May 21, 2011

Statistics can be used to mislead?

Written by Saran Soeung

Statistics have been utilised by government institutions, business entities, or non-governmental organisations to indicate their own achievement and positive or negative development to public. Each day, the media bombards audiences a myriad of statistics in order to tabulate that stock markets have increased or tumbled or the toll of crime in cities are alarmed or debunked. This kind of data is used to gain public confidence and to prove audiences that they are working very well. Even though some people believe that statistics are technically accurate and valuable, this author feels that statistics can be misleading and deliberately misinterpreted.

One obvious problem with statistics is making a generalisation from small sample size. Most often, statistics are obtained by taking sample from large group of people and it is assumed that the sample represents the whole group of people (Adam, 1995, Internet). For example, after completion of project, a special team is commissioned to conduct a survey in Kampot Province, one community where women’s right education was conducted. 100 people could be randomly selected to be interviewed. As a result, if 75 per cent confirmed that gender-based violence in their community has decreased gradually, while the rest responded that the violence has leveled off, the team accordingly could conclude that the violence in Kampot has declined steadily. It would not be surprising since it has indicated positive result; however, it is very risky to provide such a generalization as the sample size is too small and sample selection can be biased (Responsible Thinking: Misleading Statistics, Author Unknown, Internet).

Unclear definition also contributes to misleading statistics. The definition has played an important role in ensuring that the statistics are technically accurate and reliable. If it fails to do so, the quality of report will not be acceptable and biased (Pallant, 2009, 34-35). Let us take a look at an example of the quality of education of public and private universities. The public universities declared that their quality is much higher than the quality of private one. However, the private universities also declared that their qualities are outstanding. In this case, both the public and private universities did not confirm on how they assessed the quality of education. In fact, the qualities of education can be judged by qualified teachers, disciplines, studies hours, curriculums or teaching approaches. This denotes that different institutions have produced diverse statistics regarding their own definition and the result is statistically atypical.

The other problem is when all relevant variables are not used for research studies. The variable is generally a correlation between one another and it manipulates on the accuracy of the research studies. One example is about the increase of rice yield after using the compost. If one research study has implied that the rice yield has been increasing because of using the fertilizer introduced by a new company, the accuracy of the data can ambiguous. This conclusion can be inaccurate and statistically used to fool audiences. In fact, it is not only fertilizer but many other variables such as taking care of the rice field, maintaining the level of water, or using new farming techniques. Indeed, the more variables we look at, the more confidence can be used to conclude that the rice yield has increased accordingly.

In conclusion, it should be evident that statistics can be used to misinform and sometimes intentionally distort. Most importantly, if the statistics are released by a group that has strong biased, political or philosophical agenda, we should realize that the information has been carefully chosen to promote their own specific purposes. However, it will be beneficial if statistics are released by trustworthy or unbiased institutions and interpreted and compared fairly.

References:


posted by Soeung Saran @ 9:56 AM