Patrick Gomez

Unisanté, Epalinges-Lausanne, Switzerland · Santé travail et environnement

Hi,

I have just installed and used spss 29. I was using spss 27.

I am analyzing data with a crossed random effects mixed model.

I am using syntax for this type of analysis. With the exact same syntax and data base I obtain different results with spss 29 and spss 27!

Specifically, the same model (that I call model 3) run with spss 27 was not giving me a warning whereas with spss 29 I get a warning (The final Hessian matrix is not positive definite although all convergence criteria are satisfied. The MIXED procedure continues despite this warning. Validity of subsequent results cannot be ascertained.).

Another case: with a slightly simpler model that I call model 2, I have no warnings but the results with spss 27 and spss 29 are not identical (e.g. BIC is different).

Is anyone experiencing the same or similar ?

In gsem of STATA we can test random-intercept and random-slope models (multilevel) (see example38g in the manual). STATA MULTILEVEL MIXED-EFFECTS "me" deals with multilevel mixed-models, in particular MIXED for continuous outcomes.

I asked myself: Do I get the same results if I use gsem or MIXED? For the moment my answer is yes and no.....

In MIXED we have several options: we can use ML or REML estimation method; we can define different residual variance structures,....

I ran a gsem 2-level random-intercept model (id defines level 2 and session_coded defines level 1 nested within level 2) using own data

gsem (rd <- mpa_level i.session_coded i.order M1[id])

I found out that I get exactly the same results with the following mixed model

mixed rd c.mpa_level i.session_coded i.order ||id:, ml cformat(%9.4f)

However, using reml is prefarable; furthermore, an heterogenous residual variance better fits the specific data rather than the default. So the "best" MIXED model I would use is

mixed rd i.session_coded c.mpa_level i.order ||id:, reml residuals(ind, by(session_coded)) cformat(%9.4f)

With this model, the results are quite different.

Why am I asking this question? Because in a second stage I would like to run multilevel-mediation analyses using gsem but ideally I would like to keep the level of "sophistication" that I have with MIXED (reml, residual variance, etc.).

Best regards,

This study Guelen, I., et al. (2008). "Validation of brachial artery pressure reconstruction from finger arterial pressure." J Hypertens 26(7): 1321-1327. showed tha the return-to-flow calibration with the arm cuff should be avoided in the sitting position. Yet, the Finometer user's guide recommends to perform it without making any distinction depending on body position. Also the FMS support recommended me to perform this calibration by saying

"The use of the return to flow calibration will improve the accuracy of your data, the RTF calibration is used to correct any deviations caused by the way the fingercuff is applied. I would always advise to use this function."

I am confused and I do not know if I should perform the RTF calibration in my current study with young healthy participants! I do not want to add something that might increase the measurement error rather than reduce it.

Can anyone help?

Patrick

I am using GLMM by SPSS to test different predictors of recognition memory performance for affective pictures. The outcome variable is binary (Yes/No); so I use binary logistic regression. Predictors include Gender, Age, etc. I notice that the p values I get for the predictors in the table "Fixed coefficients" are different from the p values I get with the pairwise contrasts. For instance, in the fixed coefficient table the gender effect has t = 3.185 p = .001 but in the pairwise contrast t = 2.727 and p = .006. In the table for the pairwise contrasts it says "adj. sig.". I cannot find what the p value is adjusted for. In general, for all predictors the p value of the pairwise contrasts is always lower than in the fixed coefficients table.

Has anyone an explanation?

Hi,

In the PHQ by Spitzer and colleagues question number 7 asks about symptoms/behaviors that could point to anorexia. Strangely, I cannot find any information on how to score/use this question, not even in the PHQ manual.

Does anyone know something about this? Is this question useless?

How does (short-term) memory for affective material (e.g. pictures) depend on the valence and arousal of the material? Are unpleasant stimuli better remembered than pleasant ones? Are high-arousal stimuli better remembered than low-arousal ones?

Any (review) articles on this topic is highly appreciated.

Are there (recent) (review) articles on gender differences in memory for affective material (e.g., pictures)?

Hi,

I am using GENLINMIXED (SPSS version 25) to analyze the effects of categorical (e.g., gender) and continuous variables (e.g., variable A) on a binary variable (using BINOMIAL). I see I can get estimated means for my binary variable for categorical variables together with graphs, e.g.estimated mean for men and women, which is great. How can I get estimated means for different levels of my continuous predictors (main effects) and for categorical predictors at different levels of continuous predictors (e.g., interaction between gender and variable A). In MIXED there is the command WITH, which is very useful. In GENLINMXED WITH does not seem to exist :-(

Any suggestions?

Hi,

Is it possible in GENLINMIXED (SPSS version 25) to specify a model with two crossed random effects (not nested) for a binomial outcome variable? (It is possible with MIXED).

I can do this in STATA https://www.stata.com/manuals13/memelogit.pdf

but I would prefer using SPSS

PAtrick