Tuesday, November 22, 2011

And multiple analyses later. . .

When the Matrix becomes real, I will find the SPSS program within its intricate pathways. I will then proceed to initiate mortal combat with this program; not that crappy real-life Dragon Ball Z nonsense from Matrix 3, but the good ole' fashioned fisticuffs from the first movie. Overcoats, sunglasses, hilarious numbers of guns, the works.

And I shall dance upon it's pixelated corpse with giddy abandon.

Now, this might seem a bit strong, and indeed it is. I have been doing outlier data cleaning for 3 data sets for the past few days, and the problem is that in the literature that we're working off, everybody has their own method of dealing with outliers. Some researchers ignore them entirely, others delete them, others replace them with a 2.5 standard deviation cutoff. Even worse, some determine outliers within-participants, within-conditions or sometimes both. Owing to the high likelihood that a fellow psycholinguist will review our paper come submission, we decided it was a good idea to cover all the bases in case they ask why we didn't manage outliers in (x) fashion.

Cue 13 different outlier cleaning processes, over 3 data sets. Every possible relevant permutation of the process to be performed on each dataset.

Predictably, I have reached the point where I no longer believe that I and my beloved Apple Mac are separate beings. I think we've become some sort of Deus Ex Machina, albeit in the narrow field of number crunching. I think it simply is aware of my intentions and reacts accordingly; I could probably use two chopsticks instead of a mouse and keyboard and still create the desired SPSS commands. I may eventually nickname it 'Skynet' and welcome the end of humanity.

I am of course inflating the issue for the sake of comedy and a good read. The work is not hard, nor is it particularly urgent. I've set myself a deadline for the end of the week, mainly in order to give myself a semblance of motivation beyond that transmitted through caffeine-based products. It is a reminder that not all work in a PhD is gloriously interesting, mind-expanding stuff. Occasionally you do have to engage in exercises of pure tedium with the promise of a reward at the end, which in my case is a conference I hope to go to next year overseas, as well as the publication that will come out of this work.

My new master/Sith Lord/computer program calls to me to resume crafting the syntax that will deal with analysis no. 3. If anyone hears 'EXTERMINATE' or sounds of death rays coming from my office, worry not - it's all been ethically approved. See you next week!


No comments:

Post a Comment