I proposed an additional dissertation experiment and am amid data collection for it. My research assistant Deniz and I have consented 40 subjects. Two subject numbers were unable to complete the study in the allotted time (#13 and #14). We re-ran a new #14, but still need to re-run a #13. Somehow we also skipped #10. Thus, when data collection resumes tomorrow, we'll do 10, 13, and 41. The reason subject numbers are important in my study is that they're tied in the programming to the counterbalancing groups and experimental conditions. Thus, if one number is no good, we just throw it out and get another one.
Of course, the reason to go in and look at the data mid-data collection is just to ensure that everyone's completing the study as expected and that there are no bugs or glitches in the program. This process alerted me to one potential issue -- currently, if someone doesn't make a source decision (i.e., they selected NEW on the source test), I grade their source accuracy as INCORRECT. Obviously this is wrong -- when computing overall source accuracy, we should only look at OLD responses.
This has the potential to get pretty sticky and confusing as things progress, so I'll have to think carefully about these issues while working on the writeup.