Sampling methods

Probability

  • Random sampling & systematic sampling (every Nth person) ⇒ equal probability of selection
  • Sampling proportional to size (PPS) – concentrates on the largest segments of the population
  • Stratified sampling (members of each stratum (a sub-population) share some characteristic)
  • Advantage: can calculate sampling error

Nonprobability

  • Accidental, Haphazard, convenience sampling ⇒ these might not be representative of the target population
  • Purposeful – sampling with a purpose in mind
    • Modal instance sampling – focused on ‘typical’ case
    • Expert sampling – choosing experts for your samples
    • Quota sampling - proportional vs. non-proportional
    • Heterogeneity sampling – to achieve diversity in samples
    • Snowball sampling – get recommendations of others to sample, from your samples

For further details of Nonprobability sampling see: [Trochim2006] http://www.socialresearchmethods.net/kb/sampnon.php Links to an external site.

[Trochim2006] William M.K. Trochim, ‘Nonprobability Sampling’, Research Methods Knowledge Base, 20-Oct-2006. [Online]. Available: http://www.socialresearchmethods.net/kb/sampnon.php Links to an external site. . [Accessed: 03-Aug-2015]

Transcript

Now, not surprisingly, there many different methods of sampling. We talked about some of them earlier, but, typically, we can split them into the following categories: probabilistic methods - such as random sampling or systematic sampling - or very commonly sampling proportional to size.  You take that population, we say, "Hmm!  We stratified it into different segments, and this segment is twice the size of that - maybe we only need to ask half the number of people in this population as of that to balance- probabilities across the different segments?"  Or we might need to do stratified sampling to ensure that we actually have some samples from every subpopulation that we're interested in.  Conversely, we have non-probabilistic methods of sampling, such as accidental or haphazard or convenience sampling. But here we run the risk that that may not be representative of our target population.  So, in the past, for instance, students have asked other students questions - forgetting that the actual population for the study that they were trying to do wasn't about other students and so, therefore, the data that they got was really not the data relevant for the target population. Oops!  (yes) wrong data!  Irrelevant results - perhaps.  We can also think about purposeful sampling like modal instance sampling - what's the most typical case. Or expert sampling, we choose just our experts -- because we believe we will get the most information out of them in the shortest period of time. We can think of quota sampling - proportional or non-proportional sampling.  Or heterogeneous sampling - we want the greatest diversity possible. Or one of the most common methods is snowball sampling - you get recommendations from the people who you had do your questionnaire, and they say, "Oh yeah! One of my friends would be really interested in participating in your study also" - and they give a copy to [them], and it spreads out that way in a social network kind of fashion.