Statistics Are Integral to Research

Statistics are integral to research but it is important to know how to read, interpret, and use statistics so that one can best comprehend what one is reading and not be duped by those who may distort statistical data for subjective purposes.

Statistics are powerful, but used incorrectly or erroneously they can also distort information and lead to negative results. Just as words are ambiguous, numbers and images (such as tables, flowcharts, graphs etc.) can be misleading too.

Reading articles that contain statistical data involves a critical manner throughout and involves practicing critical techniques.

Firstly, one has to constantly question the source of the data. Even when extracted from a scientific article, the journal needs to be checked to see whether it is a peer-reviewed credible source. This refers all the more so for popular books and articles. All too often, people assume diets and do-it-yourself treatments that can be potentially destructive at their worst due to respect for statistics and the fact that the data was extracted from a journal or book that contained 'psychology' or 'science' as its tile. The background of the author has to be carefully reviewed, as well as the publisher, and the context of the statistics.

The presented data may indicate only a part of the picture. Questions, therefore, have to be asked, as for instance: the purpose of the data, the subjectivity of the researcher, the intention of his research, the policy or procedure that hinges on the statistics, and so forth. Illustrative of this fact are statistics that convey some political process, as for instance, the current Iraqi conflict. A good number of the sources are partisan, and the statistical data, authoritative and impressive as they seem, too readily reflect the interest of one side or the other.

Related to this is the investigation to discover whether all the data has been included. To return to the Iraqi conflict scenario again, some data may have been intentionally excluded or presented in an incongruous manner. The other side of the picture -- and the entire picture -- has to be seen for an accurate perspective to be garnered.

Statistics can also be erroneously interpreted. One of the most common fallacies is to confuse correlation with causation, but the two are actually distinct. My demonstrating that construction of snowmen and outbreak of acne occur simultaneously does not mean that snowmen produce acne. It may imply an underlying matter, such as the snow itself may contain some component that may instigate the outbreak, or the children who build the snowmen may be particularly vulnerable to acne, and during that period of the year, and so forth. Two factors happening concurrently, does not mean that one influences the other.

Readers may also take the percentage face blank unaware of the notion of margin of error. Margin of error implies that the data shows only an approximate result of the sampled population (usually 95%). There is a certain percentage of error either way, and so the end results can only be approximate, never absolutely certain.

Using statistics in writing contains its own set of rules:

Distinction must be made between the mean, mode, and median. Each refers to average, but they are different ways of measuring the average, and one can be irrelevant when applied to a certain instance (e.g. taking the average of a list of names).