorders).
Identify the independent variable
- There are no independent variables in this study
Identify the dependant variables
- Psychiatric problems and disorders
- The researchers measured the occurrence of psychiatric problems and disorders among Yazidi Kurd children and adolescents.
What research strategy was used
- the researchers were describing psychiatric symptoms and disorders of a specific group of children who underwent forced migration and related traumatic experiences. There was no manipulation, control, or treatment conditions.
Suppose the researchers wanted to conduct a follow up study to examine the relationship between the number of traumatic experiences and the rates of psychiatric symptoms and disorders among Yazidi Kurd children and adolescents. What research strategy should they use?
- In this scenario, the researchers were interested in describing the association/relationship between two variables. There was no manipulation, control, or treatment conditions.
Journal 2 Experimental study of the differential effects of playing versus watching violent video games on children's aggressive behavior
Identify the research question
- Does the active involvement of playing a violent video game increase aggression than the passive observation of watching the same violent video games?
- Found in the last paragraph of the introduction
Identify the independent variables
- Video game conditions
- The researchers manipulated the video game conditions (Active violent condition vs passive violent vs active non-violent condition). Identify the dependant variable
- Aggressive behaviour
What research method was used
- This study satisfies the four basic elements of an experiment. The independent variable is video game conditions and the dependent variable is aggressive behavior.
Suppose the researchers wanted to conduct a follow up study to determine the average number of hours a week the participants play violent video games. What research strategy should they use? - Discretional
- on this scenario the researchers are interested in describing one variable (average number of hours spent playing video games).
Assignment #3 –
1. A researcher tests the effect of stress on the tail suspension test. The following data shows the percentage of time tha a mouse is immobile in a tail suspension test.
a. State the null hypothesis
i. Stress has no impact on the percentage of time that a
mouse is immobile in a tail suspension test
b. What is the mean score for each condition
i. Nonstressed: 42.5
ii. Stressed: 45
c. Name the independent variable and dependant variable i. Independent: stress
ii. Dependant: % immobile
d. The above data is best summarized by
i. Bar graph
e. The outome of the experiment yielded a p-value of p=0.42 f. What is the probability that there is no differene between non stressed and stressed mice?
i. 42% chance there is no difference
g. Can the researcher reject the null hypothesis
h. Is there a statistically significant difference in immobility time between the stressed and non-stressed mouse?
2. With reference to a 3 x 2 x 3 factorial design answer the following questions:
a. How many independent variables are in this factorial design i. Three
b. How many effects are possible in this experimental design i. 3
3. In the sample excerpt below, indicate (yes or no) if a citation is required:
4. Autistic spectrum disorders (ASD) typically begin in childhood and are characterized by difficulties with social interaction, communication, and behavioral flexibility. There is a speculative association between ASDs and violent crime offenders (Reference A). Hippler et al. (2010) found that rather than being more likely to engage in violent behavior, individuals with ASD have a higher risk of being the victim rather than the perpetrator. It is necessary to stress these findings regarding neurodevelopment and violent crime in order to avoid stigmatizing an already vulnerable group (Reference B).
a. Reference A
b. Reference B
5. Use the abstract obtained from PubMed to answer the following questions:
a. What is the volume
b. What is the PMID for this publication
c. Is this a primary or secondary article?
6. Use this description to answer the following questions A researcher determined if caffeine can offset the debilitating effects of alcohol on
motor skills test. The factorial matrix below summarizes the results of the rat experiment.
a. List the specific conditions for the experiment
i. Top left: Placebo+ water
ii. Top right: Placebo +alcohol
iii. Bottom left: caffine + water
iv. Bottom right: caffeine + alcohol
b. How many effects
c. Is there an interaction between alcohol and caffine on motor score
d. The researcher used a mixed group design for this experiment and included 15 rats per group. How many rats were used in total?
i. 30 rats
7. Illustrate the above factorial data as a graph. Remember to label axis, legends, etc. Please draw the graph on a separate piece of paper and upload a photo below.
Methods of acquiring knowledge summary
accepted as fact because it is
believed to be true -Can become
superstition of habit
-This information could be outdated and no longer true.
-Brains are hardwired to look for patterns, it is then easy to turn patterns into habits
holding on to ideas and beliefs simply because they have been accepted as facts for a long time or because of superstition. Therefore, the method of tenacity is based on habit or
superstition. Habit leads us to continue believing something we have always believed. Often this is referred to as belief perseverance.
accepted based on a feeling something is right.
-Often based on subtle cues that you perceive
- We can’t solely rely on intuition we also have to test the information. -Often times we think our gut feeling is right
One problem with the method of tenacity is that the information acquired might not be accurate. With regard to the statement about old dogs not being able to learn new tricks, the elderly can and do learn
information is accepted as true because it “feels right.” With intuition, a person relies on hunches and “instinct” to answer questions. Whenever we say we know something because we have a “gut feel- ing” about it, we are using the method of intuition. For example, at a casino, if someone puts his money on the number 23 at a roulette table because he “feels” it is going to come up, then that person would be using the method of intuition to answer the question of which number to play.
-When an expert in the field gives
information and they are trusted -Quickest and
easiest way to get an answer.
-You accept the information because you believe the authority figure. -Journal impact factor: Just because it is published in a journal doesn’t mean the information is always correct.
It does not always provide accurate information. For example, au- thorities can be biased. We have all seen examples of conflicting testimony by “expert witnesses” in criminal trials. Sources are often biased in favor of a particular point of view or orientation.
a person finds answers by seeking out an author- ity on the subject. This can mean consulting an expert directly or going to a library or a website to read the works of an expert. In either case,
you are re- lying on the assumed expertise of another person
accepted based on use of reasoning or logic
-The premise statement is assumed to be true. If any part of the premise statement is incorrect then the conclusion could be false.
- Used to create research idea
involves seeking answers by logical reasoning. We begin with a set of known facts or assumptions and use logic to reach a conclusion or get an answer to a question. Suppose a clinical psychologist wanted to know whether a client, Amy, had a fear of darkness. A simple example of
reasoning that might be used is as follows:
All 3-year-old children are afraid of the dark. Amy is a 3-year-old girl.
Therefore, Amy is afraid of the dark.
the first two sentences are premise statements. That is, they are facts or assumptions that are known (or assumed) to be true. The final sentence is a logical conclusion based on the premises. If the premise statements are, in fact, true and the logic is sound, then the conclusion is guaranteed to be correct. Thus, the answers obtained by the rational method must satisfy the standards established by the rules of logic before they are accepted as true.
Empirical/ Observatio nal
accepted based on personal
-Observations you make become
points of data and your knowledge becomes solely based on that data.
- Data collection - Used to create
-Variables are not operationally defined.
-If using an abstract concept of measurement (bigger/harder) observations can be misconstrued. -if only based on observations everyone would have different data points. Fundamental flaw -Our interpretations influence out observations. Interpretations can be biased based on expectations.
It is tempting to place great confidence in our own observations. Everyday expressions such as, “I will believe it when I see it with my own eyes,” reveal the faith we place in our own
experience. However, we cannot necessarily believe everything we see, or hear and feel, for that matter. Actually, it is fairly common for people to misperceive or misinterpret the world around them.
attempts to answer ques- tions by direct observation or personal experience. This method is a product of the empirical viewpoint in
philosophy, which holds that all knowledge is acquired through the senses. Note that when we make observations, we use the senses of seeing, hearing, tasting, and so on.
Messages such as “opposites attracted” or “you can’t teach an old dog new tricks” have been repeatedly presented and accepted as true.
However, it is possible for older people to learn new skills. Similarly, research has shown that it is more common for people to be attracted to others who have similar interests.
Often times, the decision is determined by what do I “feel like doing. Going out or staying in?”
If you don’t know the answer to a multiple-choice question and you have just answered 4 “d” in a row, you are less likely to guess “d” even though the answer could still be “d”
It does not always provide accurate information; authorities can be biased. A behavioral psychologist will give you a different answer than a psychodynamic psychologist.
If parents ask a behavioural psychologist, why their child is throwing tantrums they will hear “you’re reinforcing the behaviour by giving into the demands of the child” If parents were to ask a psychodynamic psychologist, they might hear “ because of a failure to meet the child’s oral needs.”
People are not good at logical reasoning. Consider the following argument;
All psychologists are human. Some humans are women. Therefore, some psychologists are women.
Many people would view this as a sound, rational argument. However, this is not a valid argument; specifically, the conclusion is not logically justified by the premise statements. Replace Women and psychologist with apples and oranges All apples are fruits. Some fruits are oranges. Therefore, some apples are oranges.
Empirical/ Observation al
We cannot believe everything we experience because sometimes we want to believe something is true simply because it is what we have always done or thought to be true.
If someone told you that you could pick between gross potato chips and delicious noodles, you would pick the noodles. But if they told you the noodles were worms; you would pick the chips. This is because you believe that people do not eat worms (method of tenacity).
Non-scientific methods of knowledge are not obsolete.
Non-scientific methods of knowledge can be used in conjunction with scientific methods of knowledge.
