Surveys are a common methodology to collect data for needs assessments of target populations. However, social scientists and survey methodologists are faced with numerous challenges when the population of interest is also considered hard-to-count (HTC). HTC groups are less inclined to take part in censuses and surveys for one reason or another. Examples of HTC groups include racial and ethnic minorities, immigrants, low-literacy groups, highly mobile populations, and populations displaced by natural disasters. Sexual minorities (lesbian, gay, bisexual, transgender – LGBT) also represent a HTC subgroup often under-represented or misclassified in surveys.  Reasons for this vary. First, the constructs of sexual and gender identity are complex making it difficult to quantify in the form of survey questions. Researchers must also decide for the purposes of their study whether it is important to measure sexual attraction, sexual identity, sexual behavior or all three? Additionally, if gender identity is important, then separate questions about current gender and gender assigned at birth may be necessary. The labels and terms ascribed to the construct can also be fluid and nuanced depending upon variables such as geographic region, primary language, or a particular subculture. These can dictate different items for a nationally representative data collection versus one for a specific service provider or local community.  Survey designers must also be careful to craft items that work for both heterosexual and LGBT respondents. A second broad challenge is that some members may feel stigmatized and afraid to report sexual or gender identity on a survey questionnaire (e.g. older populations or persons residing in rural areas). This may necessitate that data collections be self-reported as opposed to interviewer-administered. Electronic data collections via mobile devices such as iPads may be workable solutions in some cases. Finally, for some populations, (e.g. teenagers), opinions about sexual identity may not be well-formed and more difficult to precisely measure. This presentation will discuss some of these challenges citing examples of data collections with a successful track record making note of recent research to improve the data quality and reduce measurement error when studying LGBT populations for purposes of social and policy research.