Skip to main content

Table 2 Comparisons of parametric and bootstrap estimates of confidence intervals for the eight dimensions of the SF-36 from the CPSW Study for Control vs. Intervention Groups

From: The use of bootstrap methods for analysing health-related quality of life outcomes (particularly the SF-36)

SF-36    CIs Interval
Dimension   Mean Difference Lower Upper Length Shape
Physical Normal (t-test) -2.6 -5.2 0.1 5.4 1.00
Function Bootstrap BCA   -5.2 0.0 5.2 0.98
Role Normal (t-test) -9.1 -16.0 -2.3 13.7 1.00
Physical Bootstrap BCA   -15.8 -2.3 13.5 1.02
Bodily Normal (t-test) -4.0 -8.2 0.2 8.4 1.00
Pain Bootstrap BCA   -8.1 0.3 8.4 1.03
General Normal (t-test) -2.4 -5.6 0.8 6.4 1.00
Health Bootstrap BCA   -5.6 0.8 6.4 0.99
Vitality Normal (t-test) -1.3 -5.0 2.5 7.5 1.00
  Bootstrap BCA   -5.1 2.4 7.5 0.98
Social Normal (t-test) -4.7 -8.9 -0.6 8.3 1.00
Function Bootstrap BCA   -8.7 -0.6 8.1 1.03
Role Normal (t-test) -1.1 -7.5 5.3 12.7 1.00
Emotional Bootstrap BCA   -7.1 5.6 12.7 1.11
Mental Normal (t-test) 0.2 -2.8 3.2 6.0 1.00
Health Bootstrap BCA   -2.8 3.2 6.0 0.98
  1. Mean difference = Intervention mean - Control mean
  2. BCa confidence intervals based 5000 bootstrap replications.