But in order to rely on information founded on one of the non-scientific methods of knowledge you must confirm the information using scientific methods of observation.
Non-scientific methods of knowledge are not obsolete.
Scientific methods of acquiring knowledge
It has a systematic approach, which means it has a set of steps.
Better quality answers with a higher confidence in validity can be achieved by
o A greater data representation. Meaning more participants/ subjects. o With more data, you can be sure your results are reproducible.
o Data reproduction is necessary and important in validity and confidence. Steps in the Scientific Method
1. Identify a research area
a. How you can think of a good idea
1. Source of Ideas
i. Pick a general topic of interest
ii. Do research to find out what is known and what remains unknown. iii. Do background research on what is unknown.
iv. Narrow it down to a specific idea.
Past research, your observations, others observations, practical problems Make sure the idea is interesting, novel, fundable, publishable.
o Novel: It must be a new idea that has not been researched already o Fundable: Will you be able to find someone to fund it?
o Publishable: Will other researchers want to read your article? Will other researchers want
the scientific method
The scientific method is a method of acquiring knowledge that uses observations to develop a hypothesis, then uses the hypothesis to make logical pre- dictions that can be empirically tested by making additional, systematic observations. Typically, the new observations lead to a new hypothesis, and the cycle continues.
As you know, when we say that science is empirical, we mean that answers are obtained by making observations. Although preliminary answers or hypoth- eses may be obtained by other means, science requires empirical verification
Step 1: observe behaviour or other phenomena
o Introduction, or inductive reasoning, involves using a relatively small set of specific observations as
the basis for forming a general statement about a larger set of possible outcomes · 1.4 The research process
· 1. Find a research idea: Select a topic and search the literature to find an unanswered question.
o Identify a general topic you would like to explore and review the background literature to find a specific research idea or question.
· 2. Form a hypothesis and a prediction
o Form a hypothesis, or tentative answer to your research question, and use a specific research prediction.
o Logical: A good hypothesis is usually founded in established theories or developed from the results of previous research. Should be the logical conclusion to a logical argument.
o Testable: it must be possible to observe and measure all of the variables involved. Must involve real situations, real events, and real individuals. A testable hypothesis is one for which all of the variables, events, and individuals are real, and can be defined and observed.
o Refutable: it must be possible to obtain research results that are contrary to the hypothesis. A refutable hypothesis is one that can be demonstrated to be false. That is, it is possible for the outcome to be different from the prediction.
o Positive: it must make a positive statement about the existence of is a relationship, the existence of a difference, or the existence of a treatment effect.
Another half mark for making it specific
Diet soda volume or amount of weight gain
Can be proved wrong
If diet soda did not increase weight gain
You have information and make a prediction about what will happen less.
You predict something based on specific examples and give a general statement of all
- systematic Plan to collect data
o once you collect the data, you have to analyze the data to see if it supports your hypothesis. If not you have to make a new observation in order to continue your research.
PREDICTION: Deduction, or deductive reasoning, uses a general statement as the basis for reaching a conclusion about specific examples.
- Deductive reasoning uses a general hypothesis or premise to generate a prediction about specific observations.
HYPOTHESIS: a hypothesis is a statement that describes or explains a relationship between or among variables. A hypothesis is not a final answer but rather a proposal to be tested and evaluated.
- Inductive reasoning uses a few limited observations to generate a general hypothe- sis. · 3. Define & measure variables
o Identify the spefic procedures that will be used to define and measure all variables. Plan to evaluate the validity and reliability of your measurement procedure.
· 4. Identify and select participants or subjects
o Decide how many participants or subjects you will need, what characteristics they should have and how they will be selected. Also plan for their ethical treatment.
o Participants in research studies are called participants if they are human and subjects if they are not human.
· 5. Select a research strategy
o Consider internal and external validity and decide between an experimental (cause-effect), or a nonexperimental strategy.
· 6. Select a research design
o Decide among between-subjects, withing-subjects, factorial, or single-subject designs. · 7. Conduct the study
o Collect the data
· 8. Evaluate the data
o Use the appropriate descriptive and inferential statistics to summarize and interpret the results.
you must use various statistical methods to examine and evaluate the data. This involves drawing graphs, computing means or correlations to describe your data, and using inferential statistics to help de- termine whether the results from your specific participants can be generalized to the rest of the population.
· 9. Report the results
o Use the established guidelines for format and style to prepare an accurate and honest report that also protects the anonymity and confidentiality of the participants.
