Sunday, July 29, 2007

Teaching Microsoft: the aftermath

A lot has happened since I debunked Alan MacCormack's research. First, a very clever attack from some "Rufus":

"Insufficient sample sizes because the survey is not meant to be representative. Additionally, for exploratory research, it is nothing unusual to use a semi-structured approach. The goal is, for example, to extract key words the can be used for proper measurement in the later, the 'real' survey."

This may all be true, but I'm not interested in whether this is a valid methodology in social sciences. What I'm interested in is whether the methodology justifies the conclusions. Obviously not, according to Rufus:

"The intention of the study was not to reject any hypotheses or to be representative."


If the study was not intended to be representative the conclusions are a priori invalid. Still, they are presented as if they are valid and applicable. Rufus continues:

"You still decided to use lots of rhetoric to 'prove' your point. For example, this one: 'MacCormack also fails to state why licensing implications are more complex than the war in Iraq, the greenhouse effect and save the whales'. This is of course a rhetoric attack."


No, this is not rhetoric attack. When a researcher makes choices I expect him to explain why he chose to do that. McCormack does not or not sufficiently.

Finally, Rufus debunks his entire rhetoric himself:

"The problem is obviously that his results shouldn't have been advertised in the first place. However, is this the professor's fault or Microsoft's?"


Obviously, the professor's. Remember that in the eWeek article he comments in great length on his research. If this research wasn't intended to be published in this stage he should have said so. He had every occasion to do so, didn't he? I have contacted Rufus through email, but have not received any response yet. Maybe what "DB" wrote isn't too far from the truth:

"Looks like Microsoft are reading this too. And if this was an exploratory survey they wouldn't be trumpeting it as the real thing."


Of course, I tried to comment on the eWeek article, only to find that I was banned:
You have been banned from posting. If you feel this is an error, please
email sean_gallagher@ziffdavis.com

Sure, I contacted Sean Gallagher. I haven't had a response yet, for all I know it may be an error. If you want to know what I wrote, here is the comment I submitted:
I've written an article on this: The Beez speaks
It is interesting that one of my comments states that although the method used is a valid practice, the conclusions as stated in this article cannot be backed by the method used.

Bad journalism.. pity.

I leave it to you to decide whether that is sufficient cause to be banned or a valid comment. I never expected to follow the fate of Groklaw. Let's see what comes next.. I can only say that I'm a real person and I'm not paid or sponsored by IBM. Just your average blogger. ;-)

[Update: It has been reported that Sean Gallagher doesn't work for Ziff Davis anymore, so guys, update your messages! And Sean, you're out of the equation, sorry for mentioning your name, man. After weeks not a single email has made it to my inbox.]

No comments: