I had a nerve wracking, gut wrenching morning yesterday as I pored over dissertation data that was messed up – it just didn’t make sense. Some scales were beautifully normally distributed but completely unreliable, others were showing me a mirror image, backwards effect from what past literature told me to expect. Something was WRONG. I contacted the techy guy I have running my survey program for me in New York (I’m a Canadian collecting American data), and asked him to send me the absolute raw data – no fancy scripts used in the export file, not even changed from the raw labels (e.g., “strongly agree” before it gets changed to a computable value of “5”). He got me the file by 7pm, but I was putting Avery to bed and wasn’t out of her room until 9 (poor little bug has been taking a long time to fall asleep lately, writhing around itching her eczema flare up). I worked on that file until after midnight, and started up again as soon as I’d dropped Avery off at daycare this morning. By 10am, I had solved the mystery through careful detective work and meticulously doing by hand all coding and reverse scoring. The data was fine! The messed up appearance was caused by an error made by my techy survey software guy when he wrote script to code and reverse score before exporting to me.
That felt good.
What felt even better? The data was better than fine. It was AMAZING. I’m collecting my data in batches so I can ensure it’s all going to plan before spending ALL THE MONEY on the full sample; in the mere 79 person sample I have right now, I was actually getting significant effects on two of my hypotheses. For anyone who knows stats, you know how AWESOME this is. Not to mention the validation I feel in the face of two of my committee members who approved my proposal despite admitting their utter skepticism that I’d find any effects. So now I’m in full-steam-ahead recruiting mode for batch two, and if everything still looks good when I look at that, I’ll be able to finish data collection in 6 weeks. I mean, my original goal was to start collecting in September and get it all in 2 weeks, but we all hit bumps in the road when dissertating. It ALWAYS takes longer than we expect.
But I least my hypotheses are being supported. Go me!