Exam 2 Study Guide
Exam 2 Study Guide PSY 290
Popular in Intro to Research Methods
Popular in Psychlogy
This 28 page Study Guide was uploaded by Mira Kawash on Thursday October 22, 2015. The Study Guide belongs to PSY 290 at University of Miami taught by Rick Stuetzle in Fall 2015. Since its upload, it has received 811 views. For similar materials see Intro to Research Methods in Psychlogy at University of Miami.
Reviews for Exam 2 Study Guide
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 10/22/15
102215 1236 PM Experimental Research Experiment 0 Investigator directly varies some factors holds all else constant and observes results of systematic variation 0 You only let one thing change and everything else is constant 0 Factor that is varied Independent Variable o Factors held constant Constant Variable 0 Result that is observed Dependent Variable Independent Variable Minimum of 2 levelsconditions to make comparison 0 Ex Dose 1 vs Dose 2 Types of Independent Variables o Situational Variables Variations in a feature of the environment participant encounters a Ex Doorintheface effect 0 Task Variables Variation in type of task performed by participants a Ex Levels of complexity of puzzle 0 Instructional Variables Variation in instructions given tot groups of participants a Ex Go as fast as you can vs Get as many right as you can 0 Combination of Different types of IVs Ex I Crowding room size a Task Difficulty easy vs difficult a Motivation win 1 vs 5 Multiple IVs allow for hypotheses re interaction effects Control Groups as IVs Getting treatment or not getting treatment as independent variable 0 Group getting treatment experimental group 0 Group without treatment control group The control group should all be identical except for the independent variable Participants in control group are identical in all ways to those in experimental group except for experimental treatment Controlling Extraneous Variables Extraneous Variable 0 Any uncontrolled factors not of interest to researcher but which might influence the behavior being studied Confound any uncontrolled extraneous variable that n Covaries with IV and I Could provide alternate explanation of results Measuring Dependent Variables Dependent Variable measured outcomes of experiment 0 Ex TV violence and aggression Quality of study depends on choice and operational definition of dependent variables Subject Variables as Iv Compare groups based on already existing characteristics of individuals 0 Gender 0 Age 0 Personality Characteristics Manipulated vs Subject Variables o How could you study the effects of anxiety on problem solving behavior with anxiety as A manipulated variable A subject variable Studies using subject variables are sometimes called ex post facto studies or quasi experiments Experimental Validity Does your experiment provide understanding about behavior you intended 0 Recall we talked about the validity of specific measures in Chapter 4 0 Here we are referring to the validity of an entire experiment Types of Experimental Validity Statistical Conclusion Validity 0 When making conclusions about the associations between the independent and dependent variables in a study Are statistics used appropriately Are Statistics interpreted appropriately 0 Example 3rd graders dividing into 4 reading groups based on teacher recommendations Scale of measurement for reading groups Principal wants to look at proportion of males vs females being placed into each group Conducts a ttest to look at mean group for boys vs girls a Problem Conclusion Construct Validity o Adequacy of definitions of IV s and DV External Validity o How well do the findings from a study generalize beyond the specific context of the experiment Do the results generalize to other populations Also referred to as ecological validity I Are the results applicable in the real world a It works on paper but what about the real world Example Adequacy of women s health initiatives based on studies will all male participants a Ex Baltimore Longitudinal Study of Aging 0 Comes from how you got your samples If you used random sampling you have high generalizability it represents the sample of interest You must generalize the results with caution 0 Age Cohort Effects Your cohort are a group of people that are similar to you in several factors Ex Baby Boomers 102215 1236 PM Subject Variables as IV Why can t we draw conclusions about causality when the IV is a subject variable rather than a manipulated variable 0 You don t control them 0 Your results can be due to chance 0 Your results weren t something that was called for o It can be due to other factors that aren t related to your IV 0 Worry about Confound Variables when it covaries with the IV To what degree is a study methodologically sound and confoundfree o Are most results most likely due to IV and not the result of some other confounding factor If so the study has high internal validity Threats to Interval Validity PrePost Design 0 History Where an event happened between the pre and the post that could influence the outcome Ex Hurricane o Maturation Change in the organism due to changing Problematic when you are dealing with children n Ex Test someone in 3rd grade or in 6th grade Cognitive and physical abilities Also problematic with the elderly 0 Regression Return If you have extreme score then you are more likely to go back towards your average score Your prescore was an extreme score naturally your next score will move back regress towards your average in the absence of any independent variables Comparing your before and after might not have anything to do with the experimenter but just because its naturally supposed to be like that There usually is a problem because the experimenter doesn t know about the extreme score 0 Testing Just taking the pretest has an effect on posttest scores