Mode and language adjustments

The goal of this user's guide (PDF, 123 KB) is to present an approach to adjusting observed VR-12 and SF-36v.1® (MOS) scores obtained with the administration modes of self-report by phone ("self-report/phone"), proxy response by mail ("proxy/mail") and proxy response by phone ("proxy/phone") and the languages of Spanish and Chinese so that they are commensurate with scores that would have been obtained if the same respondents had used self-report by mail ("self-report/mail") in English. Responding to these surveys by self-report/phone, proxy/mail, or proxy/phone or in the Spanish or Chinese language typically results in an observed health status that is different from the health status that would be expected if the survey respondent had used self-report/mail in English. We refer to the difference in scores related to mode or language as the reporting bias.

We developed models to estimate average reporting biases for each of the "alternative" modes (self-report/phone, proxy/mail, proxy/phone) and languages (Spanish, Chinese). The estimated values of reporting bias for a given mode and language can then be subtracted from the observed score obtained in an alternative mode or language so that the score is more commensurate with the score that would have been obtained in English using self-report/mail. Details of the rationale and specific derivation of the methods for this approach are contained in a technical report made available to the National Cancer Institute.

SAS Code for Calculating Adjusted PCS MCS 8 Scales and SF12 Items by Phone Proxy and Languages