The groups do not need to be equal. However, it can cause problems if the groups are very imbalanced and one group is extremely small (<=10%).
In addition, if you do MGA, your smallest group should still fulfill the minimum sample size requirements.
Search found 1287 matches
- Sun Jun 18, 2023 7:17 pm
- Forum: Method and application
- Topic: SmartPLS moderation and MGA interpretation
- Replies: 6
- Views: 6961
- Sun Jun 18, 2023 7:14 pm
- Forum: SmartPLS 4 - Questions and answers
- Topic: NaN values in outerloadings of single indicator construct categorical variable.
- Replies: 1
- Views: 3276
Re: NaN values in outerloadings of single indicator construct categorical variable.
Yes. Single indicator construct have a fixed weight and loading of 1. Thus, there is no variance and therefore also no t or p-values to calculate. The t and p-value are just not defined and therefore reporting them is also not meaningful. We have updated this in the recent version of SmartPLS to bet...
- Sun Jun 18, 2023 7:11 pm
- Forum: SmartPLS 4 - Questions and answers
- Topic: AIC in Smart PLS 4 and Model Selection Crieria
- Replies: 1
- Views: 2329
Re: AIC in Smart PLS 4 and Model Selection Crieria
Yes, based on the literature, the BIC and the and GM (Geweke–Meese criterion) are best for model selection and all other criteria are not well suited. SmartPLS does not include the Geweke–Meese criterion (GM), which is based on the model-complexity adjusted mean square error (MSE) from a saturated (...
- Sun Jun 18, 2023 7:04 pm
- Forum: Method and application
- Topic: Calculating correlation of latent Variable scores of IV, DV and its comparison with path coefficients of IV->DV
- Replies: 1
- Views: 2036
Re: Calculating correlation of latent Variable scores of IV, DV and its comparison with path coefficients of IV->DV
Path coefficients in PLS are usually standardized multivariate regression coefficients. Thus, they are the same as correlation coefficients if you have either just one IV or all IVs are orthogonal (uncorrelated). As soon as you have multiple correlated IVs they should not be the same, because the re...
- Sun Jun 18, 2023 7:02 pm
- Forum: Method and application
- Topic: 3-Way Interaction of Moderation
- Replies: 15
- Views: 26771
Re: 3-Way Interaction of Moderation
If you assume that they both moderate the relationship you should also include both in the same model. Testing separate models would be prone to omitted variable bias.
- Tue May 23, 2023 6:13 pm
- Forum: SmartPLS 4 - Questions and answers
- Topic: Moderation analysis in CB-SEM model
- Replies: 1
- Views: 2082
Re: Moderation analysis in CB-SEM model
We have not yet included a built-in feature for creating interaction effects. One way to do this in CB-SEM is create product-indicators for the new moderating construct. You can create these outside of the software and then import the dataset. But there are also other approaches and we are currently...
- Wed May 17, 2023 8:25 pm
- Forum: SmartPLS 4 - Questions and answers
- Topic: Unsuccessful exporting of model image
- Replies: 2
- Views: 1974
Re: Unsuccessful exporting of model image
Please contact the SmartPLS support: https://www.smartpls.com/support
- Mon May 15, 2023 6:33 pm
- Forum: Method and application
- Topic: Indicators and Latent Variables
- Replies: 1
- Views: 2063
Re: Indicators and Latent Variables
No. From a methodological point of view the more the better (if you assume reflective measurement). This will increase precision and reliability of the estimated construct scores. However, asking many questions to respondents can be strenuous and make them less attentive to the questions, which may ...
- Mon May 15, 2023 6:26 pm
- Forum: SmartPLS 4 - Questions and answers
- Topic: Where can I find "unstandardized" Latent Variable Scores in SmartPLS 4?
- Replies: 2
- Views: 2029
Re: Where can I find "unstandardized" Latent Variable Scores in SmartPLS 4?
In SmartPLS 4 you now have the option in the settings to select "unstandardized", which will provide unstandardized latent variable scores and model parameter.
- Wed May 10, 2023 7:48 pm
- Forum: SmartPLS 4 - Questions and answers
- Topic: PROCESS
- Replies: 1
- Views: 1792
Re: PROCESS
If the moderator is binary, it has only the two values 0 and 1. Otherwise, if the moderator is continuous, it should also show the effect at -1.
- Mon May 01, 2023 10:43 am
- Forum: Ongoing and Future Research
- Topic: Addressing Endogeneity with interaction terms (moderators, PLS-MGA)
- Replies: 1
- Views: 76150
Re: Addressing Endogeneity with interaction terms (moderators, PLS-MGA)
For instrumental variable approaches there are guidelines for normal regression models on how to deal with interaction terms and nonlinear relationships. I would think that most of this also applies to endogeneity concerns in the structural model of PLS-SEM. A very good article is Papies, D., Ebbes,...
- Thu Mar 02, 2023 12:04 pm
- Forum: SmartPLS 4 - Questions and answers
- Topic: Moderation analysis in PLS (SmartPLS4):
- Replies: 1
- Views: 1846
Re: Moderation analysis in PLS (SmartPLS4):
1. Yes. SmartPLS never standardizes the interaction term and always provides its unstandardized coefficients for them. 2. I am not sure, why it should be a problem to standardize the dependent variable. Dawson does not provide a sufficient argument in his paper why he believes that this is a problem...
- Thu Mar 02, 2023 11:52 am
- Forum: SmartPLS 4 - Requested features
- Topic: When saving datafile form IPMA include NON standardized results
- Replies: 1
- Views: 2482
Re: When saving datafile form IPMA include NON standardized results
You do not need to run IPMA anymore to get unstandardized LV scores. Simply select unstandardized in the PLS settings. You can then also export the unstandardized LV scores into a new dataset.
- Tue Feb 14, 2023 5:30 pm
- Forum: SmartPLS 4 - Questions and answers
- Topic: can't find blindfolding in smartpls 4
- Replies: 1
- Views: 3062
Re: can't find blindfolding in smartpls 4
SmartPLS 4 does not contain Blindfolding anymore, because it is an outdated approach. If you want to do predictive model assessment, you should do PLSPredict.
Please see also https://www.smartpls.com/documentation/ ... indfolding
Please see also https://www.smartpls.com/documentation/ ... indfolding
- Wed Feb 08, 2023 7:56 pm
- Forum: Method and application
- Topic: MGA: Differences between age, in different groups
- Replies: 2
- Views: 2024
Re: MGA: Differences between age, in different groups
In general, if you have a continuous variable using a split (e.g., median split, or any other arbitrary split) is less reliable than using the original continuous variable as a moderator in the model. In your particular case, you may also end up with quite small groups that have low power to detect ...