Pre-Trial Detention Under the Young Offenders Act: A Study of Urban Courts

APPENDIX: METHOD AND SUPPORTING TABLES

Operationalizing Prior Convictions

The composite prior record variable was constructed by placing all prior record variables - including presence/absence of a record, number of past convictions, number of FTA/FTC charges, number of other administration of justice charges, total number of administration of justice charges, the most serious prior sentence (i.e., custody or probation or other), whether the youth is on probation or has outstanding charges - in a factor analysis and using the values of coefficients generated for each case as the composite index. The composite index is highly correlated with its component parts but at the same time takes a variety of indicators into account.

TABLE A.1 - THE EFFECTS OF SOCIAL AND LEGAL VARIABLES ON POLICE DETENTION AT ARREST, ALL COURTS COMBINED
Variable Module 1 Module 2
B (s.e.) Sign. B (s.e.) Sign.
Constant .527 (.085) .000 1.411 (.161) .000
12 to 16 years old vs. 17+ -.293 (.127) .022 -.206 (.147) .160
White/other vs. Aboriginal/black -.452 (.117) .000 -.216 (.135) .110
Living outside family setting -.720 (.138) .000 -.700 (.157) .000
Warrant at arrest (no/yes)     -2.455 (.544) .000
Number of outstanding charges     -.199 (.054) .000
Prior record (composite)     -.426 (.072) .000
Number of current charges     -.304 (.065) .000
Shoplifting (no/yes)     1.284 (.267) .000
No indictable vs. 1+ indictable current charge     -.751 (.154) .000
FTC/FTA charge at arrest (no/yes)     -1.033 (.226) .000
Model chi-square (df) 48.96 (3) *** 368.97 (13) ***
Block chi-square (df) 48.96 (3) *** 320.01 (10) ***
% correct predictions detained 38% 65%
% correct predictions released 79% 79%
Overall correct predictions 60% 72%
Nagelkerke's R2 .05 .33

Table A.2 - The Effects of Social and Legal Variables on Police Detention at Arrest, Halifax-Dartmouth
Variable Module 1 Module 2
B (s.e.) Sign. B (s.e.) Sign.
Constant 5.673 (1.740) .001 8.705 (2.358) .000
Being female .801 (.455) .078 .346 (.142) .536
Age at offence -.301 (.109) .006 -.362 (.142) .011
White/other vs. Aboriginal/black -.264 (.338) .434 .588 (.464) .205
Living outside family setting -.639 (.343) .062 -.650 (.439) .139
Number of current charges     -.670 (.161) .000
Prior record (composite)     -.106 (.202) .600
Warrant at arrest (no/yes)     -2.722 (.924) .003
Number of outstanding charges     -.283 (.134) .035
Shoplifting (no/yes)     1.484 (.673) .027
No indictable vs. 1+ indictable current charge     -1.838 (.378) .000
Model chi-square (df) 14.71 (4) ** 108.28 (10) **
Block chi-square (df) 14.71 (4) ** 93.57 (6) ***
% correct predictions detained 7% 60%
% correct predictions released 97% 92%
Overall correct predictions 71% 83%
Nagelkerke's R2 .08 .49

Table A.3 - The Effects of Social and Legal Variables on Police Detention at Arrest, Toronto (Both Courts)
Variable Module 1 Module 2
B (s.e.) Sign. B (s.e.) Sign.
Constant 1.777 (1.284) .166 2.152 (1.689) .194
Being female .596 (.328) .069 -.227 (.389) .561
Age at offence -.096 (.082) .243 -.015 (.103) .882
White/other vs. Aboriginal/black -.713 (.253) .005 -.893 (.3103) .004
Living outside family setting -1.863 (.373) .000 -1.501 (.450) .001
Number of outstanding charges     -.831 (.196) .000
Warrant at arrest (no/yes)     -2.445 (1.103 .027
Prior record (composite)     -.782 (.200) .000
Number of current charges     -.658(.151) .000
Shoplifting (no/yes)     2.295 (.878 ) .009
Model chi-square (df) 41.77 (4) *** 164.85 (9) ***
Block chi-square (df) 41.77 (4) *** 123.08 (5) ***
% correct predictions detained 59% 79%
% correct predictions released 75% 86%
Overall correct predictions 66% 82%
Nagelkerke's R2 .16 .52

Notes:

  • *** p<.001, ** p<.01, * p<.05. If the cell is blank, then the factor was not statistically associated with the condition. "-" means that the variable was not included in the model. "n.a." means the variable was excluded because of missing values.
  • The first value in the cell is the unstandardized B coefficient.
  • N=331, Cut value=.50
  • Dependent variable = police detention (detained vs. released)
Date modified: