The results of the current study were analyzed using SPSS. An alpha level of .05 was used as the criterion for statistical significance. A univariate analysis of variance was used to evaluate the relationship of question type, interviewer type, and their interaction. A T-Test was used to evaluate the relationship between the type of interviewer in the leading condition on the number of correct answers and on the number of shifts. A chi-square was used to evaluate the relationship between question type and the number of induced intrusions.
The independent variables, interviewer type and question type, each had two levels: perceived authority and non-authority, and open-ended and leading. For the purposes of analysis, the dependent variables were correct answers, the number of shifts, and the number of intrusions.
The appropriate statistical tests determined significance in the direction of the initial hypothesis in two areas. First, the influence of perceived authority did create a significant difference in the total number of correct answers on the scale. Second, the presence of leading questions created a statistically significant greater number of intrusions. However, the influence of question type was significant in the opposite direction originally hypothesized. In addition, the interaction between question type and interviewer type was significant in a direction inconsistent with the predicted outcome.
ANOVAS
Question Type. The effect of question type was significantly significant, F(1,68)=3.7, p=.058. Subjects interviewed with leading questions got a higher number of correct answers on the BTSS.
Interviewer Type. The effect of interviewer type was significantly significant, F (1,58)=5.08, p=.028. Subjects tested by a non-authority figure had a higher number of correct responses on the BTSS.
Interaction. The effect of the interaction between interviewer type and question type was also significant, F (1,58)=3.98, p=.050. Subjects in the non-authority, leading condition had the highest number of correct responses. Subjects in the perceived authority, leading condition had the lowest number of correct responses.
T-Tests
Correct. The number of correct responses between the perceived authority and leading question condition was significant, t= -2.59, df= 27, p=0.15. Subjects interviewed in the non-authority condition with leading questions scored a higher number of correct responses than those in the perceived authority condition with leading questions.
Incorrect. The number incorrect between the perceived authority and leading question condition was also significant, t=2.59, df=27, p,=0.15. Those subjects interviewed in the perceived authority, leading condition scored a higher number of incorrect responses on the BTSS than their counterparts.
Shifts. Though there was no hypothesized direction, the number of shifts between the groups did approach significance, t=1.48, df=27, p=.07. The perceived authority, leading condition had a mean number of 1.87 shifts while the non-authority leading condition had a
mean of 1.00 shifts.
Chi Squares
Intrusions. The interaction between the number of intrusions and question type was significant in the predicted direction, x_(1,N=61) = 5.3, p=.02. The leading question conditions created 11 induced intrusions while the open-ended questions created only 4 induced intrusions.