You Are Conducting an Experiment with Group Membership as Your

Not Answered

You are conducting an experiment with group membership as your IV and attitudes toward Barack Obama as your DV, and you hypothesize that people in Group 1 will rate Obama higher than people in Group 2. In terms of the inferential statistical tests you would use to test your hypothesis, briefly explain the four possible outcomes that you could have regarding the correspondence between treatment effect and reality. In other words, use the variables above to describe the outcomes in each cell of the 2x2 matrix of treatment (effect/no effect) by reality (effect/no effect). (Be sure to label error types when appropriate).

a. You find that Group 1 rates Obama higher than Group 2, when in fact there was that effect present. This would be a correct conclusion, and the probability in which we will make this conclusion (given the presence of treatment and "real" effects) reflects the statistical power of the study.
b. 2. You find that Group 1 rates Obama higher than Group 2, but in fact there was no difference between groups in reality. This represents Type 1 error.
c. 3. You find that there was no difference in Obama's ratings between Groups 1 and 2 when in fact there was a difference present. This represents Type 2 error.
d. You find that there was no difference in Obama's ratings between Groups 1 and 2 when in fact there was no difference between groups. This would be a correct conclusion.