For the past several months, I’ve been reading the research on adults with autism. The work should help me prepare for the next book but, more importantly, help me do a little better in this current chapter of life.

In surveys of the research literature on adults with autism, several authors say there’s not much out there. There was even less research 20 years ago, so I’m not beating myself up for being late to this party.

When Sam was a toddler, I emptied library shelves, checking out books, looking for answers. Back then, autism research had barely exited the blame-the-mother stage and was focusing on young children. I found Maria Montessori’s original treatises and other general child development research and writings the most helpful. Over the years, I believe that concept—looking first to the big, fundamental ideas and then to the science that follows—has worked well for our family, which feels forever on the front lines.

For example, a recent study set out to create a sturdy vocational index for adults with autism. Why do we need such a thing? The researchers’ answers had a lot to do with shoring up future research and policymaking. But for the rest of us, who are watching our loved ones and their peers go after their employment and higher education goals right now, it can still help to be precise in describing current conditions and supports.

After all, the first step in solving a problem is identifying what it is.

In this particular study, the vocational index would score Sam’s work conditions and supports at the top, since he works full-time with the same support as any other warehouse employee. The index would score the work conditions and support of another young man we know just a little lower, because his work was part-time and he had additional support from a job coach.

That differentiation is a small step forward, but the rest of this young man’s story shows it also has its limits.

Like Sam, he has autism. Until recently, he was working on the retail floor at a pharmacy.

The pharmacy, which is part of a national chain, is participating in one of the state’s workforce programs. In addition, a former special education teacher served as his job coach. In the end, the job didn’t work out, and I’ll bet you, dear reader, already know why.

His family recognized something off in the support he was getting, but it wasn’t readily apparent what was wrong. After all, someone was there, someone whose job was to help him.

Ultimately, the job coach was meeting the pharmacy’s needs first, not the employee’s. The company needed workers. Joining the state’s workforce program allowed the company to tap a new pool of workers with little risk or investment on its part. And that showed.

As a former special education teacher, the job coach should know that a robust assessment of the worker’s skills and the workplace conditions comes first. Just based on our early experiences, I’ve got a pretty good idea how perfunctory that fellow’s assessment probably was. Sam’s first job placement was sacking groceries because that was all the state’s workforce program had to offer. They honestly didn’t look too hard at whether the job was a good fit.

It was clear that the pharmacy wrote up a task list long before any potential employee came through the program with their own strengths and skills to offer the store. Unsurprisingly, it can be a lot to ask some individuals with autism to respond to the shopping public. Sam says he couldn’t imagine doing it today. Some customers were already awful when he was sacking groceries years ago, and these days, there seem to be more awful customers and some just go off the rails with their complaints. So when this fellow’s job coach decided that he needed to pause the program and get some behavioral training instead, it was clear something else had gone off the rails.

This young man sometimes answers questions in long-winded ways, and some of the pharmacy staff and the customers didn’t like it. The coach didn’t either. We’ve all heard about the square peg that doesn’t fit in a round hole. We know the answer isn’t to send the peg out for sanding down, and down, and down. But that’s what was passing for job support for this fellow.

So the next question has to be, how do we measure support in the index, or how do we make sure what’s passing for support is actually support?