12/09/2019 | Press release | Distributed by Public on 12/09/2019 05:43
Surveys derive a sample of the population and investigate the characteristics of people in the sample instead of referring to the whole population. There are two main types of sampling methods: probability sampling and non-probability sampling. In probability sampling, every unit has a known chance of being selected for the sample. This is not true for non-probability sampling, where units do not have an equal chance of being selected.
Probability sampling involves the random selection of a sample of the population. Since units are randomly selected, their probability of being selected can be calculated. Therefore, we can estimate the sampling error and extrapolate the findings to the whole population.
For LGBT issues, surveys with probability sampling (such as labour force surveys) allow us to know what share of the population is LGBT+. What is more, they provide data on the labour market and socioeconomic situation of LGBT+ people compared to those of the rest of the population.
The collection of LGBT statistics through probability sample surveys is not widespread, but there are a number of countries doing it. In its report Society at a Glance 2019, the OECD reviews the list of probability sample surveys used by OECD countries to collect data on LGBT+ communities. It presents the results in the chapter The LGBT challenge: How to better include sexual and gender minorities?.
In non-probability sampling, units do not have an equal chance of being sampled. The sample is built based on convenience, judgement calls or research purposes.
Snowball sampling is an example of non-probability sampling widely used to study LGBT+ communities. In snowball sampling, each respondent refers the interviewer to acquaintances meeting the criteria for selection. In the case of LGBT+ studies, each LGBT person interviewed would refer to other LGBT+ people they know.
This method is useful when the sample is very small or rare, and units (or persons) are hard to find. A survey based on non-probability sampling can provide in-depth information on the respondents, but the information only covers them. The data compiled only applies to respondents. It cannot be extrapolated to the whole population. In other words, the findings cannot be generalized, at least not reliably.