Psychology 202b

Advanced Psychological Statistics II

Seventh homework assignment, 5/3/2011 (due 5/13/2011).

Here are links to level one and level two files for a simulated growth curve model. In the level one file, we have 150 individuals, each measured at several time points. The columns are ID, test score, time, and time squared. In the level two file, there is an ID column followed by a simulated family income variable.

Your task is to use the student version of HLM software to estimate the growth model. Here are some suggestions about how to proceed:

Create an mdm file. (If you have trouble with fortran formats for the data input, don't hesitate to ask for help, but do take a stab at it yourself.)

Focus on the level-one model, and determine whether growth is flat, linear, or quadratic.

Consider a model in which all growth components are random effects. For example, if you decided that a quadratic model was appropriate, make the intercept, the linear slope, and the quadratic slope all random effects. Look at the inferential results for the variance components, and decide whether there is evidence in favor of the idea that there is significant between-person variation in these parameters.

Examine the impact of income on each of your growth parameters, using a model in which only the intercept is considered a random effect. Discuss your findings.

Consider income as a predictor of all growth parameters, this time treating all of the growth parameters as random effects. What happens when you try to do this?

Finally, take a stand on the most appropriate model for this analysis, and summarize the findings.