Study Guide Research FINAL
Study Guide Research FINAL PSY 3392
Popular in Research Design and Analysis
Popular in Psychlogy
This 8 page Study Guide was uploaded by Yesenia Notetaker on Friday April 22, 2016. The Study Guide belongs to PSY 3392 at University of Texas at Dallas taught by Meredith Grant in Spring 2016. Since its upload, it has received 29 views. For similar materials see Research Design and Analysis in Psychlogy at University of Texas at Dallas.
Reviews for Study Guide Research FINAL
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: 04/22/16
New terms Complex Designs Main effect – the effect of an individual IV alone Interaction – the effect of an IV differs at different levels of another IV Independent factor – factorial design with independent measures Repeated factor – factorial design with repeated measures Ceiling effect – performance on the DV reaches a maximum Floor effect – performance on the DV reaches a minimum Small Sample Research Applied research – examination of psychological principles and treatments in real world settings Case study – intensive description and analysis of a single case Idiographic approach – study of individual Nomothetic approach – study of groups, averages Testimonial – A case study, not proof. At best, one data point. Single-Case Experiment – apply an experimental design to a single subject or few individuals Stable Baseline – before intervention, helps infer causation for subject Reversal - hoping for a reversal from baseline. ABAB Design (Reversal Design) - baseline, then intervention, then withdraw intervention and behavior should go back to the baseline, then intervene again. Multiple Baseline Designs – compare effect of intervention on multiple baselines Quasi-experimental designs One-Group Pretest-posttest - measure, manipulate, then measure. No control group. Nonequivalent Control Group Design - compare intervention to a "like" control group, but without randomization Interrupted Time-Series Design - relatively stable baseline is needed. Several tests before and after treatment. No control group Time Series with Nonequivalent Control Group - Multiple pre and post tests. Intervention and control group “True” experiment - uses manipulation and control, often in a lab setting Internal validity - extent to which we can infer causality External validity - extent to which it applies to the real world Randomization - distinguishes a true experiment, the only difference between groups is due to chance Self-selection - impact findings. Some people more willing to participate than others Waiting list control - track people on wait list as control subjects Alternate treatments - group 1 treatment a then b. group 2 treatment b then a Threats to internal validity in quasi-experiments - 1. History - event other than treatment produced the change 2. Maturation - people naturally change over time 3. Testing - people get better with familiarity, repeated testing 4. Instrumentation - Mechanical: instruments of measurement are wrong and affect all measures. Observer biases: interviewers growing better with practice or worse with exhaustion 5. Regression to the mean - extreme scores tend to move to the middle over time 6. Subject attrition - participants are lost to the experiment over time 7. Selection - differences exist in characteristics of groups 8. Additive effects with selection - a threat that applies to one group but not another Contamination - groups that communicate Experimenter expectancy - experimenter influences findings in expected directions Novelty effects - the newness of the treatment has an effect instead of the treatment himself Hawthorne effect - behaviors change because someone is interested an paying attention Communicating Research Peer-review process – Peers (experts in the field) decide if the paper will make a contribution, they can reject, suggests revisions, or accept as is. Digital Object Identifier (DOI) Know the structure of the APA report: * Title Page - Remember running head and the first few words of title so it is easy to identify * Abstract - summarize issue, methods, findings, conclusions, and implications. * Introduction - makes a case for the hypotheses and the study, not just a summary of previous research. Brief and presents hypothesis. * Method - describes participants, measures, procedures, and sometimes data analysis. * Results - present and summarize analyses, but don't interpret. All the numerical stuff. * Discussion - summarize results, discuss the meaning/interpretation of results. * Limitations – how results can be generalized and how they can't, discuss alternative explanations. * References - Listing everything that was cited * Footnotes - Rarely used * Tables and Figures - Go in separate pages than captions * Appendices - Rarely used Revisited Terms Intro and the Scientific Method Psychology – the study of behavior and mental processes Science – empiricism and appropriate skepticism Scientific method/approach - a system for acquiring knowledge through observation and experimentation. a process for interrogating the world. facts are a product of science Empiricism – claims based on evidence. evidence derived from direct observation and experimentation “Appropriate” skepticism – a Pseudoscience –beliefs that are not empirical Basic Research – to understand behavior and processes Applied Research – to change lives Operational Definition – explains construct in terms of observable procedures used Reliability – consistency Validity – truthfulness (external validity - truthfulness in real world.) Qualitative research – events and context Quantitative research – translated and analyzed as numerical data Converging evidence – findings should converge to support scientific theory.. after several studies that try to take on same thing