I read with some interest James Wong Wing o­n's treatise (in places it read more like a diatribe) in defending what he says he had said and/or not said in a recent BBC interview.

It never occurred to me that Wong would o­ne day be defending his position in the letters column, and then again in space in the opinion column. After all, he always seems so self-assured about his position, even annoyingly self-righteous.

Nevertheless, if however Wong's point - after his long-winded explanation - is to say, simply, that Malaysians are fickle-minded, hypocritical but nonetheless allowed to either keep mum over their political views or speak up as their democratic right, then who is to disagree? We all knew that anyway.

But my problem with his explanation has more to do with the methodology he used in discerning how many Malaysians were for or against the US-led war against Iraq (Wong will probably dislike hearing this, but I learnt research methods at university under his nemesis, Manjit Bhatia, in Australia).

It's really quite simple. Judging by the way Wong has utilised random sampling for his so-called survey, it seems to me it would always lead to serious problems (that is why opinion polls are too treacherous to be believed).

First of all, it never is nor should it become a true representation of 'all views' or predilections from small quantities, and should be used o­nly as a very rough guide to mount a more formidable and accurate methodology.

Second, in random sampling, what you end up doing, effectively, is to control the variables. Usually we would want to control the extraneous variables so that they do not confound the relationship to be studied or cause spurious interpretations of that relationship.

What Wong has done is that he controlled all the variable - the physical and the statistical. As the result, he has exacerbated the incidence of randomisation thus making any reading of probability extremely difficult or, as his so-called survey shows, very awry and biased. He has allowed systematic effect to occur in a very systematic way.

Third, finally, o­n this basis, Wong's defence is spurious and shallow since he has scrapped every ounce of internal validity that the numbers in his "statistics" could have otherwise produced had he been smarter about how to develop and use more appropriate research methodologies.

Once you corrupt the internal logic of any research, whether based o­n valid or invalid comparisons for measurement, you have no legs to stand o­n.

Therefore, Wong seems to have a made a meal of his attempts to explain his position and what he did and how he did it - his defence, in my view, is unsophisticated and most unconvincing.