Tom Webster, writing and speaking

Tinderbox and Questionnaire Design

Added on by Tom Webster.

I've written before in my personal blog about Tinderbox (Mac only) and how useful I find it for everything from project tracking to day-to-day note taking. I also use it heavily for qualitative research projects (which I promise I will elucidate in this space at a later date). Lately I've discovered another use that makes Tinderbox a very handy tool for quantitative researchers as well.

Some quantitative researchers prefer to write their questionnaire drafts in a word processor, others directly in a particular research tool (the actual survey administration tool, for instance). The former is not well-suited for the mechanics of a survey instrument, while the latter are generally not particularly fertile environments for brainstorming and ideation.

Tinderbox has become my tool of choice here because it does two things, very simply: it's an ace outliner, and it allows me to make links between non-adjacent outline elements. I like to write questionnaires in outliners, since they enable you to get a list out of your head as fast as possible, and allow you to easily move the list elements around to give the questionnaire a logical flow after you've dumped your brain. Survey questionnaires are not like other writing projects, however, in that they are, in a sense, "programs." Questions often depend upon the answers to previous questions, and some questions may or may not be skipped depending on prior responses. If you move a question towards the beginning of a survey that depends upon a response to another question, you need to be able to see what that dependency is so that you don't "break" the survey and can move the related questions correspondingly. You also need to check the logic of a survey instrument to be sure that you aren't skipping persons on a given question that should actually get the question, and that you aren't assuming "facts not in evidence" by asking a question before establishing a critical bit of information.

That's why I've really taken a shine to writing surveys in Tinderbox. At first, I use it just like any outliner, getting ideas in a heirarchical form. As I write questions that will cause skips, dependencies or conditional answers, however, with Tinderbox I can simply draw a link from one note to another, label that link, and forget about it. Later, when I want to start ordering the elements, a simple map view lets me quickly see what is linked to what, so that I don't break anything. Really, if you learned to program a computer in Basic (as I and many others did back in school) then you understand that a survey questionnaire is really composed of an assortment of GoSub commands and subroutines, all of which have to end logically and return to the main path of the survey. I write in outline mode, but start playing with and reordering survey elements in map view, so I can clearly see the "subroutines" and be sure that the questionnaire order is economical, logical, and requires the lowest cognitive load on the part of the respondent as possible.

With survey research tools increasingly becoming platform independent, there are lots of possibilities for Mac users in the field, and I highly recommend Tinderbox for a variety of research purposes. It can be complex, but doesn't need to be, really--I just open it up and dump text in with the default settings and then worry about order and presentation later. With an outliner I am constricted to a hierarchical view, but with a mind-mapper, I lose the logical order that a survey instrument eventually needs to have. With Tinderbox I can shift from chaos to order, depending on the thinking mode I need to utilize, and the tool shifts with me.

tinderbox-questionnaire_design.png