Because the questions were embedded within a longer questionnaire that included several sensitive topics, the standard six-item subset of the HFSSM, rather than the full 18-item core module (Bickel et al.
The modifications I implemented were based on (1) re-phrasing of HFSSM questions used by Lawn and Harvey (2004) and in other studies conducted by Indian and Northern Affairs Canada, (2) food security survey instruments used in subsistence studies by the Alaska Department of Fish and Game (ADF&G), and (3) suggestions made by Kangiqsujuarmiut respondents who pretested the survey.
Instruments such as the HFSSM are designed to assess a one-dimensional scale of food insecurity and were developed in contexts where food access is entirely or almost entirely determined by purchasing power (Bickel et al.
Although the data presented here suggest that some of the HFSSM items can be used to provide a Rasch-scale measurement of one dimension of Inuit food security, the HFSSM questions clearly do not assess country food access very well, likely for a variety of reasons, which I will return to in the Discussion.
This finding reflects the intent of the original HFSSM.
The only dimension of Inuit food security reliably measured by the food security module in this study, and probably also in other studies using the HFSSM or modified versions of it (e.
The Rasch model results for the Kangiqsujuarmiut food security data suggest which HFSSM items seem to be more effective, notably, the three severity items: "Cut/skip meals," "Eat less," and "Not eat.
According to the standard six-item HFSSM module, 41% of households in Kangiqsujuaq could be classified as having low or very low food security, while 59% could be classified as food secure or marginally food secure.