Avoid Bad Research Practices with Pre-Registration
Investigations must be carried out in ethical ways. Researchers have to comply with ethical standards that ensure the validity of their research. However, in the past, some researchers have used bad research practices. These practices mainly consist of committing small infractions in order to be able to continue with the investigation and eventually publish it.
The goal of modern research dynamics is to publish results. This goal can exert negative pressure on the researchers. Therefore, it’s common for many of them to resort to bad research practices. There have been some very striking cases throughout history, such as the Stanford Prison Experiment and Diederik Stapel’s case.
The Stanford Prison Experiment is one of the most famous social psychology experiments. However, some recordings that demonstrated the bad research practices used to carry out this experiment came to light. For example, relying on anecdotal evidence, training the guards, making implicit demands, and incorrectly interpreting the results. Ever since everything came to light, people have debated whether this study should still be included in textbooks.
Diederik Stapel was a social psychology professor at the University of Tilburg. In 2011, it was discovered that Stapel had falsified data from many of his studies. Instead of using participants in his studies, Stapel had invented data in such a way that he always got the results he expected.
These bad research practices were unknown to his students, who trusted their teacher. That is until one of his Ph.D. students found it strange that he always got the results he expected. From that point on, everything fell apart.
“Science is the great antidote to the poison of enthusiasm and superstition.”
Bad research practices
Although these cases came to light, they’re in the minority. Most bad research practices are more subtle. Here are the most common bad research practices:
- P-hacking: To test the hypotheses of statistical analyses, researchers use the p-value, which is usually 0.05. However, when they use many variables and experimental conditions, this value can be biased. Therefore, the recommendation is to reduce it, which researchers should do before conducting the study.
- Low power: A small sample size can make the study have low statistical power. In turn, low power will likely result in a false positive. In other words, the research will suggest something that doesn’t really exist.
- HARKing: It consists of changing the hypothesis once the researchers see that the results aren’t what they initially expected.
One solution to these bad research practices is pre-registration. Pre-registration consists of making the goals of the study and the methods the researchers will use public. Therefore, anyone can verify that the researchers carried out the study ethically, or as they stated beforehand.
Pre-registration is very simple. There are web pages that can help you out, such as the Open Science Framework (OSF). In this page, you can find different pre-registration templates. Likewise, you can upload all of the study material (databases, questionnaires, supplementary material, etc.) and make it public in a simple way.
“The important thing in science is not so much to obtain new facts as to discover new ways of thinking about them.”
-Sir William Lawrence Bragg-
How to pre-register
You can find different pre-registration templates in the Open Science Framework. Some ask for more information than others. Taking one of the simplest templates called AsPredicted as an example, the sections you need to fill out are:
- Data: If you’ve already collected the data or will be collecting it after pre-registration.
- Study hypothesis: Explain the hypothesis you’ll test out.
- Variables: What the study variables are and how you’ll measure them.
- Conditions: The conditions you’ll assign to the participants (for example, control and experimental).
- Analysis: The analysis you’ll do once you’ve collected the data.
- Observations: The number of observations you’ll collect. In other words, the number of participants that you need.
By making the data of the studies public, they’re visible to everyone. Thus, the possibility of using bad research practices decreases. Although pre-registration takes longer since you have to think about what you’ll research, a lot of people do it. Additionally, it allows science to be more transparent.