Table 3.

Model results from the four study a priori hypotheses of specific food cravings. Coefficient estimates are presented for negative binomial and Poisson's regression models, and odds ratios are presented for logistic regression models. Model statistics include the analytical method used, along with the appropriate pseudo-R² value. McFadden's R² should be interpreted with caution. Significant p-values (p < 0.05) are in italics.

	estimate	odds ratio	s.e.	Z-value	p > |z|	95% CI	model statistics
behavioural immune system
model 1: fruit
months pregnant	0.04	1.04	0.15	0.29	0.78	0.78, 1.41	logistic regression
nausea	0.60	1.82	0.72	0.83	0.41	0.48, 9.01	Nagelkerke R2 = 0.02
vomiting	0.35	1.42	0.85	0.41	0.68	0.31, 10.20
(intercept)	−2.81	0.06	1.33	−2.11	0.04	0.00, 0.70
model 2: unripe mango
months pregnant	0.09	1.10	0.14	0.68	0.50	0.84, 1.44	logistic regression
nausea	0.46	1.59	0.58	0.80	0.42	0.50, 4.90	Nagelkerke R2 = 0.01
vomiting	−0.45	0.64	0.73	−0.62	0.54	0.13, 2.42
(intercept)	1.00	2.73	1.13	0.89	0.37	0.31, 27.9
model 3: unripe tamarind
months pregnant	−0.03	0.97	0.12	−0.25	0.80	0.77, 1.22	logistic regression
nausea	−0.03	0.97	0.51	−0.05	0.96	0.36, 2.74	Nagelkerke R2 = 0.00
vomiting	−0.14	0.87	0.58	−0.24	0.81	0.28, 2.89
(intercept)	−0.70	0.50	0.96	−0.73	0.47	0.07, 3.22
nutrient seeking
model 1: total calories
triceps thickness	0.03		0.03	0.94	0.35	0.97, 1.08	negative binomial
BMI	−0.03		0.03	−0.93	0.35	0.91, 1.04	McFadden's R2 = 0.06
months pregnant	0.03		0.05	0.56	0.35	0.93, 1.13
toxicity/pathogenicity	1.06		0.18	5.92	0.0001	2.05, 4.14
(intercept)	5.29		0.54	9.78	0.0001	70.78, 564.17
model 2: nutritional foods
triceps thickness	0.02		0.04	0.50	0.62	0.94, 1.11	Poisson's regression
BMI	−0.02		0.05	−0.46	0.64	0.88, 1.08	McFadden's R2 = 0.21
months pregnant	0.05		0.08	0.68	0.49	0.91, 1.22
toxicity/pathogenicity	1.69		0.32	5.19	0.0001	2.93, 10.60
(intercept)	−1.57		0.91	−1.73	0.08	0.03, 1.22
resource scarcity
model 1: nutritional foods
triceps thickness	0.04		0.04	0.87	0.39	0.96, 1.14	Poisson's regression
BMI	−0.04		0.06	−0.81	0.42	0.87, 1.09	McFadden's R2 = 0.21
dietary diversity	0.00		0.01	0.27	0.79	0.99, 1.03
food insecurity	0.06		0.05	1.08	0.28	1.02, 1.23
months pregnant	0.09		0.09	1.03	0.30	0.88, 1.21
toxicity/pathogenicity	1.63		0.33	4.94	0.0001	2.74, 10.13
(intercept)	−2.06		1.32	−1.56	0.12	0.01, 1.63
social bargaining
model 1: toxin/pathogen
psychological distress	−0.01	0.99	0.11	−0.07	0.95	0.79, 1.24	logistic regression
food insecurity	0.25	1.29	0.10	2.42	0.02	1.06, 1.60	Nagelkerke R2 = 0.25
pressure to have a son	−2.70	0.07	1.03	−2.61	0.01	0.01, 0.41
number of living children	−1.18	0.31	0.49	−2.38	0.02	0.11, 0.77
pressure × living children	1.58	4.75	0.81	1.92	0.06	0.98, 25.99
(intercept)	−0.72	0.49	1.00	−0.72	0.47	0.07, 3.43
model 2: ‘hot’ foods
psychological distress	0.04		0.04	1.03	0.30	0.96, 1.13	Poisson's regression
food insecurity	−0.002		0.04	−0.07	0.95	0.93, 1.07	McFadden's R2 = 0.01
pressure to have a son	−0.29		0.31	−0.94	0.35	0.40, 1.34
number of living children	−0.19		0.18	−1.08	0.28	0.57, 1.16
pressure × living children	0.30		0.29	1.05	0.29	0.77, 2.36
(Intercept)	0.02		0.37	0.06	0.95	0.50, 2.10
model 3: unripe mango
psychological distress	0.19	1.21	0.14	1.37	0.17	0.93, 1.63	logistic regression
food insecurity	0.03	1.03	0.12	0.22	0.83	0.81, 1.34	Nagelkerke R2 = 0.07
pressure to have a son	−0.48	0.62	0.87	−0.55	0.58	0.12, 3.86
number of living children	0.45	1.56	0.61	0.73	0.47	0.50, 5.73
pressure × living children	−0.25	0.78	0.91	−0.28	0.78	0.13, 4.95
(intercept)	−0.50	0.61	1.20	−0.41	0.68	0.05, 6.15
model 4: unripe tamarind
psychological distress	0.16	1.17	0.12	1.36	0.17	0.93, 1.48	logistic regression
food insecurity	−0.10	0.91	0.10	−0.89	0.37	0.74, 1.11	Nagelkerke R2 = 0.11
pressure to have a son	−0.37	0.69	0.78	−0.47	0.64	0.13, 3.04
number of living children	−1.19	0.31	0.59	−2.10	0.04	0.09, 0.90
pressure × living children	1.27	3.57	0.84	1.51	0.13	0.68, 19.49
(intercept)	−1.73	0.18	1.01	−1.71	0.09	0.02, 1.25