a Practice effects I Pretest sensitization Taking a test before makes you more sensitized o Instrumentation When measurement instrument changes from the pretest to posttest Did the instrument itself change in the pre and the post which can affect the outcome Participant Problems 0 Subject Selection People who choose to be in the study and people who are selected to be in a study 0 Attrition Over a period of time you lose subjects Loss death PrePost Studies 0 Will people change as the result of some experience Evaluate behavior before pretest Give experience Evaluate behavior after posttest 0 Ex Effectiveness of test anxiety intervention program In first week of class freshmen will fill out test anxiety questionnaire Students scoring high are asked to participate in a training program relaxation training study skills etc After program they retake test anxiety questionnaire I At pretest mean score 90 I At posttest mean score 70 How will you interpret this I Just based on the numbers did the program work Yes I Can I say that conclusively No Using a control group allows experimenter to account for threats in prepost studies 0 Ex What would you conclude regarding history and or maturation of these cases Anxiety test Experimental Group Pretest 90TreatmentPosttest 7O Control Group Pretest 90 No Treatment Posttest 90 I History does not explain the difference n Mature does not a Not Regression either everyone moves not just one group n Intervention makes sense Experimental Group Pretest 90Treatment Posttest 7O Control Group Pretest 90 No Treatment Posttest 70 n Intervention does not explain it a History or Regression might explain this 0 Subject Selection Effects In an experiment need to be sure that participants in different conditions are equivalent to each other except for the independent variable a Ex Effect of teaching method on performance Students choose whether to take lecture or lecturediscussion format class What cancan t you conclude about effectiveness of teaching methods at the end of the semester 0 Attrition Especially when studies span a good deal of time participants do not always complete the experiment they begin You want to know who drops out of the study a Specifically how are they similar or different from the people who continued with the study Can differential attrition account for your findings Experimental Research Two general types of experimental design 0 Betweensubjects or independent groups designs Different groups of participants contribute data for different levels of the independent variable Withinsubjects or repeated measures designs 0 Same participants contribute data to all the levels of the independent variable Independent variable with levels A amp B 0 Ex effects of music volume on studying Level A quiet music Level B Loud music Between Subjects Design 0 Each participant is observed in either Level A or Level B but not both That is each group or level represents a different condition of the independent design Within Subjects Design 0 Each participant is observed in both Level A AND Level B BetweenSubjects Design Depending on your question may be the only way to design study 0 If independent variable is a subject variable Introverts vs Extroverts Males vs Females If experience of participating at one level of independent variable makes it impossible to participate in other levels 0 Ex Eyewitness memory for details preceding violent vs nonviolent events Viewing one version of a film make it impossible to then have participants view the other version without biasinfluence of prior influence Pros 0 Each participant comes in equally nai39ve about procedures Cons 0 Need a large number of individuals to fill all experimental conditions 0 Differences between conditions may be due to independent variable you hope so BUT could also be due to PREEXISTING DIFFERENCES BETWEEN THE GROUPS Need to create equivalent groups Ways to create equivalent groups 0 Random Assignment A method of placing participants into different groups a Every participant has an equal chance of being placed in each of the groups n Spreads potentially confounding factors evenly through out the different groups BUTcannot ensure equal number of participants in each condition think about if you flipped a coin to decide which group 20 participants would be in 0 Block Randomization Used to ensure an equal number of participants per group a Each condition of the study has a participant randomly assigned to it before any condition is repeated a second time 0 Matching Participants are grouped together on some trait and then randomly distributed to the different groups This is especially useful when there are too few participants available for random assignment to work well Two Conditions needed for matching a You have a good reason to believe that the matching variable is correlated with the dependent variable I You need a good way to measure the matching variable Matching Example In Experiment comparing two forms of writing emotional and control on disease progression in HIV patients a Match Patients on the following When they were diagnosed how long have they been HIV Are they someone who used to write a lot How they contracted it The amount of stress Other medical complications a Each member of matchedpair then assigned to the different methods a The matching variables selected above are good because they are variable that likely correlate with disease progression and if left uncontrolled would threaten the internal validity of the study 102215 1236 PM Experimental Research Independent variable with levels A amp B 0 Ex effects of music volume on studying Level A quiet music Level B Loud music Betweensubjects Design 0 Each participant is observed in either Level A or Level B but not both That is each group or level represents a different condition of the independent variable