Scientific Integrity
Introduction
A lot of media reports the results of studies. There are several problems that studies can have. This includes misrepresentative and small sample sizes, and general bias. A lot of genuine science has a list of its “limitations”, showing what biases it might have.
Sample Sizes
Let’s say you wanted to know who will win in an election. You ask your 5 closest friends who they were voting for. Would this be accurate?
Probably not. Your friends are likely the same age as you, live in the same place as you and have similar political opinions to you. The country you live in is filled with people of different ages, who live in different areas and have different views.
This is why it’s important to have a representative sample. The people in the study are called the sample, and representative means that these people are similar to the population. You might need a sample that is 50% women, 40% rural, or so on.
For the sample to be accurate, the sample also must be large. If you interview 5 people at random from the population, it might just be that those 5 people are the only people voting for one party. But if you pick 1000 people, the chances of that happening are very low.
A video on representative samples. It is 2 minutes and 37 seconds long.
Bias
Like with all media, you need to check where your research is coming from and who is paying for it. For example, in the 50s, cigarette companies paid for research that doubted a link between cigarettes and cancer (Brandt 2012). It is easy to get data to say anything you want if you know how, and scientists hired by these companies do.
Limitations
A lot of research has a section for its own “limitations”, or things it could have done better. If a study doesn’t report limitations, just remember that there probably are still some and look for sample sizes, correlations and funding.
Source list
