Bill and Magnus

25 Mar 1996 11:46:08 -0500

I have recently downloaded Circumgrids from:

I was told this was easier than using the alternative means:

You will need a printer and a means to unzip the program.

Regarding use of Circumgrids, I would say that if there is a
version which allows you to store data try to get it. I entered
one grid 5 times and was getting tired of doing this. This may
be less of an issue as you get used to the program but can be a
source of frustration. Data entry is easy as is editing, that
is if you spot the mistake before you move on to the next stage.
Otherwise the whole grid would have to be re-entered.

Regarding the usefulness of the analyses, the principal
components option(s) were pretty user friendly, it does separate
element and construct analyses (though I am not sure if the
Slater version does so or not ... a question regarding this
follows below). I got pretty lost with the Multiple group factor
analysis, some form of instructions regarding use would have been
welcome here). These were the only options I used in the program
so my comments have been limited to their use.


As my questions may have some general interest I have included

1. My understanding of Ingrid is that the rows and columns
(constructs and elements) are analysed together. Will the
results obtained from the Slater analysis option in Cirumgrids
be the same as if they were obtained from Ingrid, or are the rows
and columns analysed separately as per a standard PCA analysis?

2. The raw-score factor analysis option refers to Euclidean
distance, is this measure the sum of squared distances or the
square root of the sum of the squared differences?


Bob Green.