Independent institute for race relations neededIt is true to say that any comment on the views and opinions of individuals on questions and issues pertaining to race relations would be of interest in any multiracial society.

Indeed, Farish Noor in a recent much acclaimed review of a book on "The naked Social Order" and 'Racial Polarisation in Malaysia' stated that " it is race that shapes and informs political mentality and governmentality in the country".

Therefore it is to be expected that a recent Opinion article in one of the local leading English newspapers should have attracted much interest among readers. This is particularly because the items reported on covered such controversial topics and questions as those ranging from the implementation of Hudud, the Islamic State, Chinese National Schools and the "Malaysian Mindset".

However, judging from the limited background information provided as to whether the article was specifically based on a 'nationwide survey' or on a specific opinion poll, or both, it seems obvious that the article's writers had difficulty in analysing the data presented by the researches. The unfortunate outcome was that certain responses were speculative and indeed others quite misleading.

It is therefore important, to draw attention to the pre-requisites for validity and reliability in the methodology for social research if the findings are to be meaningful for better understanding, and indeed, for possible policy implications particularly in the area of race and ethnic relations.

Limitations

In the first place it is generally agreed that soliciting primary data on the basis of a poll has limitations as a method in itself. For this reason it is normally the method of market research where the expected responses would be based on specific questions such as the quality of products for consumption.

If used in the context of social research it has been mostly on Opinion Polls focused on the popularity of candidates during elections mostly in terms of who will win (with some disastrous outcomes!).

Opinion polls have also been adopted to gauge public responses to specific questions, but the questions are based on a professionally prepared scale of responses ranging from the positive eg 'agree very strongly' to negative 'totally disagree'.

It is important to remember that the respondents, particularly in the western counties where the method is popular, are quite used to be being asked questions in such polls, which cannot be said for countries such as Malaysia.

Therefore inaccuracy in responses can arise particularly where the questions themselves are on topics such as Hudud that may not be clearly understood in the same way by all respondents. Also using the telephone to solicit responses to controversial and sometimes 'sensitive' questions in the Malaysian context can also seriously affect the validity of the responses obtained.

Problem of reliability

But the biggest methodology problem is that of reliability. There seems to be little, if any, evidence to ensure the representativeness of the sample. In the first place the term 'random sample' itself does not mean a sample chosen 'at random' but is a statistical concept where such a sample is drawn from an identified sample frame so that we speak of a random sample in terms of percentage-normally 10%.

More seriously, in soliciting opinions on questions of race and ethnicity it is essential to take into account the variables of education, occupation and income for the simple reason that these are independent variables (meaning that they are causal). In other words the responses will vary depending also on the educational, occupational and income levels of the respondents. Therefore, if the responses are to be reliable separate samples based on each of these respective variables must be drawn. These are called random stratified samples.

In this way the responses obtained would be reliable in the sense that being based on education, occupation and income levels they could, through cross tabulation and other statistical methods, be 'run' against the other background variables thereby presenting patterns in the responses towards a holistic picture. More importantly, they would be reliable in the sense that if another study were undertaken with a similar sample soon after, the results should be the same.

This comment is intended to highlight two important dimensions in research studies in the area of race and ethnic relations in Malaysia today.

In the first place while it is encouraging to note that there is a new interest in undertaking research to find out what is happening on the ground, at the same time it is imperative that such research be undertaken by qualified and experienced people.

Particularly in multiracial countries such as Malaysia social research is much too important to be left to those who take a "cavalier" approach without rigourous training in methodology and in practical experiences. (For instance it was reported that the initial racial problems of Kampong Medan were to some extent aggravated by questionable research methodology being adopted)

The second and more important argument arising from the above discussion is that with a new administration headed by the Prime Minister Abdullah Ahmad Badawi, that is determined to steer Malaysia 'on course' with a blend of the traditional with the modern, there is the need to set up an independent Institute for Race Relations Research. Such an Institute could bring together the best available expertise locally, not only for theoretically driven empirical research, but also and more importantly for the training of new researchers.

Once we have the appropriate data base in place, there is much that can be achieved also by bringing in the philosophies, and the teachings of the different religions and particularly Islamic principles, in seeking better communication, critical thinking and problem solving towards an integrated Malaysian nation.

DR COLLIN ABRAHAM, who started his career as a civil servant, had spent 16 years as a sociology lecturer in Universiti Sains Malaysia. He is currently a freelance consultant.