Sample size Determination for Structural Equation Modeling (SEM)
DOI: 10.14416/j.ind.tech.2021.12.007
Keywords:
number of indicators (p) per factor (f/p), sample size, rule of thumbAbstract
Sample size was the problem that always raised to question of what size is correct or suitable for most researchers of structural equation analysis (SEM). Whether more subjects for high confidence in the accuracy of analysis/research or lesser subjects according to some contexts is plausible? Literature reviews show that sample size for SEM could be determined in a variety of ways depending upon different statistical formulas and rules including rules of thumbs. From a comparative study of sample size determination among 11 available formulas and rules, several plausible sizes are found to be numbers that range from large to small. In conclusion, a sufficient sample size for SEM is 200 but more or less than 200 is possible subject to complications of the SEM model itself and population size constrained.
References
[2] K.A. Bollen, Structural equations with latent variables, John Wiley & Sons, NY, USA, 1989.
[3] A. Boomsma, Non-convergence, improper solutions, and starting values in lisrel maximum likelihood estimation, Psychometrika, 1985, 50(2), 229–242.
[4] J.C. Nunnally, Psychometric theory, McGraw-Hill, NY, USA, 1967.
[5] P. Bentler, EQS structural program manual, CA. BMPD statistical software, LA, USA, 1989.
[6] H.W. Marsh, K.T. Hau, J.R. Balla, and D. Grayson, Is more ever too much? The number of ondicators per factor in confirmatory factor analysis, Multivariate Behavioral Research, 1998, 33, 181-220.
[7] A. Boomsma and J.J. Hoogland, The robustness of LISREL modeling revisited, Structural equation modeling: Present and future: A festschrift in honor of Karl Jöreskog, Scientific Software International Inc, Lincolnwood, IL, USA, 2001, 1-25.
[8] K.G. Joreskog and D. Sorbom, Lisrel7-A guide to program and applications, 2nd ed., SPSS, IL, USA, 1988.
[9] M. Piriyakal, Analytical techniques for Second Order SEM, Journal of Modern Management Science, 6(1), 97-111. (in Thai)
[10] J. Koran, Preliminary proactive sample size determination for confirmatory factor analysis models, Measurement and Evaluation in Counseling and Development, 2016, 49(4), 296-308.
[11] J.C. Westland, Lower bounds on sample size in structural equation modeling, Electronic Commerce Research and Applications, 2010, 9(6), 476-487.
[12] E.J. Wolf, K.M. Harrington, S.L. Clark, and M.W. Miller, Sample size requirements for structural equation models: An evaluation of power, bias, and solution propriety, Educational and Psychological Measurement, 2013, 76(6), 913–934.
[13] N. Koch and P. Hadaya, Minimum sample size estimation in PLS-SEM: The inverse square root and gamma-exponential methods, Information Systems Journal, 2018, 28(1), 227–261.
[14] P.S. Petraitis, A.E. Dunham and P.H. Niewiarowski, Inferring multiple causality: The limitations of path analysis, Functional Ecology, 1996, 10(4), 421-431.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
ผลงานวิจัยและบทความวิชาการที่ปรากฏในวารสารนี้ เป็นความคิดเห็นอิสระของผู้เขียน ผู้เขียนจะต้องเป็นผู้รับผิดชอบต่อผลทางกฎหมายใด ๆ ที่อาจจะเกิดขึ้นจากบทความนั้น กองบรรณาธิการและคณะจัดทำวารสารฯไม่จำเป็นต้องเห็นด้วยเสมอไป