Non-random sampling
Dataviz Team · Jean Russell
18 March 2021 · 2 min read
TABLE OF CONTENTS
Non-random Sampling
Note: This page is an option chapter of Statistical Modeling Part 2 - Sampling.
Non-random sampling (or non-probability sampling) refers to subjective sampling methods in which researchers draw samples according to his/her own convenience or subjective judgment. It does not strictly follow the principle of random sampling to draw samples so it cannot determine the sampling error, and cannot correctly explain to what extent the statistical value of the sample is suitable for the population.
The convenience sampling method refers to the way that researchers choose samples arbitrarily according to the convenience of researchers. For example, several pedestrians are selected for an interview at the intersection of the street. This method is the simplest, most cost-effective and time-saving method in non-probability sampling. However, if differences between individuals in the population found to be large, we also get large sampling errors.
Disadvantages: samples usually have large deviations and low credibility, and are not sufficiently representative.
Voluntary (or voluntary response) sampling, as the name suggests is a sampling method in which individuals volunteer themselves to be the sample. For example, online surveys.
Disadvantages: Some people are more willing to participate in the experiment/survey which leads to sampling bias.
Snowball sampling refers to randomly selecting some interviewees and conducting interviews with them, then asking them to recommend some other possible individuals that are part of the research target population. Thereafter, we select the individuals based on the recommendation and so on. Snowball sampling is often used for surveys of rare populations or distributed uneven populations.
Disadvantages: Sample could be limited to a group of people with similar thought and mindset, which will cause serious under-representation problems. This group of people is often just a smaller group in the subgroup that the researcher wants to study.