PSY 2311st Edition Lecture 10Outline of Last Lecture I. Designs With One IVII. Multiple IV Designsa. Structure b. Results: Main Effectc. Results: Interaction EffectOutline of Current Lecture II. Checking External ValidityIII. Internal ValidityIV. Two Kinds of Error VarianceV. ConfoundCurrent LectureChecking External Validity● By experimenting :Replicate○ new types of subjects, test situations, time period● Results that hold true across many experiments are likely to hold true in the real world too● By comparing results to real world patterns○ does “happier” chocolate correlate with : sales? what people say?● Different from face validity● Face Validity -- experiment looks like the everyday situation it is meant to tell us about● External validity is about what we learn, not what it looks likeInternal Validity● Internal Validity-- extent to which the IV, and only the IV, could have affected the DV○ “pure primary variance”○ Experiments may also measure error variance -- score differences caused by something other than the IVTwo Kinds of Error Variance● Unsystematic error variance○ affects all groups the same way○ does not affect our impression of what the IV does● Systematic error variance○ affects different groups differently○ can change our impression of what the IV doesConfound● Confound -- threatens the internal validity, factor that causes systematic error variance● Can make : effective IV look ineffective OR ineffective IV look effective● Results can be externally valid ONLY if the study is internally
View Full Document