PLS SEM using SmartPLS P3 of P4: sample size using G power, testing hypothesis, reporting results

This channel uploads videos related to academic research. Viewers will find videos related to partial least square (PLS) structural equation modeling (SEM), PLS-SEM, fuzzy set theory in data analysis, SPSS, Fuzzy IPA, Fuzzy DEMATEL, NVivo, how to publish in good journals, how to write a research paper and writing a systematic literature review, etc.
Additionally, this channel also uploads videos related to the philosophy of life.
I am an academic researcher, trainer, and teacher.
Website: sites.google.com/view/ms-islam

[@user-un9jr3pl3i]

I have a few questions related to this. Please see if you can address them. It could be the case that I misunderstood something or do not understand the concept fully. Please don't hesitate to correct me. In the paper that you are analyzing, PLS-SEM is used. But when you use statistical tests in G power, it's Linear Multiple regression. So, my question is, how can we calculate the sample size for PLS-SEM using G power 3 if we dont select PLS-SEM as the statistical test. Another question is, when do we use G power, when do we have a list of population, or when do we not? At one point, if I am not mistaken, you mentioned that our study can't be generalized, so we are not going to use random sampling. Then, you mentioned the value of alpha-err-prob to be 0.05 and said that this value means our results can be generalized. Throughout the lecture, I understood that what we are analyzing can't be generalized, and we are testing the model and don't know the population. But at the time of calculating the sample size, the settings show that our results can be generalized. I find it a bit contradictory. It could be because I am misunderstanding the concepts. Please feel free to correct me. Lastly, I have a few questions related to my own problem. I have 5 independent variables and 1 dependent variable. The total number of arrows going into dependent variables is (5+2. 5 arrows come from independent variables, and 2 arrows come from factors that are not treated as constructs, and we dont have any hypothesis for them). So, for me, the number of predictors would be 5, right? Another question is, is it enough to manipulate the value of f-square to change the sample size? I am using PLS-SEM and need to calculate the sample size, and it would be nice if the answer is within the range of 130-170. Additionally, I am testing the model and validating the effects of the hypothesis rather than looking to generalize. I do not have any population size or sample list. I need to back my answer with academic reasoning. Thanks.

[@dr.mdshamimulislam]

Yes, we need to use linear regression to calculate the sample size using Gpower. For more detailed instructions, you can refer to the Gpower manual. When it comes to generalizing the results, there are specific criteria that should be considered. For more information, you can look into issues related to sample representativeness and generalizability. It appears that you need at least 5 predictors, which is correct. Additionally, adjusting the effect size allows you to manipulate the sample size. However, it's important to note that the chosen effect size should be supported by prior research. The minimum sample size provided by Gpower is adequate for testing the hypothesized model, but it may not suffice for generalizability purposes. I apologize for the delay in my response and any errors that may have occurred in the video. Your feedback is greatly appreciated as it helps me identify areas for improvement and correct any mistakes for future videos. Thank you for taking the time to provide your input.

[@tanjil95]

G*Power 32:02