There is no single answer for this issue, so people usually use the approximation recommended by Kleine (2015) for the N:q ratio. Here, N represents the number of participants, and q represents the parameter. n can vary, for instance, 20, 10, 5 per parameter, depending on the study.

However, the best approach is always to include all the samples you can.

You can also read this publication: Sample Size Requirements for Structural Equation Models: An Evaluation of Power, Bias, and Solution Propriety - PMC (nih.gov)

Kline, R. B. (2015). Principles and practice of structural equation modeling. Guilford publications

To most accurately and comprehensively determine the optimal sample size in CFA and SEM, the best approach is probably simulation. A Monte Carlo simulation study allows you to specify a specific population model with all its expected parameter values and to simulate samples of different sizes to determine potential bias in fit statistics, parameter estimates, and standard errors as well as estimate statistical power. You can scrutinize the impact of many specific aspects on sample size requirements in a simulation, for example, missing data, data non-normality, variations in indicator reliability, etc.

See

Muthén, L. K., & Muthén, B. O. (2002). How to use a Monte Carlo study to decide on sample size and determine power. Structural equation modeling, 9(4), 599-620.

https://www.statmodel.com/bmuthen/ED231e/RelatedArticles/Article_097.pdf

See also my free mini-workshop on sample size planning with Mplus via simulation: https://www.goquantfish.com/courses/sample-size-planning-in-mplus

Thanks Dr. Jessica Cuartero Moñino.

Will go through the suggested book.

Dr. Christian Geiser,

Thanks.

Will go through the suggested readings.