Psychology 613
Data Analysis III
Prof. Bertram Malle
Spring 2007

Assignment 5

  1. Find one article in your field of study in which the researcher analyzed interaction terms.  Critique (or praise) the manner in which the analyses were conducted and interpreted and comment on whether the interaction term was the best possible way of testing the researcher’s question/hypothesis. (One page maximum.)

  2. Why can’t we save ourselves a lot of time and trouble and simply count whether each of a set if studies did or did not confirm a given hypothesis (“vote counting”) and thereby summarize the accumulated results of a research literature? (Half page maximum.)

  3. Use a random number generator (e.g., the one in the EXCEL Analysis Tools) to run through the following steps:

    a. Generate twenty samples of “experimental groups” of N = 15, drawn from a population with μ = 6, σ = 8. Also generate twenty samples of “control groups” of N = 15, drawn from a population with μ = 3, σ = 7.

    b. Combine the samples pairwise so you have twenty "studies" of N = 30 data points.  For each of these twenty studies, compute (following Lecture 9's handout) the mean difference between the two groups, the pooled standard deviation, and a bias-corrected effect size value d.  Finally, compute an average effect size across all twenty studies (don't worry about inverse weights or the like.)

    c. Plot and describe the resulting distribution of twenty effect sizes and their average. (A dotplot would be suitable.) What does the distribution illustrate? 

    d. Now repeat the same exercise with forty samples of experimental and control groups that have an N of 8 each.  Calculate the forty individual study effect sizes and the overall average of the forty studies. (A separate plot is recommended). Comment on the similarities and differences between this effect size distribution (based on forty studies) and the one from 3b (based on twenty studies).

  4. Please include the excel file (or whatever other program you used) when you turn in your homework. This makes it easiest for you to show your work and for the TAs to grade it.