Identify and describe two threats to internal validity in this study. That is, show why there are at least two plausible alternative hypotheses for the obtained change in grades other than due to the academic-improvement program.

Question

At a large university, a group of education specialists tested the effectiveness of a new academic improvement course. Students seeking help at the university counseling center because of academic difficulties were asked to participate in this 4-week program. Only students who were judged to be deficient in reading comprehension and other study-related skills were chosen for the program. Students who sought help at the counseling center for emotional difficulties were included in the study as a comparison group. These students received the usual treatment offered at the counseling center. A group of 30 students completed the academic-improvement program at the counseling center. Average test grades from the students' courses for the two groups were compared before (midterm exams) and after (final exams) the program. Analyses revealed that a statistically significant majority of the students were doing better in school after completing the academic-improvement program than before.

Answer 1

Selection, additive effects with selection, and differential regression are threats to the internal validity of this study. Selection is a threat because differences between the two groups at the outset of the study may account for the differences in final exam test scores. Additive effects with selection are possible if the two groups experienced different events (history, e.g., if the academic program students also sought help from instructors), changed over time at different rates (maturation, e.g., the two groups adapted to the semester at different rates), or the two groups differed in how sensitively the final exams assessed their academic progress (additive effect of selection and instrumentation). Differential regression may be a threat because students selected for the academic-improvement program were judged to be particularly deficient in reading comprehension and study skills. If students were selected because of extreme scores, their test performance may have improved simply because of regression to the mean.