Demographic attributes, categories and number of participants

It seems likely that surveys with more participants would cover more demographic attributes and divide those attributes into more categories. Larger numbers of respondents would encourage more slicing and dicing. Or so I thought.

To test that hypothesis, I looked at 10 law-firm research surveys. My less-than-scientific method to pick them started with the last one alphabetically on my list by firm name and pored over the surveys in reverse order until I found 10 with usable data. The by-passed surveys either did not disclose their number of participants, gave very sketchy demographic information, or both (a few were in a series by the same firm). Having trapped the eligible surveys, I counted how many categories the report included for each of the four most common demographics — position of the respondent, revenue of the respondents’ companies, location of the companies, and industry (what I have called the “Big 4”).

The first of the two charts shows how the total number of categories in those four demographic attributes compares to the number of participants in the survey.Each red circle stands for one survey’s number of participants (on the bottom axis) and total Big 4 categories (on the left axis). The wavy blue line shows a non-linear trend line. Very non-linear, and not much of a pattern!

The second chart displays the same total of categories for the four most common demographics, plus the total number of categories in any other demographic attributes on the left axis. It reflects the same 10 surveys, so the bottom axis remains the same, but a greater range on the vertical axis because it includes counts from any other demographic attributes. The trend line here shows even less of a pattern than the squiggle of first plot!

Sigh. At least with this set of surveys, we can’t support a hypothesis that more participants means more demographic attributes. Perhaps if we broadened this particular inquiry to cover more surveys we might eventually distinguish a clearer relationship, but for the moment, none is apparent.

Leave a Reply

Your email address will not be published. Required fields are marked *