· 10. Refine or reformulate your research idea
o Use the result to modify, refine, or expand your original research ide, or to generate new ideas. ·
1. Test the boundaries of the result: Suppose your study demonstrates that higher levels of academic performance are related to higher levels of self- esteem for elementary school children. Will this same result be found for adolescents in middle school? Perhaps adolescents are less concerned
about respect from their parents and teachers, and are more concerned about respect from peers. Perhaps academic success is not highly valued by adolescents. In this case, you would not necessarily expect academic success to be related to self-esteem for adolescents. Alternatively, you might want to investigate the relationship between self-esteem and success outside academics. Is there a relationship between success on the athletic field and self-esteem? Notice that the goal is to determine whether your re- sult extends into other areas. How general are the results of your study?
2. Refine the original research question: If your results show a relationship between academic success and self-esteem, the next question is, “What causes the relationship?” That is, what is the underlying mechanism by which success in school translates into higher self-esteem? The original question asked, “Does a relationship exist?” Now you are asking, “Why does the relationship exist?”
Examples of hypotheses
Drinking diet soda “affects” body weight gain
- You’re not showing its direction
Make a hypothesis example Nicotine decreases cholesterol
o Identify the independent and dependant variable in the hypothesis
o Independent variable would be the thing that youre manipulating = nicotime o Dependant variable is the thing you are observing = cholesterol
o Bad hypothesis
Nicotine affects cholesterol o Is it good/bad increase/decrease
o Can measure the variables
o Must be a bit more specific than just nicotine
What type, what form o Or just cholesterol
What type of cholesterol HDL/LDL, etc.
MINE: Participants who use nicotine patches will show a decrease in total cholesterol levels compared to participants who do not use nicotine patches.
HERS: Participants wearing a nicotine patch will have reduced levels of total cholesterol 3. Define and measure variables 3A- Types of variables
There are two main types of variables
Independent variable (X-AXIS)
o Manipulated variable
o Determined by the experimental design/ researcher o Known in advance o Determined by the treatment conditions
• Dependant variable (Y-AXIS) o Responding variable
o Measured in each of the treatment conditions
o Determined during the course of the experiment
3B- define what you are measuring
Be specific and unambiguous
Must be measurable and defined qualities. You need an operational definition to be able to
Break down concepts into something that can be measured
• Examples of things that you can measure.
o Health: cholesterol, blood pressure, BMI, heart rate o Intelligence: Exam scores, IQ test scores, transcripts
3C- Clarify how to measure variables
Measurements must have the following characteristics
o You are measuring what you claim to be
• How presice can your measurement be o Property of your instrument
Calibration, precision o How much error is there?
How much room for error can you have in what you’re measuring?
o Measurements are consistent, produce nearly identical results when used repeatedly
When you step on a scale and you get 150 the first time you should get close to this amount when done repeatedly. If you don’t than that instrument is not reliable.
o Relates to replicability
4. Identify research participants/ subjects
Neurons, cells, tissues
o human, animal, cell model • Demographics
o Species, sex, age, weight, etc.
o You can’t say that a vaccine protects from death if you only tested it on tissue. You have to prove it is safe for animals and then move on to humans. • Sample size
1. Start with a research idea
2. Formulate a hypothesis and prediction
1. Hypothesis – is something that is a generalization, generalizations use the inductive reasoning
2. Predictions- is something that is general and then you form something more specific using deductive reasoning
3. Inductive reasoning generates hypothesis because you’re making a general statement
4. Deductive reasoning forms prediction because you’re producing more specific outcomes
3. Determine how to define and measure variables
1. Independent variables is what you manipulate
2. The variable that you measure is the dependant variable
4. Identify subjects for the study
1. Participant is human
2. Subject is nonhuman
Defining and Measuring Scientific Variables
The Research Process
Basic Terminology: factor vs condition vs level
→ Differentiates between a set of groups being compared in an experiment. → Independent variables of an experiment
o Animal handling
o Drug administration → Example
o In an animal experiment “does drug A increase locomotor activity of a mouse” The factor would be the type of drug being used
The factor could also be the type of locomotor activity that is used.
→ A factor is ultimately different categories of independent variables that can occur in your experiment. Condition
→ How is the group treated in an experiment?
o Handled for 10 minutes and given 10 mg of drug o Not handled and given 10mg of drug. Levels
→ Different values of the independent variable that are selected to create the treatment condition → What is the mouse experiencing, how long is the mouse experiencing that for.
o Animal handling
Handling vs no handling of animals
Handling for 10, 20, 30 minutes o Drug administration
Drug dosage: 10, 30, 100 mg of dosage Time course: 1, 2, 3, 4 hours
→ Variables are any condition that can change or have a different values for different individuals. → Goal of experiment - Is to find out how variables can change under different conditions o Example: “how fast is a neuron firing during sleep and wake?” Variable is the neuron firing o Example: how much do mice move after amphetamine treatment?
