UW PSYCH 209 - Final Study Guide

Unformatted text preview:

Final Study GuideCh. 9: Factorial Designs- Factorial design: Includes two or more IV’s and crosses every level of each IV w/ every level of all other IV’so 2x2=Two IVs, two levels eacho 4x2x3=Three IV’s, four levels of first, two levels of second, three levelsof thirdo # of conditions = multiplication of # of levels of each IV- Main effect: Occurs when an IV has an overall effect on a DV- Interaction occurs when the way in which an IV influences behavior differs, depending on the level of another IVo Two-Way: Among two IV’s, the way that one IV influences a DV depends on the level of the second IVo Three-Way interaction: The interaction of two IV’s depends on the level of a third IV- If there is a main effect, marginal means are different for all IV’s and vice versa- If there is an interaction the difference between the marginal means is different for each IV and the link graph is not parallel- Multiple main effects & no interaction: stronger validity- Main effects w/ interactions: limited external validity- Between-subjects: each participant experience only one condition- Within subjects: each participant experiences every possible condition- Mixed-factorial designs: One or more between-subjects variables and one or more within subject variables- Person x situation design: An experimental design that incorporates at least one subject variable & one manipulated variableCh. 10: Experimentation and Validity- Categories of inferenceso Inferences about conflictso Statistical inferenceso Causal inferenceso Inferences about generalizability- Validity of measuremento Fact validity: Degree to which items on a measure appear to be reasonableo Content validity: Degree to which items on a measure represent the range of tiems that could have been appropriately includedo Criterion validity: The validity of a measure to predict an outcomeo Construct validity: Measure truly assess the construct that it is claimed to asses- Validity of experimentso Types of validityFinal Study Guide Construct validity: constructs that researchers claim to be studying are the constraints they are manipulating and measuring Statically conclusion validity: Concerns the proper statistical treatment of data & the soundness of statistical conclusions Internal validity: Does X definitely cause Y- Establishes ability to make causal inferences- Requires main effect & no alternative explanations External validity: Generalizability- Ecological validity - Mundane realism- Psychological realm: stimulate natural behavior Replication- Threats to internal validityo History: Unrelated events that happen to occur during a studyo Maturation: Natural change over timeo Testing a previous measure affects response on subsequent measureso Instrumentation: Changes in measurement deviceo Regression to the mean: Extreme scores do not remain extremeo Attrition: Participant drop outo Selection: Non-equivalent groups at the beginning of the study bias results- Ceiling effect: Discuss when scores on the DV bunch up at the maximum score level- Floor effect: Occurs when scores in a DV bunch up at the minimum score level- Control groupso Yoked control group: Each control member is procedurally linked to a particular experimental group member whose behavior will determine how long both of them are treatedo Untreated control groupo Placebo control groupo Waiting list control groupCh. 11 Quasi Experiment- Quasi Experiments vs. True Experimentso Participants exposed to an IV and a DB is measure but… Researched may not have control over content or design of treatment when it occurs or whom it affects Researcher may not have control over choice of DV or how, when or for whom they are measuredo No random assignment to groupso Allows for experimental comparisons when true experiments are not possible- Designs w/o a control groupFinal Study Guideo One-group post-test only design treatment occurs + DV is measured afterwardo One-group pretest-posttest designs: DV measured before + after treatment Shows change in DVo Simple interrupted time series design DV repeatedly measured before& after a treatment Establishes pre and post treatment trendso Nonequivalent control group designs Posttest only w/ nonequivalent control group participants in one group exposed to DV participants in nonequivalent group not exposedo Pretest posttest w/ nonequivalent control group: Pre-treatment & post treatment scores are obtained for both groups Info gathered on group similarities prior to treatmento Simple interrupted time series w/ nonequivalent control group Series of pre and post treatment scores collects for both groups- Switching replication designso One group receives a treatment while a nonequivalent group doesn’t  Control group receives treatment later DV measuredo Group A receives treatment while group B does not, then group B receives treatment & group A does not DV measured before, during and after treatments- Program evaluationso Process evaluation: Determines whether a program is being implemented as intendedo Outcome evaluation assess a programs effectivenesso Efficiency assessment: Weighs the program’s benefits & effectiveness in relation to its com’s to determine whether it is on efficient method for addressing the programo Program diffusion Implementing & maintaining effective programs in their settings or with other groups- Block randomization: Conduct single round of conditions then another for as many as needed & randomly order conditionsCh. 12 Single Subject Experiments- Difference between case studies & single – subject experiments:o IV is manipulatedo Participants serve as their own controls- Advantageso Allows for systematic study of rarely occurring phenomenao Flexible o Individual differences are examined, not lost in an averageFinal Study Guide- Power vs. Clinical significance- Baseline: measurements of DV before treatment- AB (Comparison design)- ABAB design: sequence of phases that alternate control & treatment conditionso First “A” phase establishes baselineo “B” phases present a treatmento Subsequent “A” phase remorse treatmento DV measured during each phase- Multiple baseline designs: Baseline DV is measured for multiple participants, behaviors, or settings- Treatment introduced to one person/behavior/setting at a time, others at control level (eventually all receive treatment)o Across individuals: two or more participants exposed to the same treatment

View Full Document

Pages: 4