Psych 325 1st Edition Lecture 18Thursday, April 9, 2015Chapter 8Part I: Introduction and basic designs-What is a Quasi-Experiment?oResembles a true experimentoImportant differences-Partial (or not) control over independent variables-No random assignment to conditionsoInternal validity problems-Person confounds-Problems can be minimizedoUseful if you recognize limitsoExamples:-Natural disasters…9-11 exposure and PTSD symptoms-Clinical interventions…12-week cognitive behavioral therapy and anxiety levels-Educational interventions…time management workshop and grades-Social policies---change in state speed limits and number of speeding tickets-Trying to establish a cause-effect relationship between childhood trauma (i.e. abuse) and self-esteem in a field setting (resembles a true experiment)oYou would find individuals from similar neighborhoodsoEach had a different experience as a child-we ae unable to manipulate this, for obvious reasonsoThese are the various conditions of your independent variable, which however are already establishedoYou would measure your dependent variable (self-esteem) across the four groups-Why do Quasi-experiments?oSome questions can't be studied with experiments-Can't manipulate some variables-Questions are still interesting!oUsing best of experimental and non-experimental techniques-Best internal validity possible-Also externally valid-Finding equivalent comparison groupsoLocate similar group-Not exposed to treatment-Useful comparison-ProblemsConfound--group may differDifficult to findThese notes represent a detailed interpretation of the professor’s lecture. GradeBuddy is best used as a supplement to your own notes, not as a substitute.-Locate similar peopleMatching individualsBetter, more valid comparisonProblems-Identifying variables to match-May introduce a confound-Tough to dooPages 255-6-Problem: some manipulation provides some value (DV)How do we interpret the meaning of that value?Ex: suppose we do group-based therapy with a group of highly anxious patients and find an average post-test anxiety score of 2 (out of a possible 10)-Can we conclude that the therapy did not work? NO!!!-Solution: find an equivalent control group(s)-Matching: try to find other group(s) that have similar qualities as your therapy group BEFORE the therapy (well adjusted healthy people WILL NOT be a good control)i.e. locate similar group: find a group of people with a mental disorder related to anxiety (i.e. panic disorder) who are also going through group-based therapyEx2: find people with similar traits (i.e., not extroverted) or life circumstances (i.e. similar income) as the therapy group, except not high on anxiety-Ultimate goalTo make each group or condition as equal as possible prior to the manipulation-Basic quasi-experimental designsoOne group pretest-posttest -Comparing with self-History, maturation confoundsoPosttest-only with non-equivalent groups-Compare to other groups-Confound of group differencesoPretest-posttest with non-equivalent groups-Uses all of the above-Helps rule out above confounds-Effect of natural disaster on stressoPPNEG: includes all of the elements in the design--minimizes history, maturation-However, confounds are still a problem because of:Lack of researcher controlLack of random assignment of participants to conditions-Adding patches (i.e. new control group, add pre-test)Provides more confirmation of the independent variable effectDecreases the role of