When we work with our data we have to decide who and what gets to be in the story. A number goes in the table or it doesn’t. A quote earns half a page, a passing mention, or nothing at all. The participant who said the awkward, off-message thing gets wrapped in a sentence that opens “some respondents felt.” We usually make these calls fast, and often without noticing we are making them. But when we call the results of all these calls “the analysis” we gloss over all the tiny decisions we made about what and who goes where and what and who are important.

This playful exercise slows down that decision-making. Think of it as passing the mouse to the other side.

Pick one item from a dataset you are working on right now. Not the whole set, just one thing. Uno. Singular. It might be a favourite quote, you know the one you keep trying to find a home for. It might be the outlier you are about to round away or the participant who dropped out halfway. Perhaps it is the average that wants to stand for everyone or the contradictory case that wrecks your tidy result. Maybe it is the pilot interview you are tempted to bury. Choose the one that is giving you most trouble right now.

Now write a letter to you, the researcher, in its voice. The data is directly addressing you, the about -to writer, and it has come to argue a case. Yes it’s a bit silly. But stay with me. 

Most of your characters will want to be written. A minority will want the opposite and will beg you to leave them out of it, or at least to stop making them mean what they do not mean. Let them be a bit unreasonable. They have been sitting in a spreadsheet or transcript for a long time and they have opinions.

To get them talking, you might have them tell you:

• why they think their story has to be told or has to be kept quiet
• how they have decided to approach you – with humour, charm, threat, guilt, a reasonable memo, open weeping
• who they think they are and how that squares with the job you have given them in your argument
• what they are willing to say or do to get their way
• which bits of themselves they think are the good bits, the parts worth putting on the page.

Give this exercise ten to fifteen minutes and a page. Write fast and write badly. A polished text is not the goal. You are trying to hear the argument you have been having with yourself, out loud, from the other side.

There may be a few characters you’ll meet as you write. The outlier may arrive first, certain it is the most interesting thing you have generated and appalled that you are about to demote it to a footnote. Right behind comes the mean, smooth and reasonable, faintly embarrassed by the outlier, asking only to represent everybody. The long quote you love turns up nervous, because it has worked out that you are about to make it carry a claim it never agreed to. The missing case –  the non-response, the person who left – wants you to understand that an absence is also a result and writing them out a second time is a poor choice. The number you rounded would love its decimal places back.

When you have produced a page, don’t read it as the author. Make yourself the reader. Notice which arguments moved you. The outlier that genuinely earned its keep is a different proposition from the outlier you find interesting and want to keep for the wrong reasons. The participant whose plea to be left alone you find yourself overruling is telling you something about the ethics of the chapter, not only its prose.

Whatever persuaded you on the page is a clue to the editorial values you use when nobody is watching. That‘s worth knowing. And you may also have gained new insights into analysis and what you will write. 

This is of course just a game. But there are real data decisions at stake. When we write with data we decide whose story this, then write the text as if the decision made itself. This game alerts us to the power we have when we write. 

Photo by Carlos Muza on Unsplash