[K12OSN] Request for suggestions regarding comprehensive statistics on a large survey

[Hoover Chan]

Wow, I hope you had a chance to design the study and examine assumptions and sampling first before collecting data. That makes the statistical analysis easier.

That being said, I have to admit that I too have been in situations where the data were collected first before thinking about the underlying statistical models. 

Without knowing anything about your data, I'd imagine that an analysis of variance (ANOVA) maybe couple with factor analysis may be the best approach. You can get the tools to do these kinds of analyses with packages like SAS, SPSS or Statgraphics, among the ones that I've used in the past.

Instead of a spreadsheet, you may want to consider storing your data into some kind of DBMS. Filemaker is a nice one and MS Access can suffice too. Storing the data in this form can make the retrieval process easier when it comes time to feeding these data into a statistics package.

I hope this helps some.



--------------------------------------------------
Hoover Chan                   chan at sacredsf.org
Director of Technology
Schools of the Sacred Heart
2222 Broadway St.
San Francisco, CA 94115

----- "Daniel Bo" <daengbo at gmail.com> wrote:
> I'm a long-time member  of this list (K12OS 1.0) and know that there
> are a ton of great people on it who offer great advice so my first
> thought was to come here for the help I need.
> 
> OK, enough boot licking. Let's get down to business.
> 
> I have given a survey to 1400 elementary school students to study a
> certain area of linguistics (specifically, comprehensible input
> theory
> and the role of attitude and aptitude in L2 success). The amount of
> data I have is overwhelming. I now need to do statistical analysis to
> find out which of the 18 elements are correlated. I have no idea how
> to do this other than to enter the data into Calc and do the
> calculations manually. There has to be a better way.
> 
> People do this kind of thing all the time. How do they do it?
> 
> General suggestions are, of course, welcome, but I'd really like to
> know how to isolate certain parts of the survey (being discrete) and
> see if there are correlations among e.g. just grade level or section
> number.
> 
> I haven't touched stats or matrix algebra in over twenty years, so go
> easy on the heavy math talk  --  baby steps, please.
> 
> I love you guys in advance,
> 
> Daniel
> 
> _______________________________________________
> K12OSN mailing list
> K12OSN at redhat.com
> https://www.redhat.com/mailman/listinfo/k12osn
> For more info see <http://www.k12os.org>