THE recent release of the results of opinion polls conducted by University of Dar es Salaam-based REDET has generated interesting public debate, even controversy. Of course, it is not surprising at all given the very nature of opinion polls on sensitive political issues where results are perceived differently by different political groups depending on how they are being portrayed. Since we will be hearing more of these polls as general elections draw closer it is not a bad idea to say something about them. Perhaps, the two relevant questions here are: why opinion polls? Do the results of opinion polls matter in the conduct of politics?

Why opinion polls? The essence of opinion polls, first and foremost, is to gauge people views on various public issues including national politics and functioning of political system. It is a barometer, ‘kipima joto’, of some sort in the management of public affairs. If properly designed and executed they provide useful feedback to political and bureaucratic elite responsible for day to day affairs of the nation.

Now, if opinion polls are that much important why do they generate controversy once results are released? This is a million-dollar question. There are several plausible explanations for that. More often, opinion polls have good news for some and bad news for others. No politician in his/her right senses would like to hear that his/her popularity is on the wane and therefore unelectable. In such situation polls results will most likely be disputed and no argument, scientific or otherwise, can settle that. In the end people tend to agree to disagree.

More important, however, is the polling process itself which is quite prone to disputes. For opinion polling to produce acceptable results it must adhere to certain basic principles of social science research. As such it is a science. For example, one must have what is commonly referred to as a ‘representative sample’ of the total population of interest. In the case of voters’ preferences for political parties or candidates a representative sample of most likely voters in 2010 elections. By any standard this is the most difficult part in conducting opinion polls. More often it boils down to how representative is the sample. It is this claim of being ‘scientific’ that most opinion poll results are being defended as reflecting majority view.

The science part of the opinion polls notwithstanding, the polling is also an art of its kind. Polling is not just the matter of asking which political party, or candidate, you like or don’t like. It is not simple matter of shooting questions and getting back answers or getting someone to fill in the questionnaire. It is more than that. The manner questions are framed, how they are asked and means used (telephone, face-to-face interview, etc.) can make a difference on the responses and how they are interpreted. This is the art part of opinion polls, as important as the science part, but subject to human failings hence disputes.

What sense do we make of all these? Opinion polls are not deterministic in the sense of providing the whole truth about what people think of their government, political leaders, and their social and economic wellbeing. Sometimes people can be deceptive saying one thing and doing something quite the opposite. If anything polls provide us with some guidance on human behaviour and expectations. Views or attitudes are not static. They do change in time and space. Therefore, due caution must be exercised in the use of the results of opinion polls by politicians and political parties.

Now, turning to the second question: Do opinion polls matter in the conduct of politics? The answer is both yes and no. I have lived in a system where national politics revolve around opinion polls. Opinion polling is a big industry. Any idea about policy must be poll-tested to determine its acceptability and identify potential areas of resistance. Political parties and candidates have polling experts on their payrolls before and during political campaigns. No having polling done or paying for polling services is tantamount to committing political suicide. In such system people views or opinions do matter and political parties or politicians can only ignore at their own peril. They have indeed ‘perfected’ the art and science of opinion polling. Of course, this is not to suggest that there are no failures in the process. There are a lot of them partly due to complexity of the political system.

Can we say the same thing about Tanzania or other African democracies? Not really. Public opinion polling is still at its infancy to say the least and we should commend those who are even attempting doing it. On the one hand, shortage of resources, both human and financial, and technologies are major constraints. On the other hand, however, the problem is with the process of forming opinions about public issues. Indeed, to form opinion about an issue, political or otherwise, one need to be informed. Therefore access to information is quite critical to the success of polling process and its outcomes. In a situation where people are ill-informed or not informed at all, the outcomes of opinion polls cannot be relied upon. More often, they raise more questions than answers. About REDET’s polls, mmhh!, let us wait and see, only time will tell!