I use longitudinal data for the polynomial mixed-effects model and nonlinear mixed-effects model. My question is should Longitudinal data hold a normality assumption?
Normality assumptions are not typically required for longitudinal data, it is still important to assess the distributional properties of the outcome variable and ensure that the chosen analysis method is appropriate for the data at hand.
However, it is important to consider the distributional properties of the outcome variable and assess whether any transformations or non-parametric methods are needed to meet the assumptions of the chosen analysis method. This assessment can involve checking for skewness, kurtosis, and outliers, as well as conducting residual analyses to evaluate the adequacy of the model fit.
Alemayehu Siffir Argawu , be wary of responses that are AI generated (the above, using https://copyleaks.com/ai-content-detector, suggests it may have been AI generated ... actually 99.8% value, which is the highest I have seen). Anyway, the distributional assumption depend on the model, but the typical model does not make assumptions about the outcome variable, but about the residuals. This will be covered in longitudinal methods textbooks.
As you stated your analysis method is polynomial and nonlinear mixed effect model, hence the normality assumptions is not required. However, if your outcome of interest variables are continuous and if you are want to apply Linear mixed effect model(LMM), you should chech this distributional assumptions. Thank you for the questions!