<p>Bigtrees, point about Op’s son not being willing to turn in other student is a good point and a red flag for me also. Why would you protect someone who has purposefully taken advantage of you, put you into a very precarious and compromised situation, done so without any collusion, and with whom you have no meaningful friendship/relationship? Doesn’t knowing that someone cheated off of you and refusing to acknowledge it constitute a violation of the honor code by itself? This could be just misplaced loyalty, but something just doesn’t smell right.</p>
<p>Turning someone in for cheating from you is not dishonoroable. It is not being a narc or tattle tale. It’s normal the students to feel they shouldn’t tell on another student, but this is a circumstance that coming forward with the information is the right thing to do.</p>
<p>An adult version of this type of situations is a vehicle accident. Say you are driving on the road and somebody rear ends you. In the process your vehicle sustains a couple thousand dollars damage, like a crushed bumper and it deployed your airbags. Do you call the police? Of course!</p>
<p>In this case, the other student didn’t do physical damage but damaged the students reputation, injured his potential to get into college, and will reduce the amount of activities that he may be able to participate in next year. If the S is innocent, the other student should be held accountable as well.</p>
<p>tjd-- Wow, that’s a bad paper. Notice that it wasn’t published in a statistical journal (it would have been laughed off) but in a journal of education, where, presumably, the reviewers are innumerate. The author’s method is, to put it bluntly, wrong.</p>
<p>Let’s do a simple example. We have a test with one question, with six possible answers. This test, as it happens, is so hard that everyone just guesses, so among the general test-taking population all of the answers are selected at an equal rate of 1/6.</p>
<p>You and I take the test. We both pick the same answer, answer 3. Mogull says, aha, we must have collaborated, because the chance that we’d both pick the same answer is only 1/6*1/6 = 1 in 36. But that’s clearly wrong; we have a one in six chance of matching. It’s exactly the same as if I rolled a blue die to pick my answer, and you rolled a red die to pick yours. What’s the chance of rolling doubles? One in six, and that is also the chance that you and I randomly pick the same answer to the test.</p>
<p>QuantMech’s remark about birthdays, though he didn’t explain it, is exactly on point. We need to distinguish two questions that sound almost the same: What’s the chance that two students have the same birthday? versus What’s the chance that two students have their birthday today? Similarly, in the cheating situation, we are not asking the probability that both students have the same specified set of answers. We just want to know the chance that the second student has the same answers as the first student, whatever answers the first student has.</p>
<p>Smom, I truly feel for you. Having 5 boys who have been involved in troubles, not always able to get to the truth of the matter, I can tell you I know how you must feel. I think you are doing the right thing.</p>
<p>I don’t know enough about how the evidence and the rules of the classes and school are binding on your son. The attorney will have to dig through all of that. It is not a matter of fairness most of the time, but the way things are set up. A statistical coincidence in multiple choice tests does not seem to be proof positive from the cursory search I have done, though in plagerism and other situations it can be. It’s something the attorney and the district need to address. I consider such accusations serious without merit. Whether cheating occurred or not, to get a confession by accusing and threatening is not right, in my opinion. There either are grounds for the accusation and proof or there is not. That your son says he did not cheat and does not know of any cheating that occurred, again is what you have to go on. The rest is up to the system. I am appalled that this could happen the way it did. My sympathies.</p>
<p>Thanks, Cardinal Fang. There are a couple of reasons that I keep suggesting that smom123 needs to contact a statistician. Some of them were given by goaliedad in post #180–which I recommend to anyone who is just skimming the thread. No one is more keenly aware of the limitations of statistics than a statistician.</p>
<p>A statistician will be able to discuss the assumptions that go into the calculation that yields 1 chance in 14 billion or 3 in 16 trillion, or whatever, and could re-calculate the odds with different assumptions. The fact that the students have similar educational backgrounds is relevant to the true odds, but it is not built into standard programs to calculate the odds. A statistician could also look at the effect of considering the other 7 tests, where there was apparently less correlation.</p>
<p>A statistician (and for that matter, practically anyone) can tell you that the statistics cannot say whether smom123’s son was aware of the copying, assuming that copying was actually occurring. (Have to say, it seems likely that there was copying, based on all the information here.)</p>
