Assessing the internal validity of a household survey-based food security measure adapted for use in Iran

Nutrition Journal

Table 4 Response characteristics, item calibrations, and item-fit statistics of items in the Isfahan Children's Food Security Scale

Item *	Affirmative responses(%)**	Item calibration***	SE****	Item infit	Item outfit
CH1 Few kinds of low-cost food	49.0	3.24	0.134	1.06	16.46
CH2 Could not afford nutritious meals	48.4	3.40	0.128	0.83	109.28
CH3 Not eating enough	32.0	5.86	0.094	0.52	0.98
CH4 Reduced size of meals	10.3	8.64	0.118	0.70	12.46
CH5 Skipped meals	6.8	9.53	0.145	0.85	4.97
CH6 Hungry	7.8	9.26	0.137	1.03	34.62
CH7 Did not eat whole day	3.3	11.03	0.222	1.26	1.16
Mean (item calibrations)		7.28
Standard deviation (item calibrations)		2.88
Discrimination parameter		1.50
Number of cases *****	990	501	501	501	501

*The full wording of each question specified the time reference (previous 12 months) and specified that the condition occurred because of a lack of money for food.
**Percent affirmative responses(%) of all households with children
***Item calibration indicates the severity of the item. The calibrations for the child scale were estimated with the mean and standard deviation constrained equal to the mean and standard deviation of the same items when scaled jointly with the adult items. This provides the most meaningful comparison of the severity of child items relative to adult items. The discrimination parameter of 1.5 for the child scale required to effect this transformation indicates that the discrimination of the child items is considerably higher when scaled alone than when scaled jointly with the adult items – primarily a result of the differing relationships between adult and child responses across households with children of different ages.
****Standard error of estimation of item calibration
***** The scaling sample omits households that affirmed no items or that affirmed all items, since those responses do not provide information about the severity of the items relative to one another.

ISSN: 1475-2891