TABLE 2—
Variable | Interrogative (n = 9631) | Interrogative + Social Context (n = 9596) | Noninterrogative (n = 9630) | Noninterrogative + Social Context (n = 9632) | No-Intervention Control (n = 9602) | Statistic χ2/F | P |
Demographic variables | |||||||
Age, y, mean ±SD | 60.5 ±6.0 | 60.5 ±6.0 | 60.6 ±6.0 | 60.4 ±6.0 | 60.7 ±6.1 | 1.9 | .101 |
Women, no. (%) | 5001 (51.9) | 4838 (50.4) | 4858 (50.4) | 4945 (51.3) | 4911 (51.1) | 6.3 | .181 |
SES, no. | 9604 | 9561 | 9595 | 9604 | 9567 | 1.3 | .995 |
Low, no. (%) | 2534 (26.3) | 2502 (26.2) | 2505 (26.1) | 2486 (25.9) | 2514 (26.3) | ||
Medium, no. (%) | 4393 (45.7) | 4353 (45.5) | 4369 (45.5) | 4429 (46.1) | 4370 (45.7) | ||
High, no. (%) | 2687 (28.0) | 2706 (28.3) | 2721 (28.4) | 2689 (28.0) | 2683 (28.0) | ||
Married no. (%) | 6722 (69.8) | 6635 (69.1) | 6673 (69.3) | 6779 (70.4) | 6710 (69.9) | 4.5 | .347 |
Performed FOBT in the past, no. (%) | 2673 (27.8) | 2601 (27.1) | 2668 (27.7) | 2528 (26.2) | 2624 (27.3) | 7.3 | .123 |
Test performance | |||||||
At 6 mo, no. (%) | 942 (9.8) | 988 (10.3) | 884 (9.2) | 923 (9.6) | 817 (8.5) | ||
χ2 | 9.4 | 18.0 | 2.69 | 6.74 | |||
P | .002 | < .001 | .101 | .009 | |||
Effect sizea (95% CI) | 0.05 (0.04, 0.06) | 0.06 (0.05, 0.07) | 0.02 (0.01, 0.03) | 0.04 (0.03, 0.05) |
Note. CI = confidence interval; FOBT = fecal occult blood test; SES = socioeconomic status.
Effect size for the difference between 2 independent population proportions, specifically a difference between arc sin transformations of the proportions. Each proportion is transformed:
Ø1 = 2arcsin√P1
Ø2 = 2arcsin√P2
And the effect size is:
ES = Ø1- Ø2
The arc sin transformation allows for equal “units of detectability.”33(p181)