Yesterday in the Research Bunker, I posed this question to our two analysts: “In your opinion, has convenience sampling overtaken random sampling as the preferred method of sampling in quantitative research?” Working in market research and on surveys daily, sampling methodologies become a key piece to answering our insights puzzles for our clients. Depending on which sampling route you take, there will always be pros and cons to weigh. Such considerations as time versus cost, accuracy versus speed, etc.
Our two market research analysts at RMS offer their take below on the question of convenience sampling versus random sampling:
- Vance Marriner, Senior Research Analyst at RMS – The majority of the clients I’ve worked with over the years have been fairly new to market research and/or have tended to rely on our judgment for the best approaches to methodology and sampling. If anything, when the question of sampling has come up, it’s more often the case where the client has suggested a convenience sampling approach and we have convinced them that a random sampling approach was more appropriate for their needs (or vice-versa). That said, my own opinion of convenience sampling has evolved a bit over the years, and I’m more open to it in some circumstances than I used to be in the past. Mostly, that evolution of opinion has been due to a shift in the realities of the research landscape (e.g., erosion of landlines) that have made random samples more difficult and more expensive to obtain. That, combined with the growing trend of clients valuing faster turnaround for research results, has led to a situation where the choice at hand isn’t so much convenience sampling versus random sampling, but convenience sampling versus simply not doing research at all. In those cases, I think it’s okay to work with convenience samples, but it’s essential to make sure that everybody involved understands the limitations and implications of the approach and adjusts their expectations accordingly. Basically, the trade-off is that research done with convenience samples gives you “fuzzier” answers to your research questions, but sometimes that’s preferable to not having no answers at all.
- Chris Coville, Senior Research Associate at RMS – While the selection of a sampling method is most certainly a case to case basis for each project, I do believe non-probabilistic and convenience sampling methods have grown in popularity. Depending on the project, random sampling can be difficult, costly, and a lengthy while convenience sampling can be inexpensive and provide a quick turn-around. When looking to convenience sampling methods, it’s important to remember that not all methods are equal with regards to bias and error. If you do use a convenience method, then you need to properly evaluate the data collection process and try to cut out any issues in an attempt to keep the sample as “random” as possible (e.g., if you’re conducting intercept interviews at the mall or on a college campus, don’t conduct them all in the same exact location). Convenience samples can be a great option when you’re conducting base level exploratory research or if you have multiple research components to backup your findings. Most importantly, when you’re reviewing the research results, you need remember the fact that you used a non-probabilistic sampling method. Keep that in mind before making any decisions based off of the results (e.g., if you’re using a referral sample, homophily is going to be a major concern and you can expect the respondents to have similar traits and preferences).
Research & Marketing Strategies (RMS) is a market research firm in Syracuse, NY. For more information about Vance, Chris, and the authors of the RMS Bunker Blog click here. If you are interested in consulting with RMS for your market research needs please contact our Director of Business Development Sandy Baker at SandyB@RMSresults.com or by calling 1-866-567-5422.
I think most people would agree that a random sample, even better a random probability sample, is the best way to go. But how often are those achievable in market research? Sadly, almost never.
[…] the era when almost everyone in the U.S. had landlines. For a variety of reasons, market research clients have begun to care less about considerations like statistical reliability and more about using methodologies that can give […]