On the Use of Growth Models to Study Normal Cognitive Aging

Ghisletta, P., Mason F., von Oertzen, T., Hertzog, C., Nilson, L.G., Lindenberger, U.: On the Use of Growth Models to Study Normal Cognitive Aging 

International Journal of Behavioral Development, Methods and Measures section. Manuscript accepted.

Click here to download the full article.

Click here to download the supplementary materials.

Abstract

Growth models (GM) of the mixed-effects and latent curve varieties have become popular 
methodological tools in lifespan research. One of the major advantages of GM is their 
flexibility in studying individual differences in change. We scrutinized the change functions 
of GM used in five years of publications on cognitive aging. Of the 162 publications that we 
identified, 88% test linear or quadratic polynomials, and fewer than 5% apply functions that 
are nonlinear in their parameters, such as exponential decline. This apparent bias in favor of 
polynomial decomposition calls for exploring what conclusions about individual differences 
in change are likely to be drawn if one applies linear or quadratic GMs to data simulated 
under a conceptually and empirically plausible model of exponential cognitive decline from 
adulthood to old age. Hence, we set up a simulation that manipulated the rate of exponential 
decline, measurement reliability, number of occasions, interval width, and sample size. True 
rate of decline and interval width influenced results strongly, number of occasions and 
measurement reliability exerted a moderate effect, and the effects of sample size appeared 
relatively minor. Critically, our results show that fit statistics generally do not differentiate 
misspecified linear or quadratic models from the true exponential model. Moreover, power to 
detect variance in change for the linear and quadratic GMs is low, and estimates of individual differences in level and change can be highly biased by model misspecification. We 
encourage researchers to also consider plausible nonlinear change functions when studying 
behavioral development across the lifespan. 
 

Tags: Growth Model; long-term change; normal cognitive aging, nonlinear mixed effects models, longitudinal research designs
Published Mar. 7, 2019 4:15 PM - Last modified Mar. 8, 2019 9:50 AM