The Killer Papers

On one cold February day,  the world’s greatest detective and I were in a professor’s office reading student research proposals, looking to find any papers that had been plagiarized.

Allow me to explain. Usually, the research methods professor caught and dealt with plagiarists on her own (with some help from computerized tools).  However, given that one of the plagiarists was probably also a murderer, that the class was the most dishonest she had encountered in her 36 years of teaching, and that Sherlock Holmes had offered his services, she turned the case over to him with instructions to find both the cheaters and the murderer. Consequently, Holmes and I found ourselves reading student research proposals.

                Watson: I think Amy’s paper may be plagiarized because it is on a common topic (meditation and stress). I will try to find the article she took this from.

                Holmes: How many participants did she propose testing?

                Watson: Ten.

                Holmes: Ten? Classic undergraduate mistake-- she is proposing a study that will not have enough power to obtain significant results. And since it was a proposal for a study she was not going to do, it would not have cost her anything to say she was testing 200 participants. Since it is doubtful that any published study would get significant results with only 10 participants, I think we can assume that she did not merely copy a method section from a published work.

                Watson: Well, what about Karen’s proposal? She proposes testing 100 treatment participants in one room and 100 no-treatment participants in a different room. Two hundred participants is a large sample size. Perhaps Karen "borrowed" her method section from a published paper.

                Holmes: Highly unlikely. Her proposal involves 200 participants but only 2 independent units. The 100 treatment participants are one independent unit; the 100 no-treatment participants are another unit. In other words, all 100 treatment participants were tested by the same experimenter, in the same room, and under the same conditions. Similarly, all 100 no-treatment participants were tested in a different room, by a different experimenter, under the different conditions. Put another way, the treatment participants differ from the no-treatment group in more ways than just getting the treatment:  They differ from the no-treatment participants in that they were tested by a different experimenter, in a different room, and under different conditions.

                Watson: Under different conditions?

                Holmes: Yes. Besides being tested at the same time by a different experimenter in a different room (or being tested at a different time by the same experimenter in the same room), the testing conditions in one session will not be identical to the conditions in another session. For example, suppose a participant in the treatment condition sneezed, coughed, yawned, dropped a pencil, came in late, or made a comment. That participant’s actions would affect everyone in the treatment condition but would not affect anyone in the no-treatment condition. Or, suppose a light flickered in the treatment group’s room, but not in the other room —or that there was a commotion visible from the treatment group room’s window but not from the other room’s window.  Any of these differences would affect everyone in the treatment group but no one in the no-treatment group. So, some random difference in the testing conditions could make it look like the treatment had an effect.

                Watson: There are going to be differences between testing conditions, so how could this problem be avoided?

                Holmes:  If participants had been tested individually, random differences between the treatment conditions would probably balance out. With individual testing, it would be very unlikely that the 100 treatment participants would be exposed to an outside noise but none of the 100 no-treatment participants were so exposed.

                Watson: I see. So, a study that violated independence would like that would not be published. So, Karen did not lift her method section from a published study.

                Holmes: I would hope not.

                Watson: Tom’s proposal for a study involving cognitive dissonance and therapy is extremely well-written and incredibly detailed. I have checked and found that it is not from any published work, so it is probably not plagiarized.

                Holmes: Unfortunately, it is plagiarized.

                Watson: How do you know?

                Holmes: For an undergraduate, the highly polished style—and having 20 more journal references than the assignment required—is suspicious. In addition, the proposal has extremely detailed predictions about the results—it predicts specific t values and p values, none of which are statistically significant, and none of which support the hypotheses. Finally, whereas a research proposal is written in the future tense (it describes what will happen), Tom’s proposal is written in the past tense—like a research report describing what the researchers had done. My conclusion is that Tom copied from a paper that got null results—and thus was not publishable.

                Watson: You mean Tom took advantage of the file drawer problem?

                Holmes: Precisely. Look for a professor at this school who was, until recently, doing research on cognitive dissonance and therapy. I suspect that professor will recognize this as one of her unpublished papers on the topic. If you have trouble tracking that down, consult with the school ’s IRB, which should have some record of that research.

                Watson: We aren’t making much progress. Should we try a different tactic?

                Holmes: Just as scientists try to eliminate hypotheses, let us try to eliminate suspects. We can start by eliminating papers that probably were not plagiarized.

                Watson: I feel good about Jerry’s proposal on hypnosis and smoking cessation not being plagiarized. It looks like good, solid student work.  I like this proposal. And, look, she attached a nice summary at the end.

                Holmes: Is it labeled “Abstract”?

                Watson: Yes

                Holmes: That is a bad sign. The Abstract should be at the beginning of the paper.

                Watson:  Oh! Here is another bad sign—Jimmy turned in the same paper, only without an Abstract.

                Holmes: I’m afraid that you may find yet another copy of this paper on the victim’s computer.

                Watson: Unfortunately, the victim’s laptop is missing.  

                Holmes: I think it is safe to assume that the victim’s paper was on hypnosis and smoking cessation. Furthermore, Jimmy stole the victim’s paper before the victim had written the Abstract and that Jerry stole the victim’s paper before the victim had pasted the Abstract into the beginning of the paper. The victim had the Abstract at the end of her paper, probably because she had taken the advice to write the Abstract last. Writing the Abstract last is a good idea for many reasons, one of which is that it is hard to summarize something before it exists. But it seems that the victim was killed before she could put the Abstract in the front of her paper.  

                Watson: So, is Jerry our murderer?

                Holmes: Not necessarily, but Jerry would be one of the last people to see the victim alive.

                Watson: Do we have any other leads?

                Holmes: Yes, as luck would have it, there was a computer glitch last night at the library that made it impossible to access journals online. As a result, most of the class had to go to the library last night to read actual, physical, paper journals to get the sources they needed to finish their introduction. Consequently, by looking at each student’s Introduction, we may be able to deduce their whereabouts around the time of the murder.

Watson: How?

Holmes: I see you have read the Introduction sections of Arnie’s, Bella’s, and Carla’s proposals. What do you make of those section.?

Watson: Extremely disorganized. I cannot see how the studies described fit together.

Holmes: What if I told you their Introductions were, in a way, organized?

Watson: How?

Holmes: They tell us which article each person found and read first (the first article cited), which article they read second (the second article cited), and so on. That’s the only way to explain why their introduction is not in any kind of logical order.

Watson: So, they first wrote about the first article they found, then they wrote about second article they found, and so on—without using an outline or connecting the studies to each other?

Holmes: This “strategy” for writing an introduction is a common mistake that beginning writers make.

Watson: So, we know that the last article they read was the last article in their introduction. How does that help us?

Holmes: We know those three were all in the library last night right before it closed. If the last reference in the introduction leads us to the last place they were in the library, it may help us in verifying alibis and  interrogating witnesses.   

Watson: How so?

Holmes: For example, Carla’s last reference, for some reason, is a very old one. Since the old journals are in the library’s basement, we know she was in the basement. Arnie’s last reference, on the other hand, is very recent—it is last month’s issue of the Journal of Personality and Social Psychology. In this library, the new issues, are on the second floor.

Watson: And, since the victim was found in the second floor, we know Carla was near the victim shortly before or during the murder.

Holmes: Right, Watson.

Under Holmes’ intense questioning, Carla finally admitted that she had seen Jerry and the victim in the basement. Holmes confronted Jerry with these facts. Jerry, whether due to guilt about the murder or shame about being caught plagiarizing, quickly confessed.

After this scandal, the school incorporated lessons on plagiarism in orientation and in freshman English. Since then, the school has had many fewer cases of plagiarism—and no murders!