Last Friday was my son’s lower elementary invention fair. And his project was resplendent – a powered K’nex conveyor belt designed to transport a picture of our aging Labrador, Gus, through a diorama of our living room (without the clutter). It glittered among the many cardboard boxes slashed and duct-taped into laser-shooting alarm clock robots and crocodile egg protectors. Leif was so proud, explaining again and again, “The problem was my dog has a hurt leg. I made a conveyor belt so he can get around the house.”

The thing is, I built it. Or at least most of it. And touring other kids’ homespun and authentic projects, I started to realize how profoundly I should not have. To discover exactly how much I had messed up my kid, I called Temple University psychology professor, Laurence Steinberg, author of The Ten Basic Principles of Good Parenting.

Here’s what he said: “I think the research is pretty clear that parents should not help with homework unless (1) the child asks for something specific that is beyond the child’s capability (asking the parent for help using a tool, for example), (2) the child does not understand what the homework assignment is (and you can explain it), or (3) the teacher has explicitly developed an assignment designed to have the parent and child work together on something.”

I also talked to Patricia Miller, SFSU professor and 2009-2010 president of the American Psychological Association division of developmental psychology, who said, “I think there’s HELP, and then there’s help. Based on developmental psychology research (and being a parent), I’d say that the best kind of parental help is ‘scaffolding’ — providing support in the form of prompts, hints, suggestions, reminders, etc. The child is more likely to learn something from that kind of help than from parents doing it themselves.”

When I tweeted the question of homework help, Sandra Aamodt, past editor of the journal Neuroscience and author of Welcome to Your Brain replied, “Costs of unfortunate results are low in kindergarten, will be higher later on. Let the kid figure it out.”

Points taken (not really).

Let me take a step back. I’m not necessarily a product-oriented parent. This year's Christmas cake was a drooping, kid-created "tree" inexpertly smeared with pseudo-marshmallow “fondant” (which, according to my wife, is a real thing, sometimes used decoratively on the Food Channel). The kids' Lego projects are as likely to be postmodern expressionist interpretations of "Death Star" as they are to be built from plans.

But for some reason homework is different. I want to help; I want Leif to do it well and right. And I think that by sitting with him and “helping” I can make it so. Because dedicated parents are the only group more determined than climate change deniers to cherry-pick science that supports our point of view, I went trolling through the literature this weekend looking for studies that might refute the experts. Here’s what I found:

A 2008 meta-analysis titled Parent Involvement in Homework: A Research Synthesis found “a stronger association [with higher achievement] for parent rule-setting compared with other involvement strategies.” And found that parental involvement with homework had “negative association for mathematics achievement but a positive association for verbal achievement outcomes.”

On aggregate that’s a swing and a miss for homework helping.

More: a study of 709 kids titled Homework in the Home found that “More parental support for autonomy was associated with higher standardized test scores, higher class grades, and more homework completed. More positive parent involvement was associated with lower test scores and lower class grades, especially for elementary school students. Student attitudes toward homework were unrelated to parenting style for homework.”

In other words, the best thing we can do as parents is to “support autonomy.” See Steinberg’s points above – answering specific questions or explaining the assignment is autonomy support. Strike two.

In fact, the only concrete support I could find for parents’ hands-on helping with homework is from a 2011 study of a structured intervention called Teachers Involve Parents in Schoolwork (TIPS). The program, tested at four urban elementary schools in which 70% of students qualified for free or reduced lunch, included a weekly assignment with “specific instructions for students to involve a family partner in a discussion, interview, experiment, or other interaction.”

Completion rates, attitudes about homework, and test scores all went up. But what is TIPS, really? Their discussion calls the program a “