Assignments for Module Three

In this part, you need to look at the following paper:

Sherry Vidal (1997) Regression is a Univariate General Linear Model Subsuming Other Parametric Methods as Special Cases.  Paper presented at the annual meeting of the Southwest Educational Research Association, Austin, January, 1997.  [Available here or http://ericae.net/ft/tamu/regress.htm The paper presents a number of analyses conducted on a simple data set (found here).

1. As an exercise, try to replicate the analyses described using those data.   In each case, there are some fairly clear analytical steps carried out.  Some of them, particularly the plots, are a bit complicated depending on what version of SPSS you have; just do as much as you conveniently can.  The main purpose is simply to walk through the analyses discussed in the case for yourself, and see how the numbers were generated. Copy into your report the results of your analyses, and add a paragraph or two of discussion on the experience and what, if anything, it clarified for you (NOTE: "Nothing" is a perfectly permissible answer, as long as you add some explanation of why).
   PART  TWO
   Here you are invited to try out some variations on the General Linear Model.  Use the LA School data set that you first encountered in Module 2; here is a somewhat dressed-up version of it.  The codebook (attached)
2. First, use CTBS Language NCE Score (interval) as DV, and gender (categorical; coded 0=female, 1=male) as IV; try some predictive models:
   ANALYZE : COMPARE MEANS : INDEPENDENT SAMPLES T-TEST (remember to specify the two categories of gender )
   ANALYZE : COMPARE MEANS : ONE-WAY ANOVA
   REGRESSION : LINEAR
3. Then compare the output from these three analyses.  Note which numbers are the same, and what they represent.  Interpret the output from each test specifically, and identify what each flavor of analysis tells you and what it doesn't.
4. Next , use CTBS Language NCE Score (interval) as DV, and gender (categorical; coded 0=female, 1=male) and engprof2 (categorical, recoded from engprof as 0=no non-English proficiency [ engprof =0 or 4], 1=some non-English proficiency [engprof=1, 2, 3], as IV; try as predictive models:
   GENERAL LINEAR MODEL : UNIVARIATE [ gender and engprof2 are both fixed factors]
   [Multiple] REGRESSION : LINEAR [include interac [interaction (product) of gender and engprof2 ]  along with gender and engprof2 as IVs]
5. Again, compare the output, and interpret the results.  In this case, all three IV's are what are called "dummy variables", 0/1 codings reflecting the presence or absence of a particular effect.    In each case, the variable is coded as 1 if the characteristic is true, and 0 if it is not.
6. Then, try at least one more analysis along these lines of your own choosing.  Please say why it's interesting and you chose it, how you set it up, what you found, and your interpretation of the results.

Price: $19.58

Solution: The downloadable solution consists of 11 pages, 858 words and 25 charts.
Deliverable: Word Document