Table 2.

Normality and convergent, discriminant and predictive validity of the risk-perception measures (n= 566)

CriteriaRisk-perception measure
Numerical (0-100)Verbal (1-5)Comparative (1-7)
Normality
    Shapiro Wilks test (W)*0.9760.9920.981
P < 0.0001P = 0.006P < 0.0001
Convergent validity (correlation with other risk-perception measures)
    Verbal0.62
P < 0.0001
    Comparative0.600.72
P < 0.0001P < 0.0001
Discriminant and predictive validity
    Correlation with breast cancer worry0.510.510.52
P < 0.0001P < 0.0001P < 0.0001
    Correlation with measures of actual risk Gail model estimate0.260.350.33
P < 0.0001P < 0.0001P < 0.0001
    Degree of family history0.280.270.25
P < 0.0001P < 0.0001P < 0.0001
    Correlation with breast cancer screening annual mammogram0.190.120.10
P = 0.0019P = 0.053P = 0.110
    Association with Gail risk affected by: Race (Black vs White)Not significantWhite > BlackWhite > Black
P = 0.103P = 0.034P = 0.014
    Income (>$50,000 vs ≤$50,000)Not significantNot significantHigher income > lower income
P = 0.444P = 0.061P = 0.032
    Age (<50 vs ≥50 y)Not significantNot significantNot significant
P = 0.887P = 0.555P = 0.823
    Education (completed college or more vs less than college)Not significantNot significantMore education > less education
P = 0.525P = 0.175P = 0.021
  • * Value of 1 indicates complete normality; significant P value indicates nonnormal.

  • P value for interaction term.