SPSS / PASW Statistics 18 for Mac: The Same Junk as Always

So I write this rant about the horrors of SPSS 16 for Mac. As a result, SPSS Inc. invite me into their beta program for SPSS 17 and offer a free license for it in return. Fair enough. I participate in the program, do all their required testing and submit plenty of bugs. The time window for their beta testers is four weeks. Four weeks. I find that way too short but thats how they operate.

After the end of the beta testing, I never received my free license. I sent them one e-mail about it that was never answered. So when they invited me into their SPSS (which is now called PASW: Predictive Analysis Software) 18 beta, I didn't do anything.

So now PASW 18 ships and I install it with a screen shot app at hand. The first thing I see is this readme window (underlining added by me):
Wow. They still cannot handle foreign characters. I complained about that two versions ago. A company that is producing a statistics software is unable to display a text file containing ä, ö, and ü correctly.

After the install finishes, PASW 18 Mac launches with this gem (you will have to click on it for the large version to see what I mean):

Look at the shadows of "PASW" in the top blue bar and the red line next to it. They are pixelated! These images were obviously produced for a smaller resolution but were magnified to the current size, producing this highly unprofessional experience. So the first two impressions I get give me a feeling that not much care has gone into the production of this software. Attention to detail: Nope, sorry.

The data window features new icons at the top as one can see on the following image.
Unfortunately, that was the only time I ever saw it. After the install, I quit SPSS PASW. Every time I have launched it since them (including after reboots), SPSS PASW crashes on start up:
It. Crashes. On. Every. Launch. On my MacBook Pro (2.33 GHz, 4GB RAM, Mac OS X 10.5.8), SPSS PASW 18 is inoperable. What an enormous piece of junk.

Virtual Reality Social Psychology

Today I had the pleasure of giving a talk at the Institute of Work and Organizational Psychology at the University of Neuchâtel (Institute de Psychologie du Travail et des Organisations (IPTO), Université de Neuchâtel). See here. I was invited by Prof. Marianne Schmid Mast and had the opportunity to see her laboratory and meet her cool team before my talk. In Neuchâtel, they do something I had never seen before: Virtual reality social psychology.

Social psychology studies, among other things, the interactions between individuals. Which is a challenge for laboratory research, because it adds a large amount of complexity (and costs) to experiments. For example, if a researcher wants to study the effects of group diversity on group performance, each observation requires not a single participant, but an entire group of participants. If a researcher wants to study a behavior that is evoked by a specific social interaction, it gets even more complicated: Suppose a researcher thinks that individuals will share less information with an incompetent supervisor than with a competent supervisor. An according experiment would require a supervisor who is either competent or incompetent in every experiment. Also, this supervisor should possibly exhibit the same interactions towards all experimental participants in one condition. One way of doing this is by hiring an actor for the role of the supervisor, which is often not feasible.

At Prof. Schmid Mast's Lab, they do such studies in a different way. They use virtual reality; a 3D virtual immersive environment. The experiment participant wears a head-mounted display (HMD) that gives him or her her the impression of being in another world. In this virtual world, one can interact with avatars, virtual representations of individuals. Those are programmed by the experimenters in such a way that they exhibit a certain interpersonal behavior, which is of course always constant and fully controllable by the experimenters. At the same time, the system logs data that is difficult to acquire in normal laboratory settings, such as interpersonal spatial distance between the participant and the avatars. In combination with verbal codings of the participant, one gets an extremely rich and reliable source for social interaction data.

Below is a picture of me trying it out. There is a diagram showing how the system works in the background: Four cameras pick up the infrared signal that the blue LED at the back of the HMD emits. A highly sensitive motion sensor (the little blue box next to the LED) picks up the tiniest movement of the head. These data are combined by the tracking computer in order to determine the spatial position of the participant. A second computer renders the environment accordingly and projects it in 3D into the HMD (visible on the second picture on the right screen in the background). The graphics are extremely good, I'd say they match current high-quality ego-shooter games.


The possibilities of such an experimental paradigm are endless. For example, I was in a virtual world where my own avatar was a woman. I could see my reflection in a mirror and it would correspond to my head movements (the face was almost photorealistic), but it wasn't me - it was this other person. The experience was insofar fascinating as it felt extremely real. In another scenario, I found myself standing on a narrow wooden board across a crater in the street. The crater was about 30m deep and 10m in diameter. I could see to its bottom - I was standing above this gaping hole. On a narrow piece of wood. I could feel the vertigo and my palms started to sweat. When the experimenter asked me to jump into the abyss, I couldn't. I knew that this was only virtual and the graphics looked more like the ego shooter that my brother used to play but still - I couldn't get myself to jump off the board for a couple of minutes. It felt that real.

Only an imbecile would not see the possibilities of such a technology for psychological research. Body image. Behavior therapy. Social interactions. I am sure that in a couple of years, the employment of 3D VR environments for psychological research will be as common as the use of fMRI today. Just that spending half an hour in a virtual world is much more fun than spending ten minutes inside the narrow tube of a scanner.