<p>
</p>
<p>I think you’ve made your point.</p>
<p>
</p>
<p>This will get expensive quick. An attorney costs $200 an hour typically. 10 hours of attorney time costs $2000. 50 hours of attorney time costs $10,000.</p>
<p>Cardinal Fang, I don’t agree that the referenced paper is completely off base, just that the statistics can’t be applied to reach the conclusion the paper reached. Let me explain: </p>
<p>It is the case that two students, chosen completely at random, for the question answered incorrectly at 6%, would have only a .0036 change of both having incorrectly chosen that answer. However, I think you cannot make the extrapolation, as the paper does, that the same methodology be applied to two students NOT CHOSEN AT RANDOM. In this case what you’re trying to determine is, given the pattern of incorrect answers on the first test, what are the statistical odds of exactly the same answer pattern on the second test. This is the calculation you did earlier, and I expect it is correct. It also turns out to just be the square root of the answer found using the methodology suggested in the paper. Unfortunately, still very, very long odds. That’s the correctly analogy: choose someone, and determine the odds of a match. </p>
<p>The paper also makes a huge mistake to conclude that, in a specific case, the same one wrong answer by itself is enough to statistically conclude cheating. Of course that is ridiculous. </p>
<p>And I’ll tell you why the birthday example still makes no sense to me. If you pick a student out of a class of 24 (or 30), and then pick another student, there is NOT a 1 in 3 chance that these two have the same birthday, as cited by both Eric and Quant. Not even close. The odds are (probably) what your calculation suggests they would be, about 1 in 365.</p>
<p>I brought up the birthday question. I think the relevant question, in trying to decide whether the birthday question is the same as the cheating question, is to ask whether finding that <em>any</em> two children have identical exams, is analogous to <em>any</em> two children having the same birthday.</p>
<p>Since we did not frame the question ahead of time, specifying that we are only looking at students #1 and #2, the odds are about the entire class.</p>
<p>Eric, of course you can find people who selected the same wrong answer on the test (or have the same birthday). But the odds are much longer if the test question was answered similarly by only a few others (or birthday happens to be Feb. 29th, an “unusual” birthday), and that the two must also share lots of other equivalent wrong answers (or unusual physical traits from a finite, preselected group of uncommon traits, such as red hair, left-handedness, etc.). Sooner or later it becomes legitimate to ask whether they are twins.</p>
<p>Quant, a statistician may not be necessary if there are laws in place that already define what statistically defines cheating, or if there is something in the millions of papers signed to take courses that has you agreeing to some definition. If there is not, the attorney can decide whether a statistician is necessary, and what direction to take to resolve this case the best that it can. </p>
<p>Legally, it is often immaterial whether someone actually did something or not. What matters is if it can be proved. Very sad, when someone is truly innocent or guilty and the technicalities get in the way. In this case, I have doubts that if it can be proven, and if that is the case, the accusation and ordeal should not have been cast. There have been a number of occasions in my life, in anyone’s lives, that we are sure of what happened, and yet cannot accuse someone of something because there is no way of proving it. I know who stole a cell phone from a tournament, but there is no way to prove it, for instance. In some cases, EVERYONE in the group knows who did something, but again, there is not sufficient proof to levy an accusation. That is the way, it should be, because accusing someone who is innocent is a terrible, terrible thing.</p>
<p>My son got in trouble in high school for something he did not do. It came down to one guy’s word against the other. Since both were involved and knew what was done, they were both punished, deservedly so. My son accepted the consequences without complaining whereas the other young man and his parents kicked up a big ruckus to the point that the school actually considered taking the easy way out and hanging it all on the one who just took the punishment. That is when I stepped in as a parent.</p>
<p>Even then, it would have been a rough go, except the stupid kid could not leave it at the one incident and listed a number of other acts of mischief. theft and vandalism that had occurred over the school year and attributed them to my son as well. That was a big mistake on his part, because my son could not possibly have done those things as he was not present at the crucial times. He had been active in something that required him to just show up for occasions and leave right away, so that he was in full view of everyone. In other words, he had a full alibi. </p>
