I've had a few questions, so I wanted to make sure that everyone was aware of the some of the following:

  1.  The first half of portfolio 2 has to be a SEM. It is not uncommon for students to use the same lit review, and to run the same MLM analysis from Portfolio 1 as a SEM. You're welcome to innovate (e.g., try out a higher order growth model or a growth pattern mixture model), but it's fine to redo your portfolio one analyses as a SEM
  2.  The second half of the portfolio has to be a survival analysis. If you can find or create a variable in your data set, it's fine to do that. For example, one student is analyzing 10 sessions of test scores. For the survival analysis they will define a criterion score (e.g., a score of 10), and the survival analysis will look at the hazard/survival of achieving a score of 10 (At each occasion 0 = not achieving a 10; 1 = have achieved a 10, and are excluded from subsequent occasions). The critical thing with doing that is that you have to pick a criterion THAT NO ONE HAS MET AT BASELINE (or you'll be forced to exclude participants who have met it), but that at least some people (IDEALLY 50%, but beggars can't be choosers) have met before the study is over.

THAT SAID:  Sometimes, you just can't make the same data set work for part1 and part2. It's totally fine if the two halves are utterly disconnected. The goal of this paper is demonstrate technical competence in data that are at least somewhat meaningful to you.



Michael Marsiske, Ph.D.
Associate Professor
Department of Clinical and Health Psychology
PO Box 100165,
1225 Center Drive, Rm. 3150
Gainesville, FL  32610-0165
Phone:  (352) 273-5097
Fax: (801) 720-5897
Email: [log in to unmask]

Twitter: @MMarsiske