From the Field: Managing Data 101
Over the last two weeks, as part of a program evaluation, I’ve been doing something that sounds relatively simple: data entry. Our individual client records are kept in hard copy form only, and so in order to meaningfully look at how our program has impacted HIV transmission in Malawi, the first step has been to put the data into an electronic version. The steps involved seemed simple enough: 1) collect almost client registers from 50 sites throughout the country, 2) randomly sample tens of clients from each site’s register, 3) enter the data into the electronic database. The reality is that I now have a much deeper appreciation for this part of data management!
Step 1: collecting the documentation of the data. Because the client registers are used on a daily basis, it’s impossible to remove them from individual sites for a prolonged period of time. We asked locally-based staff to make photocopies and send them to our headquarters, which proved more challenging than I anticipated. We then dealt with the additional complication of illegible copies, not knowing which pages correspond with one another, and the list goes on! We started our project two weeks ahead of schedule, and it seems we’re now falling two weeks behind….
Step 2: random sampling. Well, this is a fairly complex process that computers have made super easy. Statistical calculations help to determine how many samples you need based on total clients and the level of “power” that you want. Software also helps to generate random numbers, which were necessary for the sampling. I won’t bore you with details but this was the easiest part of the process!
Step 3: enter the data. Now the fun begins! It seems straightforward: identify all 100 or so clients required for each site’s sample (using the random numbers) and plug their data into the database. But when you’re entering 30 pieces of information for hundreds of clients, it starts to become tedious! Then there’s the challenge of reading our data recorders’ handwriting, understanding conflicting information that’s documented, and – of course – making sense of the non-English comments that sometimes appear alongside a client. The other day, I spent more than an hour on the phone struggling to communicate with health workers at a few sites to get clarification on some of their data!
So after two weeks of data entry, my hands are hurting, I have a crick on one side of my neck, and I can say with full certainty that I won’t be sad when when this part of the process is over. When I was working on Masters thesis, I thought that the quantitative data analysis component was so difficult, trying to piece together two years and seven quantitative classes worth of information as I analyzed Demographic and Health Services data. But today, after two weeks (and counting) of attempting to enter data to the highest quality in hopes of improving our impact, I have a much deeper appreciation for data entry. I have enjoyed mentoring our assistant data entry clerk and being able to take ownership of this project from the very early stages, but I’m confident that data entry isn’t where I see myself in the future. The next time I read an article with the latest statistics on HIV or the hot health topic of the minute, and the next time I perform any type of statistical analysis, I will have a much greater appreciation for the work that went into just compiling meaningful data!