> It is necessary to add END DATA command in your syntax:

Thanks for pointing that out! I slipped up when I copied the example. Sally

> MATRIX DATA VARIABLES= ROWTYPE_ group math /FACTOR=group. > BEGIN DATA. > MEAN 1 13.72 > SD 1 6.44 > N 1 66 > MEAN 2 14.65 > SD 2 7.35 > N 2 128 > END DATA. > ONEWAY math BY group /CONTRAST= -1 1 > /STATISTICS DESCRIPTIVES /MATRIX = IN(*). > --------------------------------------------------- > JosepMaria.Domenech@uab.es > Laboratori d'Estadistica Aplicada > Universitat Autònoma de Barcelona > --------------------------------------------------- > > S. Zitzer <sallyz@U.WASHINGTON.EDU> escribió en el mensaje de noticias Pine.A41.4.33.0104301429310.58370-100000@mead4.u.washington.edu... > > Voila! It's not that the material was deleted (this time), but is in > > a different place. For those who have the version 10 syntax manual > > (or want to look on the CD), it's page 756 in the Matrix Data chapter. > > > > However, that one requires the pooled error term. Here's an example that > > does not require the error term -- > > > > matrix data variables = rowtype_ group math /factor = group. > > begin data. > > mean 1 13.72 > > stddev 1 6.44 > > N 1 66 > > mean 2 14.65 > > stddev 2 7.35 > > N 2 128 > > oneway math by group/matrix = in(*). > > > > For 2 groups the Oneway anova is the same as a t-test. The above example > > can be expanded to more groups. This example is implicit on page 857 > > in the version 10 syntax manual under the heading "Matrix Output". > > > > Sally > > > > > On Mon, 30 Apr 2001, patricia hernandez wrote: > > > > > Hello List, > > > > > > Is there a way to do a t-test in SPSS when only the means, standard > > > deviations, and Ns are available? I don't have the raw data, but > > > would like to do some t-tests if possible. Thanks in advance. > > > > > > Regards, > > > Trish Hernandez > > > California Institute on Human Services > > > Sonoma State University > > > > > > >