Print1.1 Falsification
A familiar scene of falsification of evidence can found while watching a courtroom television drama, where a law enforcement officer is portrayed as having tainted key evidence during the process of investigation and the suspect on trial is let go, even if the suspect may be guilty. Why does this happen? Why should someone, possibly a criminal, be let go because a piece of evidence was falsified, even if the rest of the evidence was not changed? The reasoning is that any conclusions based on or influenced by the falsified evidence cannot be sustained. Further, falsified evidence brings the validity of all of the other evidence in the case into question as well.
A article from the news illustrates the point: "In her order, [the Judge] -- a former prosecutor -- issued a scathing indictment of the prosecutor in that case for hiding evidence that [the murdered] was allegedly, a sexual predator who had molested [the murderer] and other children. [The Judge] said "evidence has plainly been suppressed," and accused former assistant D.A. of engaging in "gamesmanship" and "playing fast and loose." The judge also said [the prosecutor] "had no problem disregarding her ethical obligations" in an attempt to win."
Another way evidence can be falsified is if it is withheld, particularly if it demonstrates a counter argument, such as DNA evidence demonstrating the innocence of a suspect. If this data is available, but withheld, then it is also a form of falsification or misrepresentation of the available data. There are many similar analogies about falsification in law that also carry over to issues about falsification of data in science, engineering, economics, etc. While what ultimately constitutes proof and certainty in a court of law ("beyond the shadow of a doubt") is not the same that constitutes proof or certainty in science (>95%), the impacts and problems of falsification are very similar.
Falsification in sciences and engineering arise from manipulating research materials, equipment, or processes, or changing or omitting data or results such that research observations are not accurately represented in the research record. Falsification often occurs when a researcher chooses to omit data that goes against confirming a hypothesis, such as omitting to report harmful, but rarely observed, side-effects in Phase 1 or 3 trials of testing a new medication. In this context, falsification of data can lead directly to harming individuals who later take the medication.
Other forms of falsification not of the research ethics kind: There are times when data may be false for reasons of instrumental calibration, such as the recent example of the particles that were thought to be traveling faster than the speed of light, when later it turned out to be instrumental calibration issues. This particular issue does not constitute falsification.There is another notion of falsification in the sciences that should not be confused with the falsification of research data, namely, the falsification of a hypothesis. This simply means that a scientific hypothesis has been demonstrated to be logically false based on existing data.
Significant concerns emerge when data is falsified
- First, when conclusions are presented on falsified or incomplete data, they can hide problems about how certain we can be about our conclusions based on such research. This applies to more than just research records, it can apply to other kinds of data tracking. For example, would you want to fly in a plane that had a falsified maintenance record?
- Second, other research may be based on falsified assumptions; and, thus, errors may perpetuate throughout a later process.
- Third, falsified research that receives funds and/or is published instead of other research (that would not be falsified) robs the funders (typically the public through government grants) of the outcomes of proper research.
- Fourth, careers based on falsified research create problems and deficits, and often significant embarrassments, for the research institutions.
Discussion Questions
- Can you think of some reasons why someone might want to falsify their data?
- In what kinds of situations might falsification be more problematic than in others?
- When might it be ok to omit certain kinds of data from the record?
- Can you think of some ways to test whether someone's data has been falsified?