Several factors were taken into account in the analysis of item allocation rates for this study. As expected, item allocation rates for the selected demographic characteristics were lower than the allocation rates for the selected social characteristics. Of particular concern are the high allocation rates for the marital status, when last worked and employment status recode characteristics, which show an increase in item allocation of 5 and 10 percentage points respectively, in 2008. The consistent increase in allocation rates for marital status is likely due to the revision made to the questionnaire in 2008, which moved this question from the front section to the back pages of the instrument with marital history. When comparing item allocation by institutional/noninstitutional GQ facilities, the study found that the highest increase in item allocation rates occurred in noninstitutional GQ facilities. When comparing item allocation rates by the methodology used to collect the data, in general, item allocation rates were lower for persons interviewed by self-response in adult correctional facilities, while person level interviews with the sample resident provided more complete data in juvenile and nursing/skilled-nursing facilities. In noninstitutional facilities, interviews by self-response seem to yield lower allocation rates than interviews in person or by telephone. We must take into consideration that in many instances respondents self-select into these modes of data collection, thus we cannot conclusively state that the use of a proxy in the collection of person level data is the problem. Reluctant respondents may provide equally incomplete data if interviewed in any mode. In some instances, proxy interviews are most likely the least cooperative respondents. The use of administrative records, although an important tool in the GQ data collection operation, may not be suitable for collecting social characteristics data in GQs. For example, it does not appear suitable to collect data in adult correctional facilities and college/university student housing, both which consistently showed high item allocation rates for all the social characteristics when data are obtained from administrative records. In conclusion, the data showed that item allocation rates in 2008 varied based upon which mode and method of data collection were used in the GQ operation. Identifying the most suitable modes and methods by specific GQ types may yield lower allocation rates.