Research data generates ambiguous conclusions
Statistics are an important part of understanding the world around us. When read correctly, they can reveal a lot of information. Though there are many common ways that statistics can be honest, it is also very easy to grossly misrepresent the findings.
One possibility for misinterpreting statistics has its roots in how statistics report averages. Averages are helpful when one tries to understand general statistics about large groups, but it is important to understand averages within context. A study looking into the effects of teaching styles on academic achievement can conclude that on average, group discussions can lead to better test scores than lecture classes. This report seems to indicate that a group discussion is a more effective form of teaching than lectures.
While the statement may be true, coming up with a conclusion is a matter of understanding how large the difference between the two possible outcomes is.
If the lecture group on average scored two points lower than the discussion group, then the difference is negligible and, once understood comparatively, this statistic does not lead to the conclusion that more discussions groups should be implemented in educational settings.
Another important fact to keep in mind is the groups that are being compared. Take a commonly misunderstood statistic: women get paid 25 cents less than men. When hearing a headline like that, it is easy to assume that the groups being compared are comparable. In reality, that commonly-cited statistic is measuring all men and all women in general. This does not take into consideration each gender’s education, job, time spent within an industry and other important factors. This is not to say that a gender pay gap does not exist. Rather, it is not as extreme as 75 cents to the dollar.
When controlled for similar men and women in similar fields, the gap shrinks to only two cents to the dollar, which is much less dramatic.
The way a study defines its terms or how its variables are operationally defined are also extremely important. The New York Times reported that one out of every four women will experience sexual assault on campus. To support this claim, the publication cited a survey done by the Association of American Universities. Among other issues such as improper sampling, this survey had a very vague definition of what defines sexual assault. Their definition included some non-arguable components of sexual assault like nonconsensual sexual contact, but it also included sexual touching which is a vague term that further included acts such as rubbing and even kissing. All of this adds up to a very alarmist and sensationalized statistic. This is not to say sexual assault is not a problem, but alarmist headlines do not help anyone.
Studies can misreport information or have issues with biases and improper sampling and therefore, as consumers of statistics, it is important to dig deeper and to view things critically.