Q-factor analysis

[Claudia]

Im am hoping to conduct a Q-factor analysis. Currently I have 1800
cases and 5 variables. Once this matrix is transposed using the FLIP
command in SPSS I will have 1800 columns and only 5 rows.

Is it still possible to do a Factor Analysis at this point?

Additionally, I understand that I will have to standardize the columns
(original cases) before proceeding, correct?

Any information on how I should proceed would be great. (i.e. - which
extraction method is preferable for Q-factor, which rotation method,
etc.)

Thanks,
Claudia

[Rich Ulrich]

On 14 May 2003 14:57:19 -0700, cbar...@uga.edu (Claudia) wrote:

> Im am hoping to conduct a Q-factor analysis. Currently I have 1800
> cases and 5 variables. Once this matrix is transposed using the FLIP
> command in SPSS I will have 1800 columns and only 5 rows.

I'm curious. What source leads you to hope to conduct
a Q-type factor analysis, especially, given data like that.

It has been my impression that Q-type were not very
useful with more than N=50 (say) because it becomes
so awkward to display trees of the results.

And then, I thought that the purpose of Q was
exactly the opposite: I thought that it let you skirt
the problem of having far more variables than people;
with few people, you cluster the people and describe
what emerges.

So, what are you working from?

>
> Is it still possible to do a Factor Analysis at this point?

[ ... ]

My experience is, factors become arbitrary when
the rows don't outnumber the columns by quite a lot.
Is there a "method to your madness" or is it just
an error in trying to apply some advice?

--
Rich Ulrich, wpi...@pitt.edu
http://www.pitt.edu/~wpilib/index.html