2.4a Nonprobability and Probability Sampling

Including every person in a study would be cost-prohibitive and unnecessary, and even the best effort to do so would still miss some people. Thus, sociologists use a sample of the group (or population) they are studying. The population includes all individuals to whom a study’s results will be generalized. In Collins et al. (2020) and Calarco et al.’s (2021) studies, the population was heterosexual dual-income parents. Therefore, their results do not apply to same-gender couples, single parents, or heterosexual single-income parents. Further, their study focused on mothers with young children. Their results might look different if they had included mothers of older children.

Probability Samples

Both research teams relied on a sample of the population of interest. Collins’ (2020) research team used probability sampling to draw analytical samples from the randomly selected 60,000 people from all 50 states and the District of Columbia that are part of the U.S. Current Population Survey. Their analysis used data from participants who were married heterosexual working parents. Therefore, the results are generalizable to the broader population sharing these characteristics. In a probability sample (or a random sample), everyone was randomly selected for inclusion in the study. Random selection means that each person has an equal and non-zero chance of being chosen to take part in the study. Large probability samples are more likely to statistically represent a geographical area or another variable of interest, like social class or marital status.

Nonprobability Samples

In contrast, Calarco (2021) research team used nonprobability sampling. Nonprobability samples do not use random selection but instead rely on sampling techniques based on convenience or availability. The study team recruited participants at prenatal clinics in Indiana as part of a parenting study before the pandemic. Because this research used nonprobability sampling, the results are not generalizable beyond the study population.