Variable is how much the mouse moves
o In order to answer the questions we have to be able to define a variable
→ It is important to define the variable
o Defining a variable ensures a clear understanding of how your question is being answered.
How is neuronal firing measured frequency, or counting the number of spikes over a set period of time. How is locomotor activity measured distance traveled in cm, # beam breaks, # of wheel revolutions
o Without defining the variables it is unclear to the reader how you are quantifying the variables. Characteristics of a variable
o Can be measured directly or indirectly
o Can be measure consistently
→ In order for the above to be true, the variable must be specific and unambiguous → Must have at least 2 levels per value
o Sleep/ wake (also known as the arousal state) o Glasses/ no glasses
o Drug dose: 1,5,10 mg
→ If you give someone a drug you test if it works by also observing participants without the drug. Operational definitions
→ Some variables are not easily observed or quantified
→ Converts an abstract
→ Example: constructs are variables that are more abstract (intelligence or health). Providing a definition for
the construct is the operational definition.
→ Good operational definitions are:
o Clear, precisely articulated
o Makes replication possible
→ Example of formulating a good operational definition
o Construct anxiety
o Variable heart rate
o Operational definition heart rate exceeding x bpm
→ Improve the validity and reliability of your measurement.
o Accurate measurement
o Whether or not what you’re measuring is even accurate → Reliability
o Consistent measurements
o You consistently get the same measure
o If you put a mouse on a scale, how consistent is your measurement.
→ A variable that describes the situation or environment.
→ These variables are characteristics within the experiment, but they do not have to be an independent
variable. It can be something that you want to control
o What aspects of the environment changes for the subjects?
→ For example
o Amount of drug
Mice injected with variable amounts of drug o Time of day
Firing rate of a neuron in the light and dark cycle.
→ Response variable is almost always a dependant variable
→ We can measure the effect of a situational variable on a response variable. You often manipulate a
situation to record a change in the response.
→ Example: does caffeine boost memory performance, weight loss, athletic performance?
o Situational variable: Whether or not the participant is drinking caffeine
o Response variable: weight loss, athletic performance, memory performance.
→ Differences between individuals
→ Constant within individuals, variable between individuals
o Genetic composition
→ Helpo explain why a relationship exists between two other variables
→ A mediating variable facilitates the interaction between your independent and dependant variable.
→ Typically found in psychological theories and in treatment/ prevention
Mediating variables in psychological theories
→ Example: whether or not social support will lead to a patient taking home more positive health practices o Situational variable: amount of social support available to a participant
o Mediating variable: how many positive health practices is that participant practicing. o They determined that increasing social support increases positive health practices.
o Mediating relationship: This relationship is mediated with the patient feeling less lonely and increasing hopefulness.
Mediating variables in treatment/ prevention
→ Example: COVID-19 - non-white Americans with coloured Americans. The infection rate and death in the coloured community is experiencing a higher rate of morbidity and infection.
o We don’t see an equal result between population and percentage of total cases. o Situational variable: race
o Mediating variable: living environment
o Response variable: COVID severity
Qualitative vs Quantitative variables
Qualitative Variable- described in words
→ Described in words
→ Important when observing changes
o Behaviour change
o Depression (bipolar/unipolar) o Drug use (
Qualitative Variables – described in numbers
→ Can be counted o Age
o Caloric intake
o Number of food pieces
Continuous vs. Discrete variables
→ The reference of discrete variable for a quantitative variable is typically a description of its category
→ Measured in categories
o Distinct, separate values
→ Usually measured as a whole number
→ Is obtained by counting
o Number of pieces of chocolate o Number of children
o Number of cells
→ Measured on a continuum
o Any value in a range
→ Whole unit or fraction
o Can have decimal placed
→ Obtained by measuring
o Calories of chocolate
o Is measured on a continuum o Weight, height, age, time
Can have decimal places o Temperature
Can have decimal places o Duration of drug exposure.
Introduction to measurement
→ Aim for variables that are well defined, easily observed, easily measured
→ Some aspects of research, typically neuropsychology, examine intangible abstract phenomena variables
that cannot be seen / measured directly = constructs
o Motivation, anxiety, intelligence
→ Create operational definitions to measure these constructs o Identify behaviours associated with your construct
Tell the reader what you are defining as a behaviour o Measure that behaviour o Resulting measurements helps define your construct
o This process = operational definition
→ A systematic or step wise process in defining what you are measuring and how you are going to measure it.