Psych 325 1st Edition Lecture 25 -Choosing an Experimental DesignoUse between subjects if…-IV has a permanent effectEX: Festinger $1/$20 cognitive dissonance study (the psychological tension linked with the $1 condition would carry over to the $20 condition-"Wait, I just lied about liking this boring task for $1. I better say that I liked it for $20EX: clinical intervention - each participant gets the treatment at some point of the experimentArousal (BETWEEN: Low OR moderately OR high) x Type of mood induction (WITHIN: each participant gets positive and negative mood)-Treatments will confound studyoUse within subjects if…-Studying changes over time-Expect lots of individual differencesoCan also use a mixed design-At least one IV is within-At least one IV is between-Placebo effecto(i.e. finding ideal dosages for a new pain medication)oSay that you give all back pain patients a pill so that each patient is randomly assigned toone dosage conditions-A placebo group is a special type of control group that assesses whether the "psychological" benefit of the experiment treatment explains differences in DV scores-In the example, if the "0" milligram group had an average pain rating of 2 then we would suspect a placebo effect. That is, just taking a pill that participants will reduce their pain causes them to report lower pain levels (equal to the ratings of the "300" milligram group). Thus, we have to seriously question the efficacy of our experimental drugoStatistical power-Definition: the level of confidence that a significant effect is really significant-Three ways to increase statistical power in research designsAdd more participants (more participants per cell decreases person confounds)Increase effect sizes (i.e. increase differences in DV scores across conditions of IV)Lower the criterion (i.e. have a stronger standard for statistical significance)These 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.-Say the one-way two group between-subject effect for level of arousal on number of condom references is not quite statistically significant (p<.08; adult mean = 5; fishing mean = 15)-If we added more participants to each of the two conditions and made the adult video more provocative we could increase our odds that the two experimental conditions are producing different outcomes (i.e. reduce criterion p below .05)-By reducing our standard for statistical significance from a p < .05 to a p < .01 wecould further increase our odds that the differences between the two group means are
View Full Document