Convenience sampling – an easy, wrong, yet very used approach
Af: Jesper Ramsgaard
We live and breathe for data. Data collection and analysis. However boring this may sound, a lot of the interesting findings and one of guidance advice are based on larger studies. It may be hard to appreciate the value of data, but in some fields of business data is king.
The next time you read an article try to note the how the selection of test subjects (sampling) was performed. Usually most psychological studies are done with undergraduate students, who are obviously available like guinea pigs at the Universities. Convenience sampling is a term used to describe data collection using subjects who are “conveniently “ available. A true statistician will start shouting and throwing things at you if this is your main source of data - especially if you are trying to generalise across a wider population.
On the other hand experiments have to be pragmatic. If you pace your studies the way a rigid statistician and expert on experimental design would approach the topic, you would probably never get anything done. There will always be trade-offs and compromises, which of-course should be reflected in your conclusions.
The main study of the ExSl project involved performing one failed data collection (using convenience sampling, stupid but that’s how it goes) and one successful were 1000 respondents participated. Each subject delivered roughly 140 data points, a total of 140.000 data points. An impressive data set that will bring your excel sheet to its knees.
Based on the experience gained in the project, however easy it looks, convenience sampling should be avoided.