<p>Still, it required some legal intervention to get the school to get off of its procedural high horse and look at the facts. They were headed in one direction and the momentum was such that taking a turn was not easy for them to do. Sad, but quite true. At that point they did not care who did what. It was who could be hung for it. Had happened before as well. Yes, we had to pay an attorney, but that did resolve the situation very quickly, in that it got their attention to what the facts were instead of how far down the path they had already come in a matter.</p>
<p>Don’t police use a technique whereby the proposed perp is put in a line of say, 9 other people, and witnesses are asked to finger the criminal ? The chance that 3 witnesses all finger the same person by chance is 10^3. Our cheating probability is about 1,000,000 times that. That is, 9 witnesses would all have to pick the same perp by chance.</p>
<p>I’m reasonably sure that courts do not require 9 witnesses to reach the burden of proof.</p>
<p>We honestly cannot tell if the OP S is a innocent bystander or involved somehow. We can’t. She assures us that he wasn’t involved, and she says that he says he wasn’t involved. But in all honesty, we just don’t know.</p>
<p>When dishonest students use good students papers as answer keys, the good student shouldn’t be punished if they have no involvement in the cheating.</p>
<p>But if the two worked together, they should both be punished. </p>
<p>We can observe that cheating happened based on the information presented.</p>
<p>But we can’t know who did the cheating unless we have more specifics like room layout, computer security, or information from the other student.</p>
<p>^^
I agree with this, but from here on out, the burden of proof is on the student to show he was not complicit. This is why I keep focusing on figuring out the method of cheating used.</p>
<p>
</p>
<p>Yes. (10 char)</p>
<p>
</p>
<p>Text mesaging. Instant messaging. Sharing desktops is real option through instant messaging, although I think that would be a little obvious. Foot tapping, pencil tapping are old tricks. </p>
<p>Or the standard look at each other’s screens because the proctor is busy doing something else. </p>
<p>Text mesaging seems the most likely. Perhaps the student used a different phone than his own.</p>
<p>Bigtrees and I are in rare agreement here. From the information presented, we know that someone cheated, but we don’t know if the OP was involved knowingly, or was the victim of a cheater.</p>
<p>However, the school has more information that we do about the specifics. We cannot dismiss the claim that the other student was able to see the OP’s computer screen-- but maybe the school knows that’s impossible. We don’t know anything about the computer security, but the school does. (The facts presented don’t look particularly likely for a breach of computer security, though. If the other student broke into the computer, why didn’t he pick better answers to the test? And why didn’t he change his grade on the test where he wasn’t in the room with the OP’s son?)</p>
<p>Yes, it is probably the case that, based on the statistical evidence, the burden of proof has shifted to the OP’s son. And how can he PROVE that he didn’t cheat? if he knows anything he’s not going to give it up, and if he did he might be deemed complicit. The other student could fess up, but why would he do that? He took an offer that apparently left him unscathed regarding his college plans. What would be his motivation for jeopardizing that? Very difficult situation, and personally I’d be looking for a compromise you can live with. Likely there is going to be a very tough life lesson learned here.</p>
<p>
</p>
<p>Oh, perhaps because there is an attorney that is inquiring because a family is considering a lawsuit over the issue. The family would be willing to let the whole thing go if the other student can explain what happened and clear the S from involvement…</p>
<p>Or perhaps simply because it is the right thing to do.</p>
<p>
So basically, any time someone figures out a way to cheat off of me he can get me in trouble as well, and there’s nothing I can do about it unless I can prove I wasn’t collaborating? How exactly would I go about doing that? </p>
<p>I understand the urgency of discouraging cheating, but not to the point where it trumps fundamental principles of justice.</p>
<p>In my opinion, the consequences of being punished for cheating are so severe that the standard of proof should be that for a criminal conviction: beyond a reasonable doubt. This means that even if it is more likely than not that a person cheated, that isn’t enough. In this case, given what we’ve heard, it does appear that it’s more likely than not that cheating occurred, but I don’t think we’ve seen evidence beyond a reasonable doubt that the OP’s son cheated (maybe for the other kid; that’s a lot closer). What the standard of proof actually is at the university in question is a different matter.</p>