I have done the analysis (3DVs and 8IVs) in SPSS' graphical interface and also using syntax, but I can't find the information for the 3 path values between the DVs anywhere in my output
Can anyone point me in the right direction?
Many thanks in advance
Bivariate analysis looks at two paired data sets, studying whether a relationship exists between them. Multivariate analysis uses two or more variables and analyzes which, if any, are correlated with a specific outcome.
I would like to suggest you to follow : https://www.researchgate.net/post/In_SPSS_how_to_write_a_code_to_repeat_a_linear_regression_analysis_for_500_times_same_data_pool_but_random_pick_each_time_Thanks
https://www.youtube.com/watch?v=ZyruUhomEnQ
http://processmacro.org/faq.html
Hi Andrew Gilbey - I am by no means a stats/SPSS 'expert', though I am currently using this method for my doctoral thesis, so thought I would share my experience.
In short, whether you are using the Mutlivariate GLM function or syntax to get a more detailed output ( https://www-01.ibm.com/support/docview.wss?uid=swg21476743 ), I do not think that SPSS will provide you/us with the path values between the DVs.
Instead, what you get in the output are the multivariate and univariate regression equations, and indeed the results of both of these tests.
There are probably other ways to find the paths between the DVs, though if I was keen to find this out, I would probably do SEM using AMOS?
Hope this helps, though I am happy to stand corrected if someone with more expertise suggests otherwise
Dear David Morse
I'm not quite sure I understand what you mean about subprogram. Does GLM/Multivariate in SPSS answer that question?
Regarding syntax, I used the instruction * Encoding: UTF-8.
GLM flighttestaverage P_MEAN_Raw WITH ADAPTOverall MAsseyEssayPercentage MAsseyWords MasseyAeroscience MAsseyInterviewMeanScore
/PRINT PARAMETERS
/LMAtrix
I had to chop off the * encoding bit which keeps appearing, as it wasn't running with that there. IT may be clear by now that I don;t quite know what I'm doing here!
Regards, Andrew
Hello again Andrew,
I understand the subprogram and basic set-up of your variable set. Unfortunately, I'm still a bit unclear as to what specific parameter(s) you desire.
If what you seek is the relationship between predicted scores on the two DVs, then use the "Save..." option to post predicted values to your data set/worksheet, then correlate them.
If you want the correlation among residuals (errors from the fitted models), those are given by the Residual SSCP matrix (available in "Options..." of GLM/Mult.).
If you want correlations between estimated regression coefficients across DVs, those can be computed from the Between SSCP matrix (available in "Options..."): r = covariance12 / sqrt(Var1 * Var2).
Good luck with your work.
David Morse
Thank you very much, David That's very useful.
My apologies for troubling you again - you have been very kind to help as much as you already have. I've got the Between-Subjects SSCP MAtrix, but I don't understand quite which information in the matrix I should extract for the formula you mention to get the three pairwise correlations between my 3 DVs. I also didn't follow what was covariance12. I appreciate you're busy and may not have a chance to answer. Thanks, Andrew
Hello Andrew,
Rerun the analysis, but before you execute it, click on the "Save..." button, and check the box for Predicted values (unstandardized). These will usually be saved by default as "PRE1", "PRE2," etc.
Then, correlate these scores: Analyze/Correlation/Bivariate and select the three saved variables.
BTW, "covariance12" was intended to mean the covariance between variable 1 and variable 2.
Good luck finishing up!
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS