Does anybody know how to report results from a GLM models? I have run a glm with multi-variables as x e.g Y ~ x1+x2+x3 on R. In the summary I get results for the interaction between each of my X and the Y and a common AIC value.
Any ideas?
Hi Ada Barbanera
https://besjournals.onlinelibrary.wiley.com/doi/full/10.1111/2041-210X.12577 - step-by-step guide
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4236119/ - mixed model example
http://psych.colorado.edu/~carey/qmin/qminChapters/QMIN09-GLMIntro.pdf
Some packages:
https://cran.r-project.org/web/packages/broom/vignettes/broom.html - tidy result
https://cran.r-project.org/web/packages/stargazer/vignettes/stargazer.pdf
https://cran.r-project.org/web/packages/huxtable/vignettes/huxreg.html
https://cran.r-project.org/web/packages/ClinReport/vignettes/clinreport_vignette_get_started.html
https://cran.r-project.org/web/packages/ggeffects/vignettes/ggeffects.html
Hi, may be useful ....
https://www.amazon.com/How-Report-Statistics-Medicine-Guidelines/dp/1930513690
My perspective on the question about reporting multivariate GLM is that you meant Analysis of Dispersion in which the observation Yij is an observation matrix of i records with j potentially correlated observations in each record. That is a multivariate situation that may be analyzed using a factorial design matrix X or some other independent single or multiple variable X matrix. The linear matrix would be
Y = XB where B is a matrix of parameters that one wants to test for significance.
This analysis is nicely described by CR Rao (1965).
The analysis is reported (long form) as an Analysis of Dispersion Table which is analogous to a more standard Analysis of Variance Table. That analysis provides a test of the significance of factors or parameters using a derived Chi Square statistic or a derived F-test.
The analysis of the effects on Y can further be analyzed described by Rao using Tests of Additional Information in which a type of Multivariate Analysis of Covariance can be carried out asking whether columns of Y and be corrected for covariation in other columns of Y. After correcting for those columns is there any additional information left in a subset of Y columns of particular interest.
Analysis of Dispersion is an extremely important extension of the General Linear Model to the real world of multiple varying observations. This earlier edition of Rao's text is the best description in my opinion.
Rao, C. R. (1965) Linear Statistical Inference and Its Applications. John Wiley & Sons, New York, 522pp.
While dealing with the question above, I have used SPSS and wondered, what components of the output I need to report, do I need to summarise them in a table, and what layout of the table would be suitable?
@ Tobias, its always advisable to summarize your SPSS output in tables for your readers because not everyone can understand the output from SPSS.
secondly, working with separate tables outside SPSS helps you remove unnecessary details from that you may not need to include in your results e.g. when using frequency tables cumulative % are not necessary.
- Badges
- Science method