As part of our marketing course, we were expected to do a project regarding working women. To complete the same, we went to a mall and interviewed people we found there. But when we went back, our guide said we should do a ‘random sampling’. I didn’t quite understand what we had done wrong. After all, we had selected people at ‘random’…
Hmm… What rule did you use to select people?
Three of us in my group moved around the mall individually, and contacted people who seemed to be in the right age group. We looked for someone relaxed and approachable .
Well, your guide was right because that is not random sampling. That is actually a pretty arbitrary sampling. And though it is termed ‘random’, there’s nothing arbitrary about it.
Random sampling is, in fact, a very specific term that involves specific rules. And these rules are meant for two purposes — one, to make the sample as representative as possible, and two, to ensure that there is no strong bias for or against any sub-group.
But what kind of rules are we talking about? How could we have used it in our project?
Well, one rule you could have used, is to approach every 10 th lady who walks into the mall through one particular door. This is not a perfect rule, but it avoids any particular bias or skew.
And what if the 10 th lady refused?
Well, then approach the next person… and if she agrees, then go ahead and interview her, and wait for the 10 th person after her.
Ok,I see where you are going… some kind of a system or method to the way we selectpeople to approach…
Yes, that is right. In fact, it is probably better to use the term ‘probability sampling’. Bear in mind, however, that there is something called non-probability sampling too.
What’s the difference?
Simply put, probability sampling is used for quantitative research, while non-probability sampling is used for qualitative research. We can explain these terms at some future date, but basically, quantitative research has larger sample sizes, while qualitative research has small sample sizes and more depth of information.
Another difference is that probability sampling has more rules, while non-probability sampling has fewer rules. It even allows the selection of the respondent based on the researcher’s convenience.
So then what we did at the mall was a non-probability sampling?
Actually yes, and it is a specific type called convenience sampling. There are a couple of other types of non-probability sampling called ‘quota sampling’ and ‘judgment sampling’.
Are there different kinds of probability samplings as well?
There are different types; in fact, there are four types.
The first is called simple random sampling, and that is actually why I said earlier that it is better to use the term ‘probability sampling’. Because we need to be sure that we are talking about probability sampling as a whole, and not specifically about simple random sampling.
The second is called systematic sampling, and the other two are stratified sampling and cluster sampling.
I want to avoid going into too much detail about these methods, which is why I’m not explaining them all. But all four methods are equally important and, except for the first one, are used a lot in commercial market research.
Which of them is the best?
See, what is the best in accuracy may not necessarily be the best in commercial practice because sometimes, the price of being accurate is too high for the value of the information.
There is no definite answer to the question, even if we set aside the commercial angle. But in general, stratified sampling is more likely to give an accurate result than the others. And I am not going into the reasons for that here, because it is too technical.
One more question... If non-probability sampling has fewer rules, it is probably less accurate as well. Then why is it even used?
The field of qualitative research focuses more on exploring the consumer’s mind, motivations and concerns, than on findings which can be extrapolated to the population. The purpose is exploratory, not conclusive.
Great! Can you tell us more about qualitative research?
Sure, but some other time. And believe me, quantitative research is equally fascinating. There is often poetry in numbers…
To read more from the FundaMental section, click here .