so i'm writing a research proposal to study academic achievement between retained students and low-achieving promoted students. academic achievement will be measured by EOG scores (which is on a scale of 1-4 with 3 being at grade level, 4 above grade level, and 1 & 2 below grade level). would a 2 X 3 between subjects factorial ANOVA be the appropriate way to analyze the results? i thought 2 X 3 because 2 (status = retained or promoted) and 3 (EOG subject area - reading, writing, and math). i would assume I would want to do ANOVA, but i'm not so certain about the 2 X 3 design seeing as the dependent variable would be whether the student was retained or promoted...or maybe i'm even wrong about using ANOVA for my statistics. any help would be appreciated!

4/7/2009 3:02:59 PM

so i'm writing a research proposal to study academic achievement between retained students and low-achieving promoted students. academic achievement will be measured by EOG scores (which is on a scale of 1-4 with 3 being at grade level, 4 above grade level, and 1 & 2 below grade level).

would a 2 X 3 between subjects factorial ANOVA be the appropriate way to analyze the results?

i thought 2 X 3 because 2 (status = retained or promoted) and 3 (EOG subject area - reading, writing, and math). i would assume I would want to do ANOVA, but i'm not so certain about the 2 X 3 design seeing as the dependent variable would be whether the student was retained or promoted...or maybe i'm even wrong about using ANOVA for my statistics.

any help would be appreciated!

4/7/2009 3:46:16 PM

yeah so i should've spaced stuff out

4/7/2009 5:22:08 PM

wouldn't a regression be more meaningful?

or ANCOVA using their GPA as a covariate?

4/7/2009 9:11:57 PM

actually, a discriminant analysis might be your best bet.

your DV is categorical, correct? Essentially you are trying to predict group membership.

[Edited on April 8, 2009 at 5:55 PM. Reason : stats R hard]

4/8/2009 5:54:54 PM

or logistic regression, all depending how you treat your IV's.

4/8/2009 5:59:31 PM

i really have no clue. stats isn't my thing. my supervisor mentioned a t-test and then pearson's R. at least i know how to do that. i'm not sure about the other stuff you guys mentioned.

4/10/2009 5:01:54 PM

If this is a part of a graduate thesis or dissertation, you can lookup the NCSU Statistics department and request a consultation. Seriously, this is part of their Masters/PhD's programs to assist with stuff like this as it mimics pretty closely what will be asked in the real world. "Hi, I have some data I'd like to analyze. How do I do it?".

This may help out a bit:
http://www.ats.ucla.edu/stat/mult_pkg/whatstat/choosestat.html

4/19/2009 8:41:44 AM