Causes trump statistics (or why quant needs qual)

I've borrowed part of the title of this blog from a chapter in Daniel Kahneman’s excellent book “Thinking Fast and Slow” which is, in my view, essential reading for every market researcher wanting to understand more about how and why we make the choices we do.

The chapter in the book explains how, in general, people’s understanding of the world is not changed by statistics. In contrast, experience of individual cases can impact our attitudes and beliefs.

To illustrate this Kahneman writes about the “helping experiment”. Essentially, a mock experiment was set up which required students to sit in individual booths. All had a microphone and talked about the problems they were facing in life to the others. However, one of the participants was a stooge and pretended to have a seizure and asked for help. Only 4 of the 15 participants responded immediately to his call for help, a further 5 came out of their booths well after the “victim” choked and 6 never came out of their booths at all. This experiment demonstrated that individuals feel relieved of responsibility when they know that others have received the same request for help.

At a later date, two separate groups of students were told about this study but one group were not told the results. Both groups were then shown short, bland videos of two people who had actually participated in the experiment. The interviewees appeared to be nice, normal, decent people. The group who were not told about the results thought that both would rush to the victim’s aid. The other group, despite knowing the results of the experiment, also said that both would help the victim despite knowing that only 4 out of 15 participants actually did so. Therefore, the statistics hadn’t changed their way of thinking.

However, a new group of students were then taken and told the procedure of the helping experiment but they were not told the results. They were shown the two videos and were told that the individuals had not rushed to the victim’s aid. They were then asked to guess how many people had actually helped and their guesses were surprisingly accurate. The psychologists who carried out this experiment, Nisbett and Borgida, neatly summarised the findings as follows:

Subjects’ unwillingness to deduce the particular from the general was matched only by their willingness to infer the general from the particular.

As a quantitative market researcher this is concerning. By definition we are producing amalgamated data (the general rather than the particular) to help inform business decisions, but, however surprising or impressive the statistics we produce, clients’ views and perceptions of their world are unlikely to change. It is only by relating this data to stories about individuals that we can truly impact beliefs.

Quant alone can therefore be a blunt instrument and qual trumps it in terms of being able provide those individual stories that can change attitudes. Involving clients in qual gives them direct access to the particular in a way that really hits home. However, care needs to be taken to make sure clients aren’t being affected by the atypical and being steered down blind alleys.

As ever, quant and qual together are a powerful combination. Quant can provide robust evidence but qual can provide the stories and the theatre that illustrate and allow the findings to become entrenched. The next time you’re commissioning or proposing for a quant project consider added a little qual to the mix.

Submit comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Recent Posts