Table 3.
Percent match comparisons by domain.
| Domaina | Media (n=17), n (%) | Scientific (n=14), n (%) | χ2 (df) | P value |
| Analysis match | 12 (18) | 52 (93) | 69.6 (1) | <.001 |
| Attrition match | 17 (20) | 63 (90) | 68.8 (1) | <.001 |
| Design match | 8 (47) | 14 (100) | 10.4 (1) | .001 |
| Population match | 61 (46) | 108 (97) | 69.6 (1) | <.001 |
| Exposure match | 48 (48) | 80 (95) | 45.9 (1) | <.001 |
| Results match | 19 (56) | 28 (100) | 16.3 (1) | <.001 |
| Comparison match | 40 (42) | 70 (86) | 31.0 (1) | <.001 |
| Limitations match | 12 (71) | 14 (100) | 4.9 (1) | .03 |
| Outcomes match | 39 (59) | 45 (80) | 5.3 (1) | .02 |
| Objectives match | 50 (98) | 42 (100) | 0.8 (1) | .36 |
| Caveats match | 15 (88) | 5 (36) | 9.3 (1) | .002 |
aMultiple domains had more than one parameter, as shown in Multimedia Appendix 2. Therefore, it was possible for a domain to have a number of matches higher than the actual sample size.