Within Subjects Design 0 Each participant is observed in both Level A and Level B WithinSubjects Design Every participant is exposed to every level of the independent variable 0 Also called a repeatedmeasures design Often used in studies of sensation and perception 0 Ex Fig 61 Muller Lyer illusions Pros 0 Fewer people needed to fill all experimental conditions Ex 2 conditions and you need 20 people per condition a Total N for between subjects design a Total N for withinsubjects design Don t need to worry about equivalent groups Cons 0 Sequence or order effects Once a participant has completed the first part of a study the experience could influence performance in later parts of the study 0 Progressive Effects After Trial I performance is better than Trial 2 practice effect After Trial I performance is worse on Trial 2 fatigueboredom o Carryover Effects AB has different effect than BA Controlling Sequence Effects 0 Counterbalancing Present conditions in more than on sequence a Primarily used to minimize progressive effects a Why is it less effective at minimizing carryover effects Types of counterbalancing when testing only once per condition a Complete counterbalancing every possible sequence of conditions is used at least once Problem n Partial counterbalancing a subset of total possible sequences are used Counterbalancing effects a Types of counterbalancing when testing more than once per condition Reverse counterbalancing for each participant present the conditions in one order and then again the reverse order 0 ABCD DCBA a Block randomization every condition occurs once before any condition is repeated a second time Within each block the order of conditions is randomized BCDA CADB n Demand characteristics Characteristics in a study that demand your attention The aspects of a study that haul your attention to what you are testing Control Problems in Developmental Research When independent variable is age 0 Two types of design to consider Crosssectional Design a A betweensubjects design Ex 3 yr 4 yr 5 yr language performance a Problem Cohort effects 0 Longitudinal Design A withinsubject design a Ex compare language performance of same kids 34 amp5 yrs ofage n Attrition can be a problem a Feasibility of collecting data for 3 years Between or Within In a study of eyewitness memory a researcher wishes to determine whether the accuracy of eyewitness memory can be influenced by the level or stress an eyewitness experiences 0 Between In a crosscultural study of prejudice a researcher wishes to known whether prejudices would develop earlier in life for Western or Eastern Cultures 0 Between In a study on the sense of touch blindfolded participants had to judge whether the apparatus touching their skin had two points or one The researcher wished to determine if different areas of the skin ex the plan of the hand vs small of the back were differentially sensitive 0 Within 102215 1236 PM Problems with Biasing Bias preconceived expectation about what should happen in an experiment 0 Who could be biased Experimenter Participant Both Experimenter Bias As experimenter you might inadvertently interact with participants in a way that will make them behave in a way that will confirm your hypotheses 0 Ex Rosenthal studies of picture perception with experimenter expectancy as independent variable Positive bias expectancy participants give higher ratings Negative bias expectancy participants give lower ratings How is this possible Some way experimenter bias might be communicated 0 Ways in which instructions are given Describing anchors on scale 0 Facial expressions when answers given Participant changes response to try to get positive reaction from experimenter 0 Personality characteristics Ex preschoolers perform better on cognitive task when experimenter is caring versus indifferent Ways to control for experimenter bias 0 Minimize contact between experimenter and participant Computerize assessment Doubleblind procedure a Have research assistants who are quotblind to hypotheses a Hard in practice though Ex High positive vs High Negative feedback Participant Bias Knowing that you are in an experiment can make you change your normal behavior 0 Hawthorne Effect read Box 62 Belief of participant that they are part of special group and focus attention n Regardless of whether changes in independent variable are positive or negative get positive outcomes NOT due to independent variable Hawthorne Study Know this study a They were examining the differences between worker conditions a They were trying to productivity in factory workers a They changed the lighting incentives temperature etc a The workers knew they were part of the study to a point a Groups that had improved studies obviously improved 0 Efforts to present self as a good subject Evaluation apprehension 0 Demand characteristics Is part of your experiment revealing your hypotheses u If so will change way participants behave Do you think demand characteristics would be more of a problem in n Between subjects designs a Within subjects designs Have more demand characteristics 0 Why Because being exposed to both levels it calls for more demand characteristics 0 Controlling participant biases Minimize demand characteristics a Use of deception to get participants to behave more naturally 0 Use of placebo control group Everything same expect content of treatment a As much special attention etc etc o Manipulation Check Ask participant what they think the hypothesis is Between Subjects Single Factor Designs 1 Independent Groups 0 Effectiveness of weightloss programs Independent variabe Weight loss consequence a Level 1 Selfreward for weight loss a Level 2 selfpunishment for weight loss Dependent variable pounds lost over 8 week program What would you expect data to look like after 8 weeks n Why a Why is this an example of a