Applicable definitions from the Online Ethics Center Glossary:

DATA SELECTION: Data is selected for presentation on the basis of clear criteria for thinking it is least subject to confounding influences or "noise."

FABRICATION: In research ethics the term, "fabrication" means making up data, experiments or other significant information in proposing, conducting or reporting research. In engineering, the term "fabrication" has a benign connotation, meaning to make something.

FALSIFICATION: In research ethics the term "falsification" means changing or misrepresenting data or experiments, or misrepresenting other significant matters, such as the credentials of an investigator in a research proposal. Distinguishing falsification of data from legitimate data selection takes judgment and an understanding of statistical methods.

FRAUD: A fraud is an intentional deception perpetrated to secure an unfair gain. The term "research fraud" or "scientific fraud" is also used to mean an intentional deception about results, a type of research misconduct. In this case, there need not be a financial benefit which the perpetrator receives, and there need not be an injured party.

It is useful to sort out the characteristics of a fraudulent act. For this it is instructive to look at the legal definition. In a civil law suit charging fraud there is a plaintiff who brings the charge against a defendant.

To win the suit the plaintiff must prove five points

• the defendant made a false representation
• the defendant knew that the representation was false or at least recklessly disregarded whether it was true or false.
• there was intent to induce belief in the misrepresentation
• the plaintiff had a reasonable belief in the misrepresentation
• the plaintiff suffered damage as a result.

The Backwards Math (A software industry example)

Jay's boss is an acknowledged expert in the field of automated software code generation tools. Jay is the leader of a group that has been charged with selecting the tool that generates code with the fewest errors, and the search has narrowed to two possibilities, Code Generator 'A' and Code Generator 'B'.

The boss is certain that the best choice is 'A', but he directs that tests be run on both, 'just for the record.' Owing to inexperienced help, the tests take longer than expected and the results show that 'B' generates code with significantly fewer errors. The engineers question the validity of the tests, but because of the project's timetable, there is no time to repeat the series. The boss, therefore, directs Jay to work the math backwards and come up with phony data to substantiate the choice of Code Generator 'A', a choice that all of the software engineers in the group, including Jay, agree with. Jay writes the report.

Consider the following with regard to the case of The Backwards Math:

1. What would you have done had you been Jay? Would you have written the report?
2. If you had refused to write the report, how could you have justified this refusal since you, along with all of the others on the team, felt that the test data was invalid and there was not time to duplicate the test? Is it not likely that experience is a better guide than one set of questionable test data? If yes, does this impact on the ethical nature of the dilemma?
3. Would it have been a good idea to write the report and also write a memo saying that you were being directed to do something which was unethical? If you did this, would it be done just to cover you if you are found out? Would that make what you did any more ethical than it would have been without the memo?
4. How about the alternative of writing the report, but not signing it? Would this be a satisfactory solution to your ethical dilemma?
5. Is this a case in which you would become an "internal" whistleblower by going over your boss' head and reporting that you had been asked to write a false report?  [Murdough Center]

Data Selection, Legitimate or Illegitimate? (A university research example)

Elton, an engineer, is conducting graduate research at a major university. In order to complete his graduate studies and produce an advanced degree, Elton must develop a research report. As a part of the effort to develop his report, Elton compiles a vast amount of data relevant to his subject.

The vast majority of data obtained by Elton strongly support not only his report's conclusion, but also the conclusions of others. However, a small proportion of the data differs from the data consistent with Elton's conclusion. Convinced that his report is sound however, and concerned that the inclusion of the anomalous data would only detract from the essentials of the report, Elton does not include these figures in the report. [Online Ethics Center]

1. Was Elton guilty of falsification of research results in omitting the anomalous data?

2. What type of additional information would be ethically relevant to this case?

3. Consider the characteristics of a fraudulent act as described above. Were Elton's actions fraudulent? In the previous scenario, were Jay's actions fraudulent by the definition?

4. Whether or not Elton and/or Jay's actions were fraudulent, do you consider them unethical?