Видео

Dear Christian, thank you for this video! I tried this model by following your syntax. The weird thing is mplus did not give me the model fit information (Chi-square, RMSEA, etc.), but it gave sample statistic, UNIVARIATE sample statistic, MODEL RESULTS, and MODEL COMMAND WITH FINAL ESTIMATES USED AS STARTING VALUES. Do you know the reasons or any possible solutions to get the model fit information in this case? Thank you a lot.

Hi Dr. Geiser, thanks so much for all your hard work. Do you think it is possible to use the latent classes derived from an LCA to predict the latent profiles derived from an LPA using the same sample of people? Or would this run into a problem as the same participant would be in one group of LCA which is then predicting that same participant's membership in another class in an LPA? I’m struggling to find any examples of this in the literature but think it might be useful

This is interesting. I appreciate if you can share the data file so that we can follow.

thank you for the video!!

Hi Christian, thank you for this video, it is very helpful! I would like to ask the variables you input to the Mplus for this model are the original one, or the centered one? Thank you!

Hi Dr. Geiser. Thanks for this really informative tutorial. I am wanting to expand on this model to also include a predictor and mediating variable. I have repeated the steps outline above in order to estimate change and TAU1 for the mediator and am trying to regress change in y on change in m and x however, my model won't converge. I have included bootstrap = 1000 in the analysis command in line with your recommendations on mediation and moderation analysis. I have also included starts = 1000 and the best LL appears to have been replicated but Mplus still will not calculate S.E.'s. Have you any suggestions? Thanks in advance

Thank you so much for this clear video! If I understand correctly, for longitudinal CFA we need to account for correlations between indicators across time points (i.e. Y11 WITH Y12... based on your example), which is why we cannot use multi-group CFA to test longitudinal CFA, since indicators/variables are not independent. In this example, they are not correlated with each other, is this just for simplicity in this video for introductory purposes, or is it also possible not to manually correlate them?

Can you use MPLUS for IRT with continuous variables? Thank you!

Hello Dr. Geiser, do you have an example of a publication using SEM that you can share as a model paper?

Hello professor, could you provide reference research papers published using this technique. It would be helpful for justification when submitting to journals.

Nice video, love the multiple versions of the same model. Around 18:30 you mention we could use lavaan for the multilevel CFA version of the model, just a quick note to mention that lavaan doesn't currently support random slopes (only random intercepts), so it wouldn't be possible to reproduce this model exactly as intended using lavaan

Thank you so much

Hi there and thnx for this excellent video can you do a video on Stratified Cronbach Alpha for multi level scales !

It will be nice if you show a table

Thank you, Dr. Geiser, for your enlightening video.👏

Thanks, Dr. Geiser! I'm doing a bifactor SEM. Part of the indicators were measured on a 5-point Likert scale and the other part of the indicators were binary. In this case, do I need to dichotomize the 5-point variables and redo the SEM as a sensitivity analysis?

A very lucid lecture for beginners learning psychometrics!

Thank you so much!

why not just use R? what a hassle and old-fashioned program, it also costs money even

Hi. I am experiencing major issues conducting this in MPlus. I made it to the end only to learn that you can't use BCH weights with ALGORITHM = INTEGRATION to regress C (classes) onto convariates. My ON command will not run with the BCH weights included, and removing them messes up the class structure. I am stuck and don't know how to proceed. If remove the weights, how do I retain the class structure for the covariate part of the analysis?

Could you help me understand how SEM is a different framework than IRT/CTT, or is SEM a subclass of IRT/CTT?

Regarding class proportions. The class output in this file seems to be based on most likely class membership, but in the LCA plot classes seem to based on the estimated model and posterior probabilities. This leaves slightly different proportions for each class. How can we get a save file with the class memberships based on the estimated model so that the latent class prevalences are consistent between the plot and dataset for further analysis.

So thankful I stumbled upon your channel. Really appreciate the clarity and intuition you teach with.

Learned a lot here, appreciate it!

can you share your email id sir?

Really well explained...Thank you sir

Great video! Thank you so much! I have a question. How to learn the proportion of missing values in our data (e.g., min.%-max.%) in Mplus or SPSS?

Thank you for the video. Does the sparseness issue only arise when there is a discrepancy between the two chi-square values? I have a sample size of 600 with 6 items, and both p-values are around 0.99. In this case, should I just interpret this as not significant, or could it indicate a sparseness problem?

Thank you so much

hi Dr. Geiser. i wonder that how do we expalin transitiion odds in mplus output

Thank you! Your videos have helped me a lot.

