Table 3.
Estimated mean body satisfaction/percentage reporting disordered eating behaviors by feminist identification groups
| Unadjusted | Adjusted1 | |||||||
|---|---|---|---|---|---|---|---|---|
| Feminist Identified | Feminist Beliefs | Not Feminist | F, p | Feminist Identified | Feminist Beliefs | Not Feminist | F, p | |
| Body satisfaction [M (SE)] | 41.0 a (±0.86) | 37.7 b (±0.47) | 38.1 b (±0.54) | F = 5.57, p= 0.004 | 40.7 a (±0.91) | 38.4 b (±0.65) | 38.6 b (±0.71) | F = 4.00, p =. 0.019 |
| Binge Eating2 | 15.6% | 17.1% | 15.1% | F = 0.41, p =. 0.66 | 14.8% | 15.5% | 14.6% | F= 0.09, p =. 0.91 |
| Less Extreme UWCBs2 | 46.1%a | 56.5%b | 47.5%a | F = 5.50, p= 0.004 | 47.9%a | 56.3%b | 47.1%a | F = 5.16, p =. 0.006 |
| Extreme UWCBs2 | 14.4%a | 20.6%ab | 22.1%b | F=2.37, p =. 0.094 | 16.1% | 21.1% | 21.7% | F= 1.30, p =. 0.27 |
| Dieting2 | 50.0%a | 60.1%b | 56.5%ab | F= 3.02, p =. 0.05 | 51.8% a | 60.0% b | 57.1% ab | F = 2.04, p =. 0.13 |
a,b
Within rows, cells with different superscripts are significantly different (p < 0.05)
1
Adjusted for race, educational attainment, age, and BMI
2
For categorical dependent variables, least square means can be interpreted as predicted probabilities of engaging in each behavior. Due to the large sample size, the Central Limit Theorem effectively treats them as normally distributed continuous variables.