Hi,
I'd like to conduct Latent Class Growth Curve analysis in order to explore inter-individual differences in the evolution of my variable. How could I estimate the a priori sample size required to conduct this type of analysis? Does anyone might have recommandations please?
Thanks
A priori determination of sample size allows for reasonable power to detect the mean of the quadratic slope. Based on the Quadratic Growth conditions considered to ensure a satisfactory power rate higher than .80 across conditions, quadratic LCM studies with four measurement points should be based on samples of at least n = 250 participants but ideally n = 400. When 6 measurement points are available, samples sizes of at least n = 100 seem sufficient, albeit n = 150 would be better. Finally, when 6 measurement points are available, sample sizes as low as n = 50 seem sufficient although n = 100 would be ideal. For these reasons, Bayesian LCM may be considered as a potential alternative when sample sizes are small or when high rates of nonconvergence are observed.
The sample size requirements to detect an intervention effect are study-specific, and though longitudinal (latent growth) studies may involve 'small samples', it is very important to go for large samples sizes that can make small effect sizes detectable! Effect size, I may recommend between 0.6- 0.8.
Sample size can be calculated using formula or by simulation/iteration (most use Monte Carlos) which you can seek the help of a good 'particular software coder' with strong knowledge of Statistics and modeling.
The G*Power software package (version 3.1.4), this software is freely downloadable and capable of estimating sample size for many of the different statistical tests.
You can use the power calculation a priori, because the latent growth model is based on the method of structural equations, this online calculator is oriented for those models according to the number of latent and observable variables.
https://www.danielsoper.com/statcalc/calculator.aspx?id=89
A priori determination of sample size allows for reasonable power to detect the mean of the quadratic slope. Based on the Quadratic Growth conditions considered to ensure a satisfactory power rate higher than .80 across conditions, quadratic LCM studies with four measurement points should be based on samples of at least n = 250 participants but ideally n = 400. When 6 measurement points are available, samples sizes of at least n = 100 seem sufficient, albeit n = 150 would be better. Finally, when 6 measurement points are available, sample sizes as low as n = 50 seem sufficient although n = 100 would be ideal. For these reasons, Bayesian LCM may be considered as a potential alternative when sample sizes are small or when high rates of nonconvergence are observed.
Pradeep Paraman thanks for your answer, do you have some references concerning the estimations you suggest?
Maas, C. J. M., & Hox, J. J. (2005). Sufficient sample sizes for multilevel modeling. Methodology: European Journal of Research Methods for the Behavioral and Social Sciences, 1, 85-91.
Fan, X. (2003). Power of latent growth modeling for detecting group differences in linear growth trajectory parameters. Structural Equation Modeling, 10, 380-400.
Fan, X., & Fan, X. (2005). Power of latent growth modeling for detecting linear growth: Number of measurements and comparison with other analytical approaches. The Journal of Experimental Education, 73, 121-139.
Good luck.
Cristian Ramos-Vera thank you for your answer! Would you have some references that I could rely on to justify the using of this type of calculator please?
Thanks
The link reports hundreds of references that have used and referred to the use of this online calculator
https://scholar.google.es/scholar?cites=2432655585030014484&as_sdt=2005&sciodt=0,5&hl=es
https://scholar.google.es/scholar?cites=10571790929695343515&as_sdt=2005&sciodt=0,5&hl=es
As a rule of thumb, for any variable in the design, you will need at least 30 individual recordings.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size