study that has to be conducted as a between subjects design 2 Matched Groups 0 Effects of sleep deprivation on responses to interrogation questions 0 Independent variabe sleep deprivation Level 1 Deprivation 21 hours awake in lab Level 2 Normal sleep pattern at home 0 Matching Variable To control for what How does this help clarify conclusions 0 Independent variabe score on suggestibility tests 3 Nonequivalent Groups 0 Are cognitively gifted children also gifted socially and emotionally 0 Independent variable degree of giftedness Level 1 Gifted IQ 130 Level 2 Average IQ 90110 0 Independent variabe Social Emotional Problem Solving Test Score Within Subjects Single Factor Designs 4 Repeated Measures 0 What are the effects of a moving environment on children s balance Independent variabe direction of motion of Moving Room a Level 1 forward motion 10 trials a Level 2 backward motion 10 trials 0 Dependent variabe direction of infants body lean or fall How do you know if the two levels differ Interested in the difference between the mean scores 0 Does one level of your independent variable do better than the other Inferential statistics like a ttest will tell you whether this difference is larger than you expect due to chance alone 0 T test for independent groups o T test for dependent groups Chapter 5 102215 1236 PM Chapter Outlines Hypothesis Testing Procedure for deciding whether the outcome of a study results for a sample supports a particular theory or particular innovation which is thought to apply to a population Hypothesis A prediction often based on observation previous research or theory that is tested in a research study Theory A set of principles that attempt to explain one or more facts relationships or events behavioral and social scientists often derive specific predictions hypotheses from theories that are then tested in research studies The central theme of hypothesis testing has to do with the important distinction between sample and population discussed in the previous chapter Hypothesis testing is a systematic procedure for deciding whether the results of a research study which examines a sample support a hypothesis that applies to a population The Hypothesis Testing Process Step 1 Restate the Question as a Research Hypothesis and a Null Hypothesis about the populations Research Hypothesis Statement in hypothesis testing about the predicted relation between populations usually a prediction of a difference between population mean The opposite of the research hypothesis is that the populations are not different in the way predicted Null Hypothesis Statement about a relationship between populations that is the opposite of the research hypothesis a statement that in the population there is no difference or a difference opposite to that predicted between populations a contrived statement set up toe examine whether it can be rejected as part of hypothesis testing The research hypothesis and the null hypothesis are complete opposites if one is true the other cannot be In fact the research hypothesis is sometimes called the alternative hypothesis Step 2 Determine the Characteristics of the Comparison Distribution If the null hypothesis is true Population 1 and 2 are the same Comparison Distribution Distribution used in hypothesis testing It represents the population situation if the null hypothesis is rue It is the distribution to which you compare the score based on your sample s results The comparison distribution is the distribution that represents the population situation if the null hypothesis is true Step 3 Determine the Cutoff Sample Score on the Comparison Distribution at which the Null Hypothesis Should be Rejected Cutoff Sample Score In hypothesis testing the point on the comparison distribution at which if reached or exceed by the sample score you reject the null hypothesis is also called critical value Convention Levels of Significance plt005 p lt001 The levels of significance widely used in the behavioral and social sciences Statistically Significant Conclusion that the results of a study would be unlikely if in fact the sample studied represents a population that is no different from the population in general an outcome of hypothesis testing in which the null hypothesis is rejected Step 4 Determine Your Sample s Score on the Comparison Distribution Once you have the results for your sample you figure the Z score for the sample s raw score You figure this Z score based on the population mean and standard deviation of the comparison distribution Step 5 Decide Whether to Reject the Null Hypothesis To decide whether to reject the null hypothesis you compare your actual sample s Z score Step 4 to the cutoff Z score Step 3 If the researchers reject the null hypothesis what remains is the research hypothesis Implications of Rejecting or Failing to Reject the Null Hypothesis First when you reject the null hypothesis all you are saying is that your results support the research hypothesis Terms such as prove or true are too strong because the results of research studies are based on probabilities What you do say when you reject the null hypothesis is that the results are statistically significant You can also say that the results support or provide evidence for the research hypothesis Second when a result is not extreme enough to reject the null hypothesis you do not say that the results supports the null hypothesis You simply say the result is not statistically significant A result that is not strong enough to reject the null hypothesis means the study was inconclusive Showing the null hypothesis to be true would mean showing that there is absolutely no difference between the populations When a result is not extreme enough to reject the null hypothesis the results are said to be inconclusive It is also important to bear in mind that just because