After the study closes, a nonresponse analysis compares HMS respondents and nonrespondents for demographic differences that could be correlated with nonresponse bias. It is important to assess whether HMS respondents are collectively representative of their school’s general population or if certain groups are disproportionately represented.
This analysis informs the generation of nonresponse bias weights, as necessary. Nonresponse weighting ensures that the data represents the full sample.
This step is especially important for this research: the pilot study and HMS 2007 revealed that these two populations differed both in demographic characteristics and the distribution of mental health measures.