id	author	title	date	pages	extension	mime	words	sentence	flesch	summary	cache	txt
p5547p90t1d	Rachel Baird	The Effect of Misspecifying Random-Effect Time-Varying Predictors as Fixed on Estimates of Other Parameters	1904		.txt	text/plain	243	7	26	Although negative variance estimates are uninterpretable, results of the simulation show that estimates of the fixed effect of the time-varying predictor are as accurate for these cases as for cases with positive variance estimates, and that treating the time-varying predictor as random and allowing negative variance estimates performs well whether the time-varying predictor is fixed or random in reality. A second simulation study shows that treating the time-varying predictor as random may have poor convergence, except when allowing negative variance estimates.	cache/p5547p90t1d.txt	txt/p5547p90t1d.txt