Wednesday, May 05, 2010

Back in Malta!

Greetings all! The very fact that I am writing from Malta means that things are either going really well, or really badly. Luckily for me, this time, things are going great! If you happen to page back through my blog posts from January, 2009, you'll read about great despair, stress, and failure. Well, we have learned from our mistakes and at least, five days into Malta our experiments are going wonderfully. The key: prepare *everything* before even thinking about boarding the plane. Stimulus creation is hard. Item list creation is hard. Scripting is hard. Programming software is hard. It all takes enormous amounts of time from numerous experts, so coordination is also key. About a month ago, when I realized how close this trip was getting, I abandoned my sabbatical relaxation and moved into high gear for experiment preparation. Our Maltese speaker had been in Tucson in August and recorded all of our items for us, so those were ready. Or so we thought; it turned out we still needed a number of nonwords in order to have the correct set of counterbalanced items. In a fit of efficiency, I built a head-mounted microphone and overnighted it, along with a high-quality digital recorder, to Malta, where our speaker was able to record those few remaining items and then send back the .wav files. And voila - we were ready for the next step: creating item lists. Keep in mind that we are using two different methodologies (each one a different type of priming) and testing two separate related-priming types in addition to identity and unrelated priming. The result: four separate experiments, each with three counterbalanced lists, for a total of twelve individual experiments to run. Yikes! Luckily, I am married to the best programmer on the planet, and he devoted untold hours to writing a perl script that takes as an input file a set of words and nonwords and creates prime-target pairs. These get output into a text file read by Praat, which takes the individually spliced items and sews each prime-target pair into a single sound file for our regular priming experiment. For the other priming methodology, we use a Praat script written by another one of the best programmers on the planet to create some pretty cool-sounding items; you should come participate in the experiment so you can hear 'em :)

Once all of the stimuli were created by Praat, it was time to work with e-prime, the software that runs each experiment. E-prime is interesting, and consistent in at least one way: you can always count on needing a day or two to get it working properly. In our case, the stumbling blocks turned out to be fairly simple, and I was able to trouble-shoot everything myself (which did make me feel pretty competent, I have to admit, especially since e-prime tech support isn't great about responding within 24 hours). For instance, the button box used by subjects to respond to each stimulus has to be connected to the computer and this requires an annoyingly large set of steps, such as installing the proper driver for the serial-to-USB adaptor (who has a serial port anyway these days, come on!) and setting the port setting correctly in each experiment (i.e., 12 times).

Then it came time to program each experiment - they all have the same basic structure, it's just that with the Latin square design each counterbalanced list has different stimuli (aside from the foils or throw-away/distractor items). Because the version of e-prime we're using doesn't recognize UTF-8 character encoding, and because the Maltese alphabet involves a few non-ascii characters (ġ, ċ, ż, and ħ) we use a special system where we substitute numbers for those characters in the names of our stimuli sound files. However, each experiment also involves a set of instructions that subjects read prior to beginning the experiment, and since we can't use plain text to write Maltese we have to first create a powerpoint slide, turn it into a .bmp file, and use that .bmp file in the experiment. Not a hugely demanding step, but an extra step nonetheless.

In any case, by the time I left Tucson at 6 am on April 21, I had twelve working experiments. I think there are cracks in the ceiling at home from where I hit my head jumping for joy. I felt pretty good spending four days burning down the prairie in Wisconsin (controlled burns are part of our prairie restoration project of 400 acres in southwest Wisconsin, and also a great way to let go of stress), and then spent two wonderful days in Chicago, where Andy and I met with two colleagues working on another joint project. When I got on the flight to Frankfurt on April 26, I was feeling tentatively good about everything.

Four days getting over jet lag were spent in Italy, and then on May 1 (may day, may day!) I took the train to Rome and boarded my Air Malta flight. On Monday, I arrived at the university, ready to tackle installing e-prime on two computers in two different labs. The tech support people here are amazing. They worked for two hours on Monday, and though we ran into some problem with the installation, they were pretty sure how to fix them, and eventually did. It's amazing how complicated installing drivers for things like a USB key can be, not to mention for the button box. We also discovered a very bizarre bug in one of our twelve experiments (thank goodness not more than that!) in which a dynamic link (to our image .bmp files) turned into static links with a pathname from my laptop that wouldn't go away. Unexplainable, but nonetheless the tech guys manually edited my experiment script directly and fixed the problem. They returned bright and early on Tuesday morning and finished the installation, and I copied over my experiment files. And held my breath while I started testing them far, they all seem to be working! In addition, hordes of students showed up to sign up as subjects, so the next two weeks are completely full, which is incredible; this isn't a place with a tradition of running experiments, but my colleagues here have done such a good job helping me recruit subjects.

Yesterday, I ran my first five subjects - I hadn't even planned to run anyone yet yesterday! But since I was ready, why not? Today, twelve more. Tomorrow is quite full too, etc. It's such a new and wonderful feeling for this to be working - it's almost as if I don't care what the results will be (but not quite). The only little snag as I've discovered is that getting 5-euro bills is challenging - they are quite scarce! I pay each subject 5 euros, and few of them can change a 10- or 20-euro bill.

In non-work news, I'm having a really lovely time so far. I'm getting to visit with good friends, and have also been able to attend an opera in Valletta one night and then a chamber recital in a church the next morning. I'm having dinner with friends in Valletta tonight, and attending another concert Saturday night. The weather is perfect, too - this is a great time of year to be here.

Who knows what the next four weeks will hold; hopefully more of the same. Perhaps I'll write about it...


Post a Comment

<< Home