On 25 Sep 2002 17:18:56 -0700, [EMAIL PROTECTED] (R Cole) wrote: > This is really a two part question, and I could use any advice available > on my questions. > > First - Can anyone recommend a good desk reference for the use of SPSS > (11.0) in conducting and interpreting analysis�of survey-type research? > It has been about three years since I completed my statistics coursework > and I could use a good refresher.
1. My stats-FAQ has answers to some particular questions. And I have links to some people who have (a) excellent bibliographies, and (b) excellent discussions of their own. 2. You can check the homepages of other folks who post answers in the stats-groups. 3. You can do a google search, say, <"survey research" tutorial > or maybe FAQ in place of tutorial. > Second - I am doing a study using a two part instrument where I will use > factor analysis to identify statistically significant factors in the > first area (new work in a survey developed for this study). "Statistically significant factors" is not the way that I regard factors. And I haven't seen that emphasis in these Usenet stats groups. In 'sci.stat.edu' we seem to take the 'edu' as part of educating, but not as the specialty of 'educational research' -- which is an area that (I think) tends to treat factors with a lot more honor than we do. > > The second part will be to determine if there are any significant > correlations between this instrument and the results of a previously > validated instrument that determines the relative strength of > characteristics in four quadrants. > "quadrants" - is that the term from astrology? Sorry. I mean to say, this terminology is failing to communicate. Questions of "...relative strength" might be getting into serious measurement issues.... > The goal is to determine if there are correlations between the > identified factors in the two instruments such that indications of high > values in the four characteristics correlate with any of the factors in > the first part of the instrument. Preliminary analysis indicates there > will be 7 factors from the first instrument to correlate to 4 factors in > the second. > > I have considered ANOVA and Regression Analysis but am not sure which > would be most practical. Well, it sounds like you have some correlations to look at. Regression analysis is an "analysis of variance" in its details; ANOVA by initials usually means you have groups. Do you have groups? Or, you think it would be useful to create groups, as part of the description? > > Any advice would be appreciated. I feel I am missing something obvious, > but am hitting a mental block in my methodology development. > > Thanks, > Richard > [EMAIL PROTECTED] > > > . > . > ================================================================= > Instructions for joining and leaving this list, remarks about the > problem of INAPPROPRIATE MESSAGES, and archives are available at: > . http://jse.stat.ncsu.edu/ . > ================================================================= http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
