What Happens When I Write a 'Report'

I have been writing a report for a quantitative study as part of my day job at Willow Research of Chicago, Illinois. 

All told, the report will have fewer than 1000 words (about what’s in this blog post), and run around 25 PowerPoint slides, not including an appendix. 

Preparing a draft has taken the better part of my workday for a week’s time.

Thinking about what I’ve been doing and how I’ve been doing it reinforces my belief that for students to learn how to write we must refrain from using prescriptive methods like the five-paragraph essay. This work also makes me grateful for my flexible, adaptable liberal arts education.[1]

To explain how it can take a week to write 1000 words I should walk through what writing a quantitative report for a market research firm entails.

The writing starts with the “deck,” a file of tables derived from respondent answers to our survey. In reality, the writing starts with the client objectives and research proposal (the questions that the report will seek to answer), but let’s ignore that for now.

The deck for this project runs 85 or so pages over 48 tables. In my experience, this is on the low end of average. Each table reflects a single question that is displayed at the top, with the “banner” below. The banner is a series of columns that reflect categories that we believe might be relevant to the analysis. This always includes the total, but covers other things as well, such as age, gender, race, etc. 

The rows are the response options that could be chosen for the question. A yes/no question will have two rows. A “check all that apply” or rating question may run many rows long.

Suffice to say, when you first see a data run, it can be difficult to find any kind of pattern, to find meaning inside the data. A single page of a 48 table deck with a 16 column banner will have dozens of data points, every single one of which potentially holds meaning.

Anyone who says that “the data speaks for itself” has never spent any time dealing with data.

It is tempting – and this is how I did things early in my career – to start just summarizing each table individually, finding some interesting thing on each page, writing a bullet point or two around it and then pasting the table below. Composition instructors will recognize this technique in student papers that will cite one source per paragraph with little connection between paragraphs and limited coherence as a whole. When I still used traditional grading and more prescriptive methods, I would often receive this kind of paper, where the individual moves inside of the paragraphs seemed generally okay, but the overall effect of the piece was muted at best. Students got a B if the paper was mechanically clean, and a B- or less if the paper had issues. I don’t think students learned much of consequence from producing them.

Since Willow is a research company that prides itself on delivering actionable information to clients’ questions, a B- isn’t going to cut it, so instead, I look through the deck and start noticing things. 

I note where there are interesting differences between groups, seeking out patterns in the data. I note where some finding surprises me. I note lots and lots of things. Next, I try to tie some of the things I noticed together into inferences, conclusions I can draw from putting multiple noticings together. Here our research objectives come in handy, as I see how my inferences help answer the questions underneath the original objectives.

This process takes a couple of days, longer if there’s anomalies in the deck that suggest the data need additional cleaning.[2] I often check in with colleagues at this stage, float some of my inferences by them as a gut check. If I’m drafting the report, I’m more immersed than anyone else for the moment, but others have had a hand in the objectives and questionnaire, and will have spent some time looking at the data.

Now I “write.” There is obviously a big picture structure in that we start with objectives and methodology before the “findings,” but within the findings, I don’t really know how it’s going to be structured until I get into the thick of it.

We have a snazzy template that makes the organization and presentation of the information graphically engaging, but now it’s on me to find the “story” of the data that will convey the findings to the client with maximum impact. This involves pulling data from tables that may be far apart in the deck, but belong together in the narrative. If you were to observe me during this period you would see lots of my moving from the screen to the data tables and back and me asking myself, “Does that make sense?”[3]

I am thinking. 

I get a slide down, then another and another. Doing the fourth slide makes me realize I need to add something to the second slide, and maybe the third slide should go first come to think of it. 

It is a non-linear process, as the thinking I’m doing at any given moment may send me back to a previous part of the report to revise, extend, or even delete, but over time, the slides accumulate and progress happens.

The final part of the draft is the executive summary with the key findings, implications, and recommendations. Think of it as after having written the story, I can now go back and fully explain what it all means. 

This draft is only the beginning of the process. Once completed the rest of the team will dive in, fixing formatting, adding in graphical elements to enhance presentation, checking my assumptions, adding more insights, double-checking my numbers against the original table. If I had to guess the first draft represents 80-85% of the final version, but the last 15-20% is vital, not the difference between a B- and an A so much as the difference between not passing and an A.[4]

Without being experienced in thinking about writing through the lens of audience and purpose, I could never do this work. Thankfully, because I have been well trained in this area, I find that I can write reports about any subject.

All I need is data and sufficient time to think.

--