solution

Martin et al. (2007) conducted an experiment to determine how people’s behavioral intentions can be affected by being told that the behavior is approved by a majority or only a minority of other people (Source Status). In their experiment, they could control the number of people in each Source Status, but not the level of the other variable (Attitude Change), leading to unbalanced data. The data is summarized here for the dependent variable Behavioral Intention.

(a) Construct a profile plot that will allow you to inspect the apparent effects in the data.

(b) The authors provide the following test statistics:

Main effect for Source Status: F(1,49) = 4.98, p

(c) The authors make two statements:

(1) As predicted, when participants did not change their attitude, there was no difference between the majority and minority conditions.

(2) However, when participants did change their attitude, those in the minority condition reported higher behavioral intention. Each of these statements corresponds to a specific contrast. Construct the t statistic for each of these contrasts. If you control the experiment wise significance level for this pair of contrasts at 5%, would your results be consistent with the authors’ statements? Explain how you controlled the experiment-wise significance level. Hint: The MSE for this data was 2.248.