TABLE 3.
Women's dietary diversity (24‐h recall) a | W‐model 1 | W‐model 2 | W‐model 3 | W‐model 4 | W‐model 5 | W‐model 6 (interaction terms—women) | W‐model 7 (interaction terms—men) | W‐model 8 (market; subgroup n = 1,171) |
---|---|---|---|---|---|---|---|---|
Women's dietary knowledge | 0.19 * [0.092, 0.29] | 0.12 * [0.0052, 0.24] | 0.15 * [0.0046, 0.29] | 0.12 * [0.0059, 0.24] | 0.14 * [0.016, 0.26] | |||
Women's vitamin knowledge | 0.13 * [0.035, 0.22] | |||||||
Men's dietary knowledge | 0.18 * [0.087, 0.27] | 0.11 * [0.0030, 0.23] | 0.12 * [0.0058, 0.23] | 0.22 * [0.034, 0.42] | 0.12 * [0.00013, 0.24] | |||
Men's vitamin knowledge | 0.14 * [0.032, 0.24] | |||||||
Distance to market (min) | −0.0017 * [−0.0033, −0.00017] | |||||||
Interaction terms (knowledge and education) | ‐ | ‐ | ‐ | ‐ | ‐ | Not significant (see Table S2) | Significant (see Figure 3 and Table S2) | ‐ |
AIC | 4,303.0 | 4,310.0 | 4,303.2 | 4,311.0 | 4,301.0 | 4,306.1 | 4,296.3 | 3,631.0 |
Children's dietary diversity (24‐h recall) a | C‐model 1 | C‐model 2 | C‐model 3 | C‐model 4 | C‐model 5 | C‐model 6 (interaction terms—women) | C‐model 7 (interaction terms—men) | C‐model 8 (market; subgroup n = 613) |
---|---|---|---|---|---|---|---|---|
Women's dietary knowledge | 0.19 * [0.018, 0.36] | 0.12 [−0.077, 0.32] | 0.22 b [−0.026, 0.47] | 0.12 [−0.082, 0.32] | 0.19 b [−0.018, 0.39] | |||
Women's vitamin knowledge | 0.033 [−0.12, 0.19] | |||||||
Men's dietary knowledge | 0.19 * [0.018, 0.36] | 0.12 [−0.077, 0.32] | 0.13 [−0.070, 0.33] | 0.016 [−0.37, 0.40] | 0.099 [−0.11, 0.31] | |||
Men's vitamin knowledge | 0.21 * [0.020, 0.40] | |||||||
Distance to market (min) | −0.0017 [−0.0043, 0.00082] | |||||||
Interaction term (knowledge and education) | ‐ | ‐ | ‐ | ‐ | ‐ | Not significant | Not significant | ‐ |
AIC | 2,594.5 | 2,599.2 | 2,594.2 | 2,595.6 | 2,594.9 | 2,595.9 | 2,598.8 | 2,117.2 |
All models were adjusted for household size, household wealth quintile, woman's age, man's age, woman's education, man's education, geographical region and kebele‐level clustering (treatment effects were not significant). Children's models additionally adjusted for age of the child. Full model results are shown in Tables S3–S5.
P < 0.10.
P < 0.05.