How Significant is Data Fraud in the Scientific Community?

​Author: Maggie Johnson

One of our interns had an industrial psychology summer internship through which she received data on user experiences. When presenting her data to a group of board members of the sales division, one of the members asked if she could “make the retention numbers say something else.” In response, she gave them a clear no, as that was against her ethical and moral standards. This led her to question how often this is asked of employees and how often data fudging occurs. Many different industries fund research, so we wondered if this research is manipulated to show the companies that fund studies what they want to see. Because of this question, we are cautious to agree with data provided by privately funded studies without digging deeper. As such, we also wondered why people would manipulate their data to begin with.

We discussed the possibility of profit being the biggest motive. An article we read discussing this topic further suggests that many employees agree to edit data to fit what their employers want to see because they felt pressured or there was financial incentive to do so. Another story, however, mentioned a man who edited the background data of a client because they weren’t eligible for certain things, and this change in information allowed his client to get treatment and resources. Considering the pressure some employees may face, we understood why they may choose to fudge data. For example, if someone has a family to support and their job is in jeopardy, and they refuse to manipulate the data in a way that works best for the company, they could lose everything by saying no to the request. 
 
In terms of research in general, not in the realm of private funding, there is also the idea of “publish or perish”. In other words, there is a lot of pressure on researchers to publish a lot of articles, especially if they want to be considered for tenure (i.e., permanent employment) at an academic institution. This can be difficult, though, because journals that publish scientific findings tend to only want to publish significant results. So, even if research is being conducted for a full year, if those results are not significant or if your hypothesis does not align with your results, there is a strong chance that the data you’ve been working on won’t get published. In addition, it can be difficult to get funding if you don’t acquire enough preliminary data that shows strong evidence for potential results from a future study.  
 
In EEG research, like the research we conduct in this lab, there are a lot of different ways to process or prepare the data to be analyzed. Different filters can be used to process the raw data of the study and different filters can show different results. When reading articles, however, you can not necessarily see what filters a researcher used to achieve their results, nor how many times they may have changed filters to make their results significant. To avoid future data manipulation and “fudging,” we suggested a consistency in protocol in conducting research—especially EEG research. Methods used to process data should be mentioned in grants and preliminary stages of studies to lay down a path to conduct the research without variation. This way, we can understand what was used to process data and ensure data was not manipulated in a way that best suits the researcher.