ReviewAPPLIED MATHEMATICS

Statistical thinking for 21st century scientists

See allHide authors and affiliations

Science Advances  14 Jun 2017:
Vol. 3, no. 6, e1700768
DOI: 10.1126/sciadv.1700768

Figures

  • Fig. 1 Two thousand bootstrap replications of difference between AML and ALL proportional hazards coefficients.
  • Fig. 2 Observed proportion P of malignant nodes for 522 patients having P > 0; 322 patients (38%) had P = 0, as indicated by the large dot.
  • Fig. 3 Estimated prior density for frailty parameter θ, with median value θ = 0.09.
  • Fig. 4 Posterior probabilities of frailty parameter θ for three hypothetical patients.

Tables

  • Table 1 Proportional hazards analysis of the abandonment data.

    Estimated date coefficient of 1.660 is strongly negative, indicating decreased abandonment as study progressed.

    EstimateSEZ valueP valueBootstrap
    SE
    Distance0.2100.0722.9020.0040.068
    Date–1.6600.107–15.5080.0000.088
    Age–0.1540.084–1.8340.0670.082
    Sex–0.0270.076–0.3470.7290.078
    ALL0.1460.0821.7710.0770.083
    AML–0.0700.081–0.8640.3870.088

Navigate This Article