The grading thread

[Parasaurolophus]

Managed 20 presentations, 5 exams, and 7 essays today. Should be able to finish the essays tomorrow.

[AvidReader]

Managed 20 presentations, 5 exams, and 7 essays today. Should be able to finish the essays tomorrow.

Yay! Well done!

AR.

[evil_physics_witchcraft]

I didn't meet my grading goal yesterday, so I need to make up for it today. :(

[Sun_Worshiper]

I'm grading research papers from my masters-level methods class. For this assignment each student had to collect some publicly available data and analyze it. Consistent with the class content and assignment expectations, each of the students ran a simple regression with three or four x-variables covering a single year of data. So far so good...

...Then I get to a student who has created a panel dataset covering ~40 years and a dozen variables. The analysis is way beyond what we have done in the class and the presentation is similar to what you would see in an academic journal: Journal ready regression tables, a host of robustness tests (models with unit and year fixed effects, dummies for key historical years, etc.). This analysis would be excellent for a PhD level class paper and would have a shot at a decent academic journal. Needless to say, I am quite suspicious that this student did not write this paper and instead had somebody to do it for them, but the plagiarism checker doesn't detect anything and I don't have any other evidence except that the paper is miles ahead of than anything we've done in the classroom (and the students' own work up to this point has been fine, but no indication that they have had advanced stats training).

So what should I do?

[traductio]

Ten papers into my goal of 19 (which will finish off the class and my semester!).

Weird plagiarism case this morning -- has anyone else seen this? The student cut and pasted from a webpage, rearranged a few words, and then tacked on a parenthetical citation for a different source. The different source was almost plausible, so I looked it up (despite the incomplete bibliographic reference). It did not in any way support the claims the student attributed to it, even though it was on an adjacent topic.

This seems like a lot of work to hide plagiarism. In fact, if the student hadn't added the spurious citation, I would have returned the paper saying "fix yr citations, ya scoundrel!" and left it at that. (That's my usual approach.) But this felt so much more deceptive and was consistent throughout the paper, once I started looking. If the student had just cited the source they pasted from, and if they had put as much effort into rephrasing as they put into finding plausible but spurious articles, they could have walked away with a B instead of a zero.

[arcturus]

Ten papers into my goal of 19 (which will finish off the class and my semester!).

Weird plagiarism case this morning -- has anyone else seen this? The student cut and pasted from a webpage, rearranged a few words, and then tacked on a parenthetical citation for a different source. The different source was almost plausible, so I looked it up (despite the incomplete bibliographic reference). It did not in any way support the claims the student attributed to it, even though it was on an adjacent topic.

This seems like a lot of work to hide plagiarism. In fact, if the student hadn't added the spurious citation, I would have returned the paper saying "fix yr citations, ya scoundrel!" and left it at that. (That's my usual approach.) But this felt so much more deceptive and was consistent throughout the paper, once I started looking. If the student had just cited the source they pasted from, and if they had put as much effort into rephrasing as they put into finding plausible but spurious articles, they could have walked away with a B instead of a zero.

Yes, I have seen this before. Students who have been through our (mandatory, if you have been caught once) how-to-identify-plagiarism workshop learn that they must cite sources. So, they copy-and-paste from one source and cite others to avoid the most obvious plagiarism detection.

I learned this through the example of a student who followed this approach for a very low stakes assignment in my class. Since it was a second case of reported plagiarism (and more egregious through this deceptive practice), for what could have been a 0 (for not handing it in) on an assignment worth 0.2% of the total grade (yes, less than a percentage point!), he risked expulsion from the University.

[arcturus]

I'm grading research papers from my masters-level methods class. For this assignment each student had to collect some publicly available data and analyze it. Consistent with the class content and assignment expectations, each of the students ran a simple regression with three or four x-variables covering a single year of data. So far so good...

...Then I get to a student who has created a panel dataset covering ~40 years and a dozen variables. The analysis is way beyond what we have done in the class and the presentation is similar to what you would see in an academic journal: Journal ready regression tables, a host of robustness tests (models with unit and year fixed effects, dummies for key historical years, etc.). This analysis would be excellent for a PhD level class paper and would have a shot at a decent academic journal. Needless to say, I am quite suspicious that this student did not write this paper and instead had somebody to do it for them, but the plagiarism checker doesn't detect anything and I don't have any other evidence except that the paper is miles ahead of than anything we've done in the classroom (and the students' own work up to this point has been fine, but no indication that they have had advanced stats training).

So what should I do?

This is one of the reasons I appreciate my University's academic integrity process. We are required to have a meeting with the student *before* we file any paper work. These meetings do not have to be labelled as an academic integrity meeting in advance, so I do not have to accuse the student before getting additional facts. I usually just refer to wanting to meet about their recent work in my class. During the meeting I ask about their process of working on the paper or the assignment. Did they consult with anyone (I emphasize that in my classes I want students to work together, etc). Did they have any difficulties finding source material, etc. This usually provides a moment in the conversation where it is appropriate to ask the more detailed questions that would likely reveal cheating/plagiarism/etc. I try to make it as non-confrontational as possible. Most students confess to cheating/plagiarism/etc without much prompting. If I find that the evidence (or confession) indicates a violation of the University's academic integrity policy, I file a report, which includes a summary of our meeting and a record of the sanction imposed (no report needed if I decide there isn't sufficient evidence). Student's have the right to appeal, but I find that most just accept the results of our meeting.