a result is statistically significant it does not necessarily mean that it is important or has practical or theoretical implications OneTailed and Two Tailed Hypothesis Testing Directional and OneTailed Tests Directional Hypothesis Research hypothesis predicting a particular direction of difference between populations for example a prediction that the population like the sample studied has a higher mean than the population in general OneTailed Test Hypothesistesting procedure for a directional hypothesis situation in which the region of the comparison in which the null hypothesis would be rejected is all on one side or tail of the distribution A onetailed test can be onetailed in either direction Nondirectional Hypothesis and TwoTailed Tests Nondirectional Hypothesis Research hypothesis that does not predict a particular direction of difference between the population like the sample studied and the population in general Twotailed Tests Hypothesistesting procedure for a nondirectional hypothesis the situation in which the region of the comparison distribution in which the null hypothesis would be rejected is divided between the two sides tails of the distribution When to Use OneTailed or TwoTailed Tests If the researcher decides in advance 0 use a onetailed test then the sample s score does not need to be so extreme to be significant compared tow hat would be needed with a two tailed test In principle you plan to use a onetailed test when you have a clearly directional hypothesis You plan to use a twotailed test when you have a clearly nondirectional hypothesis If the twotailed test is significant then the researcher looks at the result to see the direction and considers the study significant in that direction If you do get a significant result with a twotailed test you are more confident about the conclusion Decision Errors Decision Error Incorrect conclusion in hypothesis testing in relation to the real but unknown situation such as deciding the null hypothesis is false when it is really true Decision errors are possible in hypothesis testing because you are making decisions about populations based on information in samples Type I Error Type I Error Rejecting the null hypothesis when in fact it is true getting a statistically significant results when in fact the research hypothesis is not true You make a Type I error when you conclude that the study supports the research hypothesis when in reality the research hypothesis is false Type II Error Type II Error Failing to reject the null hypothesis when in fact is is false failing to get a statistically significant result when in fact the research hypothesis is true Chapter 6 102215 1236 PM The Distribution of Means The comparison distribution has been a distribution of individual scores such a the population of ages when individual babies start walking A distribution of individual scores has been the correct comparison distribution because we have used examples with a sample of one individual The score you care about when there is more than one individual in your sample is the mean of the group of scores Distribution of Means Distribution of means of samples of a given size from a particular population also called a sampling distribution of the mean comparison distribution when testing hypotheses involving a single sample of more than one individual The scores in a distribution of means are means not scores of individuals The distribution of mean is the correct comparison distribution when there is more than one person in a sample Building a Distribution of Means The only information you need is 1 the characteristics of the distribution of the population of individuals 2 the number of scores in each sample Determining the Characteristics of a Distribution of Means The three characteristics of the comparison distribution that you need are as follows 1 Its mean 2 Its spread which you can measure using the variance and standard deviation 3 Its shape Mean of a Distribution of Means The mean of a distribution of means of samples of a given size from a particular population it comes out to be the same as the mean of the population of individuals Rule 1 The Mean of a Distribution of Means is the same as the mean of the Population of Individuals Population MM is the mean of the distribution of means It uses the word Population because the distribution of means is also a kind of population Population M is the mean of the population of individuals Variance of Distribution of Means Variance of the population divided by the number of scores in each sample Rule 2a The variance of a distribution of means is the variance of the population of individuals divided by the number of individuals in each sample A distribution of mean will be less spread out than the population of individuals from which the samples are taken For a particular random sample to have an extreme mean the two extreme scores would both have to be extreme in the same direction both very high or both very low The more individuals in each sample the less spread out will be the means of the samples Here is Rule 2a stated as a formula Population SDAZM is the variance of the distribution of means Population SDA2 is the variance of the population of individuals and N is the number of individuals in each sample Standard Deviation of a Distribution of Means Square root of the variance of the distribution of means same as standard error SE Rule 2b The standard deviation of a distribution of means is the square root of the variance of the distribution of means Population SDM is the standard deviation of the distribution of means Standard Error SE Same as standard of a distribution of means also called the standard error of the mean SEM It has this name because it tell you how much the means of samples are typically in error as estimates of the mean of the