Is it a good idea to force a panel dataset to ‘balanced panel’ in panel multiple regression? Forcing it here means deleting some observations or populating missing values with zeros so that the panel becomes balanced. 4:32

Dear Dr. Geiser, Thank you very much. I was struggling to find a method to analyse the data for my research when I came across your video. I hope you read this comment and give me some advice: I plan to study the creativity of 5-year-old children and explore the impact of educational practices by preschool teachers and parents on children's creativity. The research sample consists of 60 preschool teachers from 30 classes (using questionnaires), 120 5-year-old children (random 4 students in a class, creativity tests), and 120 parents of these children (using questionnaires). Do you think I can use the simple SEM model and manually synchronize the data of the children with the teachers (calculating the average score and matching them) or should I use Multilevel Regression Analysis? I find your guidance in the video very suitable for my research. Thank you very much!

Thank you so much for your helpful video. I was wondering about your codes. In analysis command, you put STARTS=1000 100; STITERATIONS=50; , are these something I always need to put in? I had LRTstarts= 0 0 300 20. I have 8 binary variables (yes, no). would you help me understand this difference?

Hi Christian, extremely helpful video, as always - thank you! I have a question regarding negative residual variances in ESEM models, run with Mplus. Is it possible to fix negative residual variance for a given item to zero (or to a non-negative value)? Trying to do so yields an error message from mplus ("cannot be done with EFA"), but based on the Mplus Forum, I feel like there might be an option for doing so. Using "model constraint" doesn't seem to be possible. I just want to see if the model runs when I fix this one error variance as I am getting the message that the model does not converge due to the item in question. Thank you!

Thank you so much for teaching us good stuff ❤

You're a boss. Thank you for this.

Hello Mr Geiser, first of all thank you for this video (and all the others 😅). In my tests I want to get the confidence interval in an interclass correlation. But I can't get it with CINTERVAL, thank you for your help.

Hello and thank you for the great Video! When I run my CFA I get the following Error Message, which is quite similar to the fist one shown in the video: THE MODEL ESTIMATION TERMINATED NORMALLY THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES COULD NOT BE COMPUTED. THE MODEL MAY NOT BE IDENTIFIED. CHECK YOUR MODEL. PROBLEM INVOLVING THE FOLLOWING PARAMETER: Parameter 108, SN BY SN23 THE CONDITION NUMBER IS 0.617D-10. THE ROBUST CHI-SQUARE COULD NOT BE COMPUTED. Unfortunately there's only the parameters 1 to 68 in the tech 1 section in the output, so i can't find parameter 108. I'm a bit lost here. Do you have an idea hot to fix this?

Thanks! Is it possible to use FIML for dichotomous variables? Some of my variables are independent and others are dependent

Thank you, Dr. Geiser! @QuantFish Can you list the exogenous variables' variances as parameters when using WLSMV to get Mplus to estimate more of the missing data? Or is this not possible since WLSMV uses a partially pairwise deletion estimation? In other words, is listing the exogenous variables' variances as parameters only possible when using ML and MLR estimation methods? Thank you for all your videos. They are very helpful and informative!

Hey Christian, thanks a lot for the video which is a great start to get an idea of what PGCM are. Could you maybe point me to good introductions/tutorials on PGCM (preferably in R). One thing I would be especially interested in is whether or not there are some guidelines on how big your sample should be (of course depending on the complexity of the model) in PGCM.

Is there a specific situation where a Chi Square difference test MUST be used instead of the HTMT or the Fornell-Larcker criterion for discriminant validity testing?

I found this video excellent, really. I think it is the first time I truly grasped the concept. Thank you so much!

Morning Christian, do you have a video on creating a Factor Scores file and use it in path analysis??

Thank you for the video, it was really informative! I have a question though. After obtaining the optimal fit model, I included gender as a covariate to compare the class distribution by gender. However, I noticed that the results differed from the initial analysis. For example, if the distribution by class was 0.4, 0.4, and 0.3 in the initial analysis, the result after adding the covariate showed distributions of 0.2, 0.3, and 0.5 respectively. Could you explain why there's a difference in the results when gender is included as a covariate?

Thank you. Might I ask this missingness technique is only available for regression models (not things like mixture modelling)?

Sir please upload a detailed video on multilevel SEM with different sample unit in R

Hi! Thank you for your videos! However, when I try to use the same way to write my model syntax, it's giving me warning saying that " Variable on the left-hand side of an ON statement in a | statement is a WITHIN variable. The intercept for this variable is not random." And it's referring to my outcome variable. Furthermore, I'm also getting several errors in the model command, and it's telling me that within-level variables cannot be used on the between level. As I specified my outcome variable beta1J | Y on x1 x2 x3, when I write Y on W1 W2 under the %between% section, it's producing errors.... I'm wondering where there's a fix for that... Thanks again.. :)

compute in spss please