[OneMoreYear]

Powered through final exams from class everyone hates, now on to the set of papers from class everyone hates.  Gonna need more chocolate, but unless someone truly crashes and burns (or someone panics and plagiarizes), everyone who is currently passing should still pass.
And one last individual competency exam for tomorrow!
I will never be so happy to be put a semester in the rear view as I will be with this semester (I think it's a combination of the specific courses and my general level of burn-out). 

[Parasaurolophus]

Marked around 20 exams today, plus my last eight essays. So: I'm all done with essays, barring a straggler or two! Yay! All that's left are the exams which get completed between now and the end of the exam period, plus all the participation marks.

[evil_physics_witchcraft]

I really wish I could get away with writing 'WTF' on some of these lab reports.

[the_geneticist]

I'm grading research papers from my masters-level methods class. For this assignment each student had to collect some publicly available data and analyze it. Consistent with the class content and assignment expectations, each of the students ran a simple regression with three or four x-variables covering a single year of data. So far so good...

...Then I get to a student who has created a panel dataset covering ~40 years and a dozen variables. The analysis is way beyond what we have done in the class and the presentation is similar to what you would see in an academic journal: Journal ready regression tables, a host of robustness tests (models with unit and year fixed effects, dummies for key historical years, etc.). This analysis would be excellent for a PhD level class paper and would have a shot at a decent academic journal. Needless to say, I am quite suspicious that this student did not write this paper and instead had somebody to do it for them, but the plagiarism checker doesn't detect anything and I don't have any other evidence except that the paper is miles ahead of than anything we've done in the classroom (and the students' own work up to this point has been fine, but no indication that they have had advanced stats training).

So what should I do?

Set up a meeting and ask them to walk you through their process.  Ask why they chose to use analysis [this type] instead of [other type].  Have them walk you through a figure or two.  That should make it really obvious if it's their own work or not.
Who knows, maybe they have passion for this particular topic & already knew a lot?  Chances are slim, but we've seen it before.
(honesty, my money would be on purchased paper, but I'm willing to be surprised).

[Harlow2]

Half done with Big Assignment 1.  Only 2 people who read and heard directions and did the opposite, but even they still managed to make their points adequately.

[ergative]

Welp--time to get on the final projects.

[Caracal]

I'm grading research papers from my masters-level methods class. For this assignment each student had to collect some publicly available data and analyze it. Consistent with the class content and assignment expectations, each of the students ran a simple regression with three or four x-variables covering a single year of data. So far so good...

...Then I get to a student who has created a panel dataset covering ~40 years and a dozen variables. The analysis is way beyond what we have done in the class and the presentation is similar to what you would see in an academic journal: Journal ready regression tables, a host of robustness tests (models with unit and year fixed effects, dummies for key historical years, etc.). This analysis would be excellent for a PhD level class paper and would have a shot at a decent academic journal. Needless to say, I am quite suspicious that this student did not write this paper and instead had somebody to do it for them, but the plagiarism checker doesn't detect anything and I don't have any other evidence except that the paper is miles ahead of than anything we've done in the classroom (and the students' own work up to this point has been fine, but no indication that they have had advanced stats training).

So what should I do?

Set up a meeting and ask them to walk you through their process.  Ask why they chose to use analysis [this type] instead of [other type].  Have them walk you through a figure or two.  That should make it really obvious if it's their own work or not.
Who knows, maybe they have passion for this particular topic & already knew a lot?  Chances are slim, but we've seen it before.
(honesty, my money would be on purchased paper, but I'm willing to be surprised).

I think the approach described is a good one because it avoids anything adversarial.

I had an extremely unpleasant experience in undergrad where a professor told me they suspected I hadn't written the paper and proceeded to ask me to define various words and terms I had used. I had written the paper myself and I knew what the terms meant, but I didn't really have a lot of experience talking about my work at that stage and I was flustered and he didn't think my answers indicated I knew what the paper was about.

I think you can avoid this by following the suggestions above and just make sure you keep things friendly and open ended. For example, if you just started with "I was really surprised by how advanced the stats work was, have you taken classes on this stuff before?" you might well get an answer from a student who didn't do anything wrong that would alleviate all your concerns. If the student's response to that doesn't make any sense you could move on to asking them to take you through the process.

The one thing to keep in mind is that talking about your work is really a skill most of us hone in grad school. Undergrads don't really do that much of it, so you really want to be careful to not assume that just because a student doesn't do a great job of explaining their process that must mean they didn't do the work. Basically, I think you have to have a very high bar before you would want to make any accusations or take anything further. If the student is clearly completely unaware of what is in their paper or can't provide even the most basic explanations of their work, you have grounds to go further.

[Parasaurolophus]

Today I count up "participation" marks. Dreary, but easy. Plus whatever exams have come in.