Toward Understanding Homelessness: The 2007 National Symposium on Homelessness Research. Characteristics of Help-Seeking Street Youth and Non-Street Youth. Relationship Between Recidivism of Running Behavior and Risk/Problem Issues


Using correlative analyses,[7] we found a significant but relatively small relationship between recidivism of running behavior (repeated running away) and the number of problem domains reported by callers (Spearman's rho = 0.194, p < 0.01). Youth who, at the time of the call, reported two or more prior episodes of runaway behavior at the time of the call (repeat runners, n = 3,022) comprised 25% of the sample.

Further analyses, exploratory in nature, were performed to examine the relationship between reported problem issues and recidivism of running behavior. Based on the results of frequency analysis of reported risk issues, we selected general domains of problematic issues that substantially differentiated between repeat runners and non-runners. These issues included family dynamics, alcohol or drug use by the youth, alcohol or drug use by the family, physical abuse or assault, involvement of the youth in the judicial system, problems with youth or family service agencies, peer or social issues, school or education issues, issues related to GLBTQ status, and issues related to transportation. To assess the relative importance of these variables to repeat runner or non-runner status, these variables were entered into a regression model[8] (see Exhibits 12 and 13). Alcohol or drug use by the family was found to be a non-significant predictor of both repeat runner and non-runner status for our sample and was removed from the final regression model.

Exhibit 12
Variables Predicting Status as a Repeat Runner
(Youth Reported Having Run Away from Home at Least Twice)
for Youth Crisis Callers to the National Runaway Switchboard, 20002005
Variables B SE Odds Ratio 95% CI Sig.
Involvement of the youth in the judicial system 1.48 0.10 4.25 3.48 5.18 p < 0.01
Alcohol or drug use by the youth 0.88 0.09 2.42 2.02 2.90 p < 0.01
Family dynamics 0.60 0.07 1.82 1.60 2.07 p < 0.01
School or education issues 0.60 0.06 1.82 1.62 2.04 p < 0.01
Problems with youth or family service agencies 0.57 0.05 1.76 1.59 1.95 p < 0.01
Physical abuse or assault 0.32 0.06 1.38 1.22 1.55 p < 0.01
Peer or social issues 0.21 0.05 1.24 1.12 1.36 p < 0.01
Issues related to transportation 1.24 0.07 3.44 3.00 3.94 p < 0.01

 In general, status as a non-runner or repeat runner cannot be consistently predicted based on the types of problematic issues indicated, although repeat runners had experienced a wide variety of problematic issues. Experiencing problems in any of the problem domains included in the regression model significantly increased the probability of status as a repeat runner (p < 0.01), with the exception of the GLBTQ issue category, which was the only predictor reported more often by non-runners (3.7 percent) than by repeat runners (1.3 percent). Conversely, experiencing problems in any of the problem domains included in the regression model, with the exception of GLBTQ issues, significantly decreased the probability of being a non-runner (p < 0.01). Overall classification was inconsistent; on the basis of the nine significant predictors, correction classification rates were 96 percent for non-runners, but only 21 percent for repeat runners.

View full report


"report.pdf" (pdf, 1.3Mb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®