Malta needs to focus on more high-quality research

Before starting to talk about research, I think it would be apt to first introduce myself. I'm a Statistics graduate and have been involved in research for the last five years. Moreover, I have recently completed research on "Missing data in sampling surveys", analysing and trying to predict non-response by adopting a scientific approach. I will shortly be initiating my Masters in Statistics.

My interest in examining and studying surveys induced me to reflect and consider how much we are focusing properly on research in Malta, starting first by analysing political surveys. One may say that there are certain mistakes which are surely intolerable in the science world, since inaccurate results are consequently unapplicable.

Here, I'm talking about the non-response issue, which is predominant in political surveys. Most of the time, after quoting results in political surveys in Malta, the non-response part or missing data is deleted.

Most people who do surveys do not realise that deleting non-response is scientifically incorrect for several reasons. First of all, from a particular study, it was found out that a large part of non-respondents are floating voters. Having said this, one automatically asks, how can you remove floating voters in political surveys, knowing that floating voters in Malta determine almost everything in an election?

Secondly, by removing non-respondents from a survey, one is reducing representativeness of the sample. A lack of sample representativeness implies more error and less quality in the research.

Another argument is sample size. We often see surveys having Malta's population as the target population, with a sample of only 200 or 300. There are also cases where there is a non-response of 40 per cent, meaning that only 120/180 people have responded to the survey.

The quality of such a survey is very low and there is large room for error. Consequently, having such a small sample, you will end up, for example, choosing a person from Kirkop, two from Swieqi and so on. Having such small numbers from localities will surely produce low-quality research and result inaccuracy.

The most frustrating element when discussing the quality of research in Malta is the abusing part, where statistics are used for fulfilling one's hidden agenda. For instance, how many times do we hear about surveys carried out to determine the best programmes broadcast in Malta?

Most of the time these surveys are done by respective presenters, who all say their surveys or 'independent' surveys show that their programme is the best on the island. But something is wrong! This cannot happen.

If five surveys dealing with the same issues are being carried out at the same time, these should conclude the same results. This is proven scientifically, and no opportunist can change these facts. This is abuse. Don't manipulate people by false statistics.

I also encourage politicians to help reduce the exploitation of statistics for manipulation, even though, unfortunately, I am aware that most of the time politicians are the main actors in trying to manipulate people by false statistics.

I encourage all those interested in thinking more about high-quality research. Furthermore, I believe Malta needs to invest more in research. Lack of research can be seen in the case of road construction and the installment of traffic lights - ending up having more traffic congestion after construction is completed.

By investing in research - studying how many cars pass from a particular street, at a particular time of day, or how long cars take to pass from a particular street - this would help estimate if the road construction or traffic lights would create more traffic or not.

Statistics and research shouldn't be taken lightly, especially since the matters dealt with are scientific. After all, science is based on facts and one cannot invent theories that are not based on explanation and proof.

If we really want to carry out high-quality research, we need to be more focused because when statistics and research are used well, for better results can be achieved, which is extremely worthwhile.

vincentmarmara@gmail.com