A question that our clients often ask is "Should I commission quantitative or a qualitative research?"
William Trochim, in his excellent Knowledgebase, makes the point that, for two reasons, there's little difference between the approaches:
- All research is quantitative, because anything can be counted - even purely verbal responses, perhaps after sorting comments into similar groups.
- All research is qualitative, because answers to even the firmest numeric questions may conceal a variety of meanings.
So the real difference between qualitative and quantitative is not so much the method, but the researcher's approach.
Despite that, some methods are more quantitative (e.g. surveys, automated counting), and other methods are more qualitative (e.g. in-depth interviews and gruop discussions). However there can be qualitative surveys - using mostly open-ended questions. There can also be quantitative group discussions (using the consensus group technique).
Our advice: choose a more quantitative method when most of the following conditions apply:
- The research is confirmatory rather than exploratory i.e. this is a frequently researched topic, and (numerical) data from earlier research is available..
- You are trying to measure a trend (almost impossible with qualitative research).
- There is no ambiguity about the concepts being measured, and only one way to measure each concept.
- The concept is being measured on a ratio or ordinal scale.
And choose a qualitative method when most of these conditions apply:
- You have no existing research data on this topic.
- The most appropriate unit of measurement is not certain (Individuals? Households? Organizations?)
- The concept is assessed on a nominal scale, with no clear demarcation points.
- You are exploring the reasons why people do or believe something.
Two extreme examples:
You are studying the trends in weather in the town where you live. There aren't many variables: temperature ranges, wind speed, rainfall, barometric pressure, and perhaps a few others. Most of the variables are measured mechanically, and a lot of historical data exists. You wouldn't even consider doing qualitative research on this.
You have invented an entirely new type of product that has a mix of positive and negative environment effects, which would be sold to large organizations, but not directly to the public. However public feelings about the product might be strong enough that organizations would be deterred from buying it. The product is very complex, and takes a long time to explain. You find it difficult to pose a clear question, but it would be something like "To what extent might public resistance deter organizations from buying this product?" Though you could do a survey and get a result, it would be very misleading. This problem has so many unknowns that it clearly calls for qualitative research.
In reality, most topics of research fall between the two extremes. Our position, in general, is that it's best to take both approaches. Quantitative data enumerates, and qualitative data explains.
When a budget is large enough, it's useful to alternate between qualitative and quantitative research. A good general approach is:
1. Begin with a set of consensus groups, to see what's on the minds of respondents in the topic area.
2. Turn the statements from the consensus groups into a set of attitude questions, and include these in a survey.
3. Interpret the survey results with a series of 20 or 30 in-depth interviews, to gain an understanding of the issues.
4. This may throw up some new questions, which can be asked in a follow-up survey.
After several cycles, alternating between qualitative and quantitative research, you'll have a thorough understanding of the topic.
But there's often no need to do several pieces of research, because (as Trochim points out) almost every research method can be either qualitative or quantitative. A survey can include a lot of open-ended questions, a series of in-depth interviews can use a random sample and collect hard data, and consensus groups (drawn from a random sample) will always collect both types of data.