In need to find something?
Custom Search
Related Posts with Thumbnails

Thursday, March 15, 2007

Much ado about data entry error

Technorati tags: , , ,

Walski wanted to post this a lot earlier today, but it's been a busy, busy day - meetings, then some more meetings, getting Yellow Fever shots... Yes, Walski's going somewhere, but more on that another time.

Image hosting by PhotobucketRemember that sensational "every female teenager boinks in school" report that caused such a furor. Walski had highlighted this in another post last weekend. It appears now that the report was in error, and that the final result presents a totally different picture.

The Sun today ran a story (on page 8) which stated that the correct figure of repsondents who admitted to having pre-marital sex should have been only 3%, and not the 886 out of 887 as initially reported.

Problem solved, right? Everything hunky-dory, right? WRONG!

To Walski this admission raises more questions. For example, how on earth is it possible to consistently enter the same data field wrongly 886 out of 887 times?

“The consultant team identified the source of the difference as a mistake in entering the right code during data entry from the questionnaire to the Statistical Package for Social Science (SPSS) software,” he said in a statement.
(quoting Inmind [Institut Perkembangan Minda] CEO Norizan Sharif)

Dunno... to Walski that is too much of statistical impossibility to be a data entry "mistake". The virtual screams of "cover-up" could almost be audible to these cynical ears.
(lies, damn lies and statistics, in the full post)

Assuming, however, that it was a genuine mistake, in the first place, the whole study opens up another can of fresh worms, if you were to ask Walski.

But first, chew on this paradox: if the study found that 90.5% of those surveyed exhibited lying tendencies, isn't it very possible that only 9.5% (or 84 of them) answered the survey questions truthfully, and in effect, makes the entire study a sham?

Researcher: Have you ever had pre-marital sex?
Female: Yes
Researcher: Do you lie?
Female: Yes
Researcher: Did you just lie?
Female: No

Incidentally, Institut Perkembangan Minda (Mental Development Institute) or INMIND, is a think-thank organization, operating on behalf of the Selangor state government. This probably explains why Khir Toyo claimed to not know anything about the survey being commissioned, in the Bernama news report where he labelled the report as "rubbish". The survey was initially attributed to the Selangor government directly. Also, as another aside, INMIND's Chief Executive Officer, Norizan Sharif, maintains a blog, Teratak Kembara, under the pseudonym Nobisha (which is probably a concatenation of Norizan bin Sharif).

Looking at the results, it's no surprise that Walski's buddy Howsy quite aptly insinuated that with a high number exhibiting "anti-social" demeanors of lying (90.5%) and playing truant (74.1%), these high-risk femate would probably make good candidates for politicians! And MPs.

But really, what are these female students a high-risk of? Spontaneously combusting? Becoming terrorists? OR, not fitting into the ideal picture of how the Selangor state would prefer its females to behave.

In reality though, the kids behaving the way they allegedly are could be the case where they're merely acting out the angst of teenagehood, compounded with the fact that school for them is a drudgery. Interestingly, nowhere in the news reports does it even venture to explain why these youngsters exhibit such "anti-social" behavior. Apparently it's important enough only to find out what kinds of behavior, and not why.

Neither was there any mention about the intelligence level of these troubled youth - very likely not a focus of this survey either.

Highly intelligent children trapped in an unchallenging environment do tend to act up, in ways that are sometimes construed as being "anti-social".

And isn't being late for school also possibly attributable to the worsening traffic conditions around Selangor? How anti-social of Selangor traffic.

Instead of finding out why behavior such as these have been observed, the purpose (per today's report in The Sun) of the study was "as an additional input for the development of training modules for girls".

Sounds like a morally high-handed venture, wouldn't you think? Who the fuck cares why, as long as they can modularly train the girls to be good, obedient, compliant, future breeding automatons for the great state of Selangor. In which case, who gives a shit what their intelligence level is.

In other words, transforming a whole breed of sociopaths with "undesirable" attributes, into sociopaths with acceptable behavior.