population of individuals Rule 3 The shape of a distribution of mean sis approximately normal if either a each sample is of 30 or more individuals or b the distribution of the population of individuals is normal Shape of Distribution of Means Contour of a histogram of a distribution of means such as whether it follows a normal curve or is skewed in general a distribution of means will tend to unimodal and symmetrical and is often normal With fewer extremes there is less asymmetry A distribution of means tends to be symmetrical because a lack of symmetry skew is caused by extremes The more individuals in each sample the closer the distribution of means will be to a normal curve The Three Kinds of Distribution a The Distribution of a population of individuals b The Distribution of a particular sample taken from that population c The Distribution of means of samples taken from that population Hypothesis Testing with a Distribution of Means The 2 Test Z Test Hypothesistesting procedure in which there is a single sample and the population variance is known Figuring the Z Score of a Sample s Mean on the Distribution of Means The method of changing the sample s mean to a Z score is the same as the usual way of changing a raw score to a Z score The Z score for the sample s mean on the distribution of means is the sample s mean minus the mean of the distribution of means divided by the standard deviation of the distribution of means Hypothesis Tests about Means of Samples 2 Tests in Research Articles Standard errors are often shown in research articles as the lines that go above and sometimes also below the tops of the bars in a bar graph these lines are called error bars Sometimes the bars are not for standard error bars but instead are standard deviations or confidence intervals Advanced Topic Estimation and Confidence Intervals Estimating the Population Mean When it is Unknown The variation in means of samples from a population is the variation in the distribution of means The standard deviation of this distribution of mean s the standard error of the mean is thus a measure of how much the means of samples vary from the overall population mean Range of Possible Means Likely to Include the Population Mean Confidence Interval CI Roughly speaking the region of score that is the scores between an upper and lower value that is likely to include the true population mean more precisely the range of possible population means fro which is its not highly unlikely that you could have obtained your sample mean Confidence Limits Upper or lower value of a confidence interval To get a sense of how accurate our estimate is we can use our knowledge of the normal curve to estimate the range of possible means that are likely to include the population mean This estimate of the range of means is called the confidence interval The 95 and 99 Confidence Intervals 95 Confidence Interval Confidence interval in which roughly speaking there is a 95 chance that the population mean falls within this interval You want the area in a normal curve on each side between the mean and the Z score that include 475 99 Confidence Interval Confidence interval in which roughly speaking there is a 99 chance that the population mean falls within this internal You use Z scores for the middle 99 of the normal curve Steps for Figuring the 95 and 99 Confidence Intervals 1 Estimate the population mean and figure the standard deviation of the distribution of means 2 Find the Z scores that go with the confidence interval you want 3 To find the confidence interval change these Z scores to raw scores Chapter 7 102215 1236 PM Effect Size The point is that knowing statistical significance does not give you much information about the size of the effect Significance tells us that the results should convince us that there is an effect that it is not due to chance Significance does not tell us how big this nonchance effect is Effect Size Standardized measure of difference lack of overlap between the populations Effect size increases the greater differences between means Effect size indicates the extent to which two populations do not overlap The amount that two populations do not overlap is called the effect size because it is the extent to which the experiment procedure has an effect of separating the two populations Figuring Effect Size Raw score effect size because the effect size is given in term of the raw score on the measure Standardized Effect Size That is to divide the raw score effect size for each study by each study s population standard deviation Here is the rule for calculating standardized effect size Divide the predicted difference between the population means by the population standard deviation Population 1 M is the mean for the population that receives the experimental manipulation Population 2 M is the mean of the known population the basis for the comparison distribution Population SD is the standard deviation of the population of individuals Effect Size Conventions Effect Size Conventions Standard rules about what to consider a small medium and large effect size based on what is typical in behavioral and social science research also known as Cohen s conventions Summary of Cohen s Effect Size Conventions for Mean Differences Verbal Description Effect Size Small 020 Medium 050 Large 080 A More General Importance of Effect Size Knowing the effect size of a study lets you compare results with effect size found in other studies even when the other studies have different population standard deviations Knowing the effect size lets you compare studies using different measures even if those measure have different means and variance 5 Metaanalysis Statistical method for combining effect sizes from different studies Statistical Power Statistical Power Probability that the study will give a significant result if