Week 9 Lecture Notes PDF

Preview

Summary

lecture notes from Ann Culligan's Class Fall 2019, three lectures on a typical week monday wednesday friday ...

Description

Kaitlyn Paskill Psych 209 Notes Nov. 2019, Week 9 Week 9: In-Class Notes Monday 11/18/19 Lecture -

If there is no crossing in the lines on the graph in a factorial design there is likely no interaction

Main Effects and Interactions - Interactions qualify the outcomes of the main effects - Both main effects and interactions must be carefully considered before drawing conclusions about the relationships between variables - Ex: age ONLY affects perceptions of managerial competence when gender is taken into account - These qualifications determine the external validity of results - Multiple main effects and no interactions: stronger validity - Main effects with interactions: limited external validity (can't really generalize) - Variables in a factorial design - Variables with only two levels reveal only a linear relationship - Ex. 2x2 design - Variables with three or more levels reveal possible non-linear effects - Ex. 2x5 design - All independent variables may be manipulated (or they might not be) - Subject variables may be included as independent variables - Non manipulated, selected - At least one other IV MUST be manipulated

Tuesday 11/19/19 Section - Need to know 2x2, 4x4, and 2x3x4 factorial design concepts well for the final there will be a large chunk of points related to this on the final - 2x2 - Two independent variables and two levels of each of them - Example: Effect of Music Volume on Happiness levels While Studying or Leisure - Iv’s: activities of study and leisure - Levels: music volumes of loud and soft - Main effect - Row versus column means (means=average)

Kaitlyn Paskill Psych 209 Notes Nov. 2019, Week 9

-

- Need to know whether there is a rate of change As long as the main effect is not zero, for the purposes of this class, state that there is a main effect, even if it is not statistically significant You can have a main effect for one IV and not the other, or both, or neither Can tell that there is an interaction if the lines are crossing on the graph Describing factorial designs - There is a main effect of x on y. On average, participants who x reported greater y than when…. - There is an interaction between x on y….

Wednesday 11/20/19 Lecture Analyzing statistical relationship between variables - ANOVA is used to compare the means of call conditions - Reveals main effects and interactions - Determines statistical significance - Simple main effects compared to the effects of an IV at just one level of another IV - Ex. is there a difference in aggression between low, medium, and high frustration levels only at 9 am? - Simple contrasts are used to compare the effects of one IV on individual levels of the other IV Validity - The goal of experimental research is to be able to make valid claims about a measured phenomenon - Valid claims rely on the validity of each step of the research process - Construct validity - Are we measuring what we intended to measure? - Are our operational definitions true to the construct? - How do we know whether our measures of the IV and DV are valid? - Face validity - Does it appear to accurately measure the construct? - The test appears to measure what it intends to measure - Content validity - Does the test include all components of the construct? - Did we get all the topics? - Statistical validity - Internal validity

Kaitlyn Paskill Psych 209 Notes Nov. 2019, Week 9 -

External validity

Friday 11/22/19 Lecture Validity - Construct Validity - (How do we know whether our measures of the IV and DV are valid?) (FOR YOUR STUDY USE THE SAME METHODS THAT HAVE BEEN USED BY OTHER SCIENTISTS) - Face validity - Does it appear to accurately measure the construct? - Content validity - Does the test include all components of the construct? - Predictive validity - Does the test accurately predict future outcomes? - Convergent validity - Do test results correlate with other measures of the same construct? - Discriminant validity - Are the results of this test dissimilar to those of tests measuring unrelated constructs? - Statistical Validity - Did the researchers perform appropriate statistical analysis for the type of data involved? - Did the researchers interpret the data correctly? - Did they use the right test? - What did we measure and what did these measures really mean? - Internal Validity - Does X definitively cause Y? - Internal validity establishes the ability to make causal inferences - Requires - A main effect of one or more IV’s - Absence of other plausible alternative explanations - With data and evidence - Not possible to always eliminate absence of other plausible alternative explanations - Threats to internal validity - History - unrelated events that happen to occur during the study, the events of the world

Kaitlyn Paskill Psych 209 Notes Nov. 2019, Week 9 -

-

-

-

-

-

-

Maturation - natural changes over time, history unto the subject, included in any longitudinal study Testing - a previous measure affects response on subsequent measures Instrumentation - changes in the measurement device - (forgetting to recalibrate your scale for example) Regression to the mean - extreme scores to not remain extreme, when we are not doing repeated measures - (someone comes in extra sad or anxious for whatever reason however we typically regress towards the mean normally) - Why are you so extra? Attrition - participants drop out - Differential attrition - group that drops out is meaningfully different from the group that stayed in Selection - non-equivalent groups at the beginning of the study bias results Demand characteristics - cues that influence participants beliefs and behavior - Subjects think they know what the study is about and they believe they know how they should behave and they behave in that way rather than in reaction to the independent variable Experimenter expectancy - researchers expectations may unintentionally influence participants Placebo effects - participants expectations influence their responses - Treatment or expectation? - Our beliefs shape our reality - Double blind studies can correct for this Multiple conditions manipulated along with the IV, may influence the DV

Week 9 Review Questions When a test measures what it is intended to measure and does not correlate with measures of other constructs the test has - Predictive validity - Statistical validity - Face validity - Discriminant validity If there is NO difference in the marginal means for each level of an independent variable that means that - There is no interaction effects between IV’s - There is a main effect of that IV

Kaitlyn Paskill Psych 209 Notes Nov. 2019, Week 9

-

-

-

-

-

- There is an interaction effect between the IV’s - There is no main effect of that IV Internal validity tells us that - The test accurately measures what it is meant to measure (construct validity) - X definitively causes y - The appropriate statistical test was used to analyze the data - The results can be generalized to the population How to calculate main effect How to calculate interaction - Differences in differences In a mixed factorial design - All participants are exposed to every level of the IV - All participants are exposed to every level of one or more IV AND each participant is exposed to (or represents) just one level of one or more other IV’s (some measured as between and some measured as within subjects) - Both qualitative and quantitative data are collected - Each participant is exposed to just one level of the IV If there is NO difference in the marginal means for each level of an IV that means that - There is no interaction effect between IV’s - There is a main effect of that IV - There is an interaction effect between IV’s - There is no main effect of that IV Study, testing the effects of Ketamine and its delivery method on depression, all the mice in the study were given Ketamine and tested for symptoms of depression. Half of mice injected by male researcher and other half by female. In study, treatement with Ketamine was a ____ variable and sex of researcher doing the injection was a _____ variable - Within subjects(everyone gets K); between subjects(not everyone exposed) - Experimental; subject - Between-subjects; within-subjects - dependent; independent If a study has three or more conditions, _____ will be used to compare the mean scores of each - T Test - Pearson's R - Wilcoxon signed-rank test - ANOVA...