looking before leaping (or running statistical analyses)

It’s the job that’s never started as takes longest to finish. Intent on completing the data summary long before the deadline, I mustered up some willpower and set to work. By the following day, I had color coded graphs, supportive statistical data, and captions. But something didn’t seem quite right. Had I analyzed the data correctly?

My questions were answered on Wednesday during office hours. It became evident that my hurry to finish this paper had led to errors through including particular data sets; I’d gotten a little ahead of myself. But no worries, I told myself. No problem. I’d just delete this data cluster, compare these two columns, and I’d be done. So I started (again).

It was only after relabeling the graphs and rerunning the statistical analyses at around 3 a.m. that I realized another thing I’d left out. The negative controls. I’d forgotten the negative controls.

So to say the least, I spent the majority of the later part of the week reworking graphs, rewriting captions, mulling over possible explanations for the results- whatever time wasn't spent on this was dedicated towards imbibing coffee.

My most major problem was a lack of forethought (how much of this was due to sleep deprivation- of that, I'm not certain); imprecise planning led to more work than necessary. It would have been much better to plan out what I intended to convey to readers in this study prior to processing data.

In the next module, I intend to follow a more organized method for tackling data summaries. What groups should I be comparing to each other? What kinds of data I want to include for the experiment to be more understandable? I think these aspects are a little clearer now, and I’m looking forward to making these changes in the next module.