the research hypothesis is true The power of a study is the probability that it will produce a statistically significant result if the research hypothesis is true If the research hypothesis is false you do not want to get significant results When a study has only a small chance of being significant even if the research hypothesis is true we say the study has low power Determining Statistical Power Power Table Table for a hypothesis testing procedure showing the statistical power of a study for various effect sizes and sample sizes What Determines the Power of a Study The statistical power of a study depends on two main factors 1 how big an effect the effect size 2 how many participants are in the study Power is also affected by the significance level chose whether a onetailed or twotailed test is used and the kind of hypothesis testing procedure used Effect Size The difference in the means between populations we saw earlier I part of what goes into effect size Thus the bigger the effect size is the greater the power is Effect size however is also affected by the population standard deviation The smaller the standard deviation is the bigger the effect size is Sample Size The more people there are in a study the more power there is Sample size affects power because the larger the sample size is the smaller the standard deviation of the distribution of means becomes If these distributions have a smaller standard deviation they are narrower The distributions of means can be narrow and thus have less overlap and more power for two very different reasons One reason is that the populations of individuals may have small standard deviations This reason has to do with effect size The other reasons is that the sample size is large This reason is completely separate Sample size has nothing to do with effect size Both effect size and sample size influence power Figuring Needed Sample Size for a Given Level of Power The main reason researchers consider power is to help decide how many people to include in the study A researcher wants to be sure to have enough people in the study for the study to have fairly high power Other Influences on Power Three other factors besides effect size and sample size affect power 1 Significance Level Less extreme significance levels such as p lt10 or plt20 mean more power More extreme significance level 001 or 0001 mean less power Less extreme significance levels result in more power because the shaded rejection area on the lower curve is bigger Thus more of the area in the upper curve is shaded More extreme significance levels result in less power because the shaded rejection region on the lower curve is smaller 2 OneVersus TwoTailed Tests Using a twotailed test makes it harder to get significance on any one tail Thus keeping everything else the same power is less with a twotailed test than with a onetailed test 3 Type of HypothesisTesting Procedure Sometimes the researcher has a choice of more than on e hypothesistesting procedure to use for a particular study The Role of Power When Planning a Study If you do a study in which the power is low even if the research hypothesis is true the study will probably not give statistically significant results So when the power of a planed study is found to be low researchers look for practical ways to increase the power to an acceptable level A widely used rule is that a study should have 80 power to be worth doing Power of 80 means that there is an 80 chance that the study will produce a statistically significant result if the research hypothesis is true How can you increase the power of a planned study Increase the effect size by increasing the predicted difference between population means Increase effect size by decreasing the population standard deviation Increase the sample size Use a less extreme level of significance such as plt10 or plt20 Use a onetailed test ChU39Iwai L Use a more sensitive hypothesistesting procedure The Role of Power When Interpreting the Results of a Study Role of Power When a Result is Statistically Significant Statistical Significance versus Practical Significance Clinically significant means that the result is big enough to make a difference that matters in treating people If the result is statistically significant you can consider there to be a real effect The next question then is whether the effect size sis large enough for the result ot be useful or interesting This second question is especially important if the study has practical implications If the sample was small you can assume that a statistically significant result is probably also practically significant Significant It means that you be pretty confident that there is some real effect But it does not tell you much about whether that real effect is significant in a practical sense that it is important or noteworthy Role of Power When a Result is Not Statistically Significant Not getting a significant result may have come about because the research hypothesis was false or because the study had too little power Suppose you carried out a study that had high power and you did not get a significant result In this situation it seems unlikely that the research hypothesis is true In this situation where there is high power a nonsignificant result is a fairly strong argument against the research hypothesis Nonsignificant results from a study with low power is truly inconclusive A nonsignificant result from a study with high power does suggest either that the research hypothesis is false or that there is less of an effect than was predicted when figuring power Result Statistically Sample Size Conclusion Significant Yes Small Important result Yes Large Might or might not have practical importance No Small Inconclusive No Large Research hypothesis probably false
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'