Re: compare means layers

[Williams, Dana]

Dear John,

Sorry it's taken awhile for me to respond with more details about this 
bug. I think the inconsistency between PSPP and SPSS originates in the 
MEANS syntax command (not PSPPIRE like I originally thought), 
specifically the placement/presence of the BY subcommand. I wonder if 
the PSPPIRE dialogs could be left alone if the syntax [sub]commands 
behind them were reversed. Anyway, here's an attachment with screenshots 
of the inconsistent output from both programs. Hope this helps. Congrats 
on the release  of v1.0, btw!

Dana


On 3/7/2017 6:23 AM, John Darrington wrote:
> On Mon, Mar 06, 2017 at 08:06:26AM -0800, Dana Williams wrote:
>       Some of my Stats students were following along on PSPP as I demonstrated
>       using the compare means function on SPSS (descriptive, not t or F). All
>       of this is through the GUI (for both SPSS and PSPP). When adding a
>       second layer in SPSS, the independent variable is nested with the first
>       IV, in a single output table. However, in PSPPIRE, using the layer tool
>       separates the layers into separate tables, but adding multiple IVs into
>       a single layer combines (and nests) them together. The functionality of
>       PSPPIRE's layer tool seems opposite of SPSS's. Is that correct or am I
>       missing something?
>       
>
> No I don't think you're missing anything.  I number of those people have
> complained about the same issue.
>
> If someone can provide a detailed description with diagrams about how these
> dialogs could be improved then maybe it could be considered for a future
> release.
>
> J'

--
Dana Williams, Ph.D.
Associate Professor
Department of Sociology
California State University, Chico

PSPP_compare_means_bug.odt
Description: application/vnd.oasis.opendocument.text