Citation

Teare J, Smith G, Osgood D, Peterson R, Authier K, Daly D. J. Child Fam. Stud. 1995; 4(1): 89-101.

Copyright

(Copyright © 1995, Holtzbrinck Springer Nature Publishing Group)

DOI

10.1007/BF02233956

PMID

unavailable

Abstract

For 472 consecutive days we monitored the number of youths residing in a crisis shelter for adolescents, the average length of stay for the youths residing in the shelter on each day, and the number of problem behaviors occurring within the shelter on a daily basis. We analyzed these data using a combination of time series and logistic regression techniques to fit a model that would predict the occurrence of a problem behavior on any given day. After controlling for significant time trends in the data, our results indicate that both the number of youths in the shelter and the proportion of youths who have resided in the shelter longer than 14 days are significantly associated with a problem behavior occurring on any given day. The number of youths in the shelter and the probability of a problem behavior occurring correlated significantly. However, as the proportion of youths who resided in the shelter longer than 14 days increased, the probability of a problem behavior occurring decreased.