from different angles Ethics Institutional Review Board (IRB) - internal review board? can approve, reject, or modify Risk Benefit Ratio – subjective process of weighing the risks and benefits of a research project that is used to determine whether the research should be conducted. (does benefit outweigh risk? is the quality of the data going to provide and interpretation and be meaningful? is there a lower risk procedure available?) Minimal Risk – the harm-discomfort of a study is not greater than would be expected in daily life or during routine physical-psychological tests. Informed Consent – persons explicitly expressed willingness to participate in a research project based on a clear understanding of the nature of the research, of the consequences for not participating, and of all other factors that might be expected to influence the person's willingness to participate. goal: ensure participant competence, knowledge, and volition Assent – verbal agreement Deception – when information about a study is withheld either through omission (withholding information) or commission (intentionally misleading a participant). cannot use deception if experiment is expected to cause physical pain. Debriefing – after participation, experimenter informs about any undisclosed aspects of the study. Publication credit – scholarly contribution, acknowledgement of people that did stuff Plagiarism – not as easy, because some things become part of what you think that you can accidentally plagiarize Observational Designs Descriptive Method – Understanding a phenomena without experimentation. These can establish correlations, but not causations. Direct Observation – Without intervention: naturalistic observation. With intervention: participant observation, structured observation, field experiment. Naturalistic Observation – direct observation of a behavior in a natural setting without attempt of observer to interfere. The goal is to describe behavior as it naturally occurs. The observer is a passive recorder and this type of observation can establish external validity. Participant Observation – Observation that is conducted by a person participating in the situation. Undisguised: it is obvious that the person is doing research. Disguised: Person doing research does not show/say is doing it for research to avoid participants reaction to the observer effect. Self-observation: reality tv. Structured Observation - The setting up of a situation to observe a specific event. Field Experiment – When the experimenter manipulates one or more Independent Variables in a field setting. Indirect Observation – Using real world evidence to test hypotheses without participant involvement, usually being part of the approach, either at the beginning or the end of the research. Physical Traces – Using remnants, fragments, and products of past behavior as evidence. Selective Survival – when some archives/traces survive whereas others do not (e.g. Cutting out images of former romantic partners) Remnant Use traces (natural/controlled) - Use traces: physical evidence that results from use. Natural: results from naturally occurring events (e.g. highlighting a textbook). Controlled: involve manipulation of the experimenter (e.g. potato chips with colors) Remnant products – creation of artifacts/behavior (e.g. portion sizes to examine French Paradox). Archival Records – Public and private documents for individuals, organizations, countries, etc. Can be Running records: continually kept and updated (e.g. college transcript, census data, tax documents, Facebook entries). Episodic records: describe specific events (e.g. birth certificates, marriage licenses, divorce proceedings. Natural treatments: track effects of societal events (e.g. did cotton prices cause great depression, cotton prices, slave trade, other.) Media: news reports, books, movies, advertisements. Meta-analysis – Take data from previously conducted studies and analyze it. Gives power and ability to draw conclusions that could not have been made without all the data. Secondary data analysis – Take a data set that has been available for a period of time and analyze it in a different way Coding – Take quantitative data and divide it into categories. Convert qualitative into quantitative. Inter-observer/inter-rater reliability – different raters agree at 85%. If the coding system is good, trained raters should agree. Survey Research Survey – All respondents complete the same items, verbally or in writing Correlation (magnitude/direction) - Magnitude is strength of correlation. Closer to –1 or 1 is strong, closer to 0 is weak. Sign is direction. If positive, both variables move together, if negative, both variables are opposite or inverses. Population – set of all cases of interest (Representative) Sample – Subset of population drawn from sampling frame. Representativeness: extent that a sample exhibits distribution characteristics of the population. Generalizable – data from sample can be applied to the general population Cross-sectional – one or more samples are drawn from a population at one time. Describe difference in section, not changes from one to another. Longitudinal – Same sample of respondents surveyed more than once. Attrition – slowly dropping out of respondents Getting to know your data “Data cleaning” - Taking the time to look closely, determining whether there is duplicate cases, missing values, impossible values, or outliers. Duplicate cases – Entering the same participant more than once. Outliers – Data points that are distinctly separate from the rest of the data. Reverse scoring – Making the answers go in the same direction to be able to do averages without changing the significance. (Used when sneaky question in opposite directions are used) Stem and Leaf Plot – Preliminary analysis. Histogram – Preliminary analysis. The Bell Curve/Normal Distribution – Symmetrical, average score in the middle of distribution, 68% fall within the middle 2 SDs. 95% Fall within the middle 4 SDs. Skewness (positive, negative) - Positively skewed: majority of data is on the left, mean is on the right because the extreme values will pull. Negatively skewed: majority of data is on the right, mean is on the left because the extreme values will pull. Kurtosis – Degree of Peakedness. Mesokurtosis: normal distribution, middle. Leptokurtosis: Peaked distribution, thin or slender. Platykurtosis: blunt distribution, broad or fat. Categorical variable – Countable, discrete values (nominal, ordinal) Continuous variable – infinite values between two points (ordinal, interval, ratio) Measures of central tendency (mean, median, mode) - Mean: mathematical average. Median: middle value. Mode: most common value. Measures of dispersion (range, variability, standard deviation) - Range: lowest to highest values. Variability: how spread out data is. Standard Deviation: normal change. Confidence interval – uncertainty associated with the parameter. Null Hypothesis Significance Testing Statistically Significant – unlikely to have occurred by chance Null Hypothesis - States there is no impact on the dependent variable caused by the independent variable, there is no difference between groups, etc. If research finds a statistical significance, researcher can REJECT THE NULL HYPOTHESIS. Alternative Hypothesis - States there is an impact on the dependent variable caused by the independent variable, there is difference between groups, etc. If research does not find a statistical significance, researcher FAILS TO REJECT THE NULL HYPOTHESIS, does not PROVE THE ALTERNATIVE. Prove vs. Support – this is probability statistic, it does not help prove. It only says there is a difference, not the effect size. P-value - Probability of obtaining the observed effect if the null hypothesis were true. Usually, p=.05 which means that if we kept taking samples over and over, there is a 5 % chance that our true population mean would NOT fall within the range. (Opposite of confidence interval). In general, the smaller the p- value the greater the change the effect is "real" and will be replicated in another study. Alpha level - the p value threshold that needs to be crossed in order to reach "statistical significance". Usually, alpha level is p<.05, which means there is less than 5% probability that the observed effect would occur in the sample if there was no real difference in the population. Less than 5% that research is wrong. Type I Error - Rejection of Null hypothesis when it is actually true. False positive. Pregnant guy. Can be reduced by having a more strict p-value, Can be caused by alpha level or too many tests. It means finding something that doesn't exist. Type II Error - Failing to reject the Null hypothesis when it is actually false. False negative. Pregnant woman. Usually caused by a small sample size. It means not finding something we should have found. Effect size - An estimate of the size of an effect that is mostly independent of the sample size. Effect size looks at difference between groups. Two common types of statistical analyses: Cohen's d and Pearson's r Power - the probability that the null hypothesis will be correctly rejected when it is false. The ability to detect statistically significant effects Experimental Designs Independent Variable - to be manipulated, 2 or more levels, usually treatment and control Dependent Variable - to be analyzed, see if IV had on effect on this. Internal Validity - he degree to which differences in the DV can be attributed to the IV versus another variable. Helps infer causality. It means that confounds have little to do with the outcome of the experiment, the effect is very valid in a lab setting. This does, though, reduce external validity Covariation - the DV values differs at different levels of the IV, the experimental and control score differently. The DV changes with the IV, either positively or negatively (correlation). Time-Order Relationship - Temporal precedence, one thing comes before another. The manipulation causes the DV to change, the DV does not change before anything happened. IV before DV always. Confounding variables/extraneous variables - other variables that may cause the effect. Balanced groups/samples - if one group substantially differs from the other, it could be something from the group itself Random assignment - Randomly assign participants to a condition. Participant Reactivity - behavior changes because participants believe it's what they have to do Experimenter effects/demand characteristics - experimenter adjusting the behavior because of condition he is dealing with. Can be counteracted by double blind procedures Independent groups (Between subjects) - Examine between group differences for 2 or more groups. Using different groups for the experimental/treatment group and for the control groups. Separating group of participants for each level of an IV. Has different types of "groups". Independent samples t-test (for independent groups) - Use t-test when IV has two levels, manipulation and control Factorial ANOVA (f-test) - Use factorial design when IV has more than 2 levels, one control and two or more degrees of IV levels Experimental Designs Practice effects - Participants getting better or worse at a task over time, disadvantage of repeated measures designs Paired Samples t-test (within subjects) - 2 groups/conditions. Like independent samples t-test. Repeated Measures ANOVA (f-test) - More than 2 groups/conditions, one IV with two or more levels. Anovas can only interpret a "simple main effect", can't have an interaction
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'