Once again... it didn't quite work out as I planned

Another semester in the books. Another set of mixed results.

I honestly don't know why I ever expect anything other than mixed results... my hopes and expectations only occasionally match what students produce and it's been that way since 1976. Fifty years should be enough to make it clear.

This spring I tried a new strategy with my CSU online Strategic Writing & Communication class, where I created a methodical "scaffolding," forcing each student to complete key steps in the process of researching and writing a recommendation report. The theory was that successfully completing those several steps (from a topic proposal to an annotated bibliography to an "executive summary" and more) would make finalizing the report relatively painless. 

I also added a final memo assignment, where students were asked to reflect on the process and how it all worked. There was a section asking about AI being used as a tool in the course, as well.

Interestingly, nearly everyone in the course commented on how they liked the "scaffolding" and discovered so much about the complexities of writing a report. They almost all liked the fact that I did not categorically ban AI but instead asked for it to be used in specific ways, always with transparency.

So, that should mean that the final reports were better than they have been in the past, right?

Ha! Not right. 

Many of the reports ended up basically having very little "there" there. It's a researched recommendation report, focused on a "client" that each student could choose... with my hope being that even pretending to have a client would help sharpen the arguments and the advice. 

I am still analyzing how so many ended up earning 65-75 percent on what amounted to the final exam for the course... and still were able to earn an A-. 

But my first theory is that I created so many steps, each earning x points along the way, that even a dismal score of 60/100 didn't really tank their grades. Good for them, I suppose, but disappointing to me. My goal was for the final recommendation report to be the culmination and the clear evidence about what was learned in the course. 

Right now, I plan to spend some time before fall semester crafting one MORE assignment that builds to the final report, one that asks students to share a few paragraphs that include in-text citations gleaned from their research. If I had any revelation, it was that they simply did not realize that they can't just plop findings and writing from a source directly into an academic document. 

It's not that their writing made no sense. It's that it made sense without any clear support.

I suspect that AI is partly to blame, as students must have dumped at least some of their research into the chats and received logical but unsupported conclusions from the robots. And that led to lots of bullet point lists of positives and negatives and conclusions. Most of those, however, did not connect with the "client" they had chosen but remained frustratingly general.

So, I know the task.

A big question is whether I can keep this particular course ahead of the rapid improvements in generative AI... and whether it matter all that much.