Jun 122013
 

Earlier this month, NewMR held its first Explode-A-Myth session (find the recordings by clicking here) and my contribution was a discussion why there is no method that is a melange of qual and quant, because the underlying paradigms are different.

Through the Q&A session at that event, and in particular a question from Betsy Leichliter, I gained a clearer understanding of the core difference between qual and quant. Betsy asked “So should the ‘qual’ or ‘quant’ labels be driven by the method of analysis, not necessarily the method of “data collection”?”. I think this question from Betsy is the best answer to the question about what is the difference between qual and quant I have seen.

Within reason, any data can be assessed quantitatively or qualitatively. Of course, there are some limits to both approaches. A very small amount of data is likely to produce findings that are hard to generalise. We can count the sales of brand X, in one store, on one day, but it is hard to draw any inferences about the world from that. Similarly, ten-thousand open-ended responses could only be assessed qualitatively with a large team, or a large amount of time.

The quantitative approach is based on an assumption that there is a ‘real’ world, which we can measure objectively (or, at least, that we can get fairly close to that ideal). The underlying beliefs are a) it is the method that provides the results (different researchers should provide the same answer if they use the same method on the same data), and b) that the researcher is discovering and reporting something that exists.

The qualitative approach, as it has developed over the past thirty years, is based (for most researchers) on a constructionist paradigm (there are several different models, but they all tend to be constructionist). The researcher does not discover truths, the researcher creates a narrative that provides useful insight into what is happening. The researcher is part of the analysis, different researchers will provide different narratives, and the value of the narrative depends on the ability of the researcher to observe what is happening, to synthesise an analysis, and to create a narrative that conveys something useful to the end client.

The key difference between quant and qual is the difference between discovering and creating, overlaid with ritual of using numbers for quant and words for qual.

Jan 072013
 

Market research is being deluged with new sources of data, from social media, from electronic communications, and from research communities. Whilst some of this information is suitable for quantitative analysis, large parts of it are unstructured, for example tweets, posts, comments, and uploads. Whilst this data presents an interesting opportunity for market research, it presents a sequence of inter-connected problems and challenges:

  1. Many of the market researchers who are most proficient with unstructured data, the qualitative researchers, are not instinctively drawn towards online data, preferring to deal with people in a face-to-face environment.
  2. Many of the researchers most attracted to large amounts of online data, the hard core quantitative researchers, have little appreciation of the different epistemologies of quantitative and qualitative research.
  3. Many of the software vendors, perhaps in a rush to market, have released products before they were really ready and with massive over-claims.

In order for market research to fully leverage the potential benefits of the discourses being generated, market research needs to address the qualitative deficit. The qualitative deficit is the shortage of talent, software, and approaches designed to utilise massive amounts of qualitative data.

Some of the new approaches and skillsets that are needed relate to the process of quantifying messages within discourses – Google’s Flu Trends being a straightforward example (where Google use phrases that indicate people are searching for flu remedies to quantify the incidence of flu). This use of unstructured data will be of value in areas such as brand and ad tracking.

However, the bigger deficit relates to taking qualitative information and extracting qualitative findings. In online discussions the meaning is often unrelated to the frequency of words and phrases, the meaning rests in the structure of the conversation and the outcomes of the conversation. Feedback for product design, the identification of opportunities, core reasons for product dissatisfaction are likely to be found in the meaning of discourses, as opposed to counts of terms and phrases.

Whilst part of the answer will be new software, my feeling is that research urgently needs to expand the number of researchers with a good understanding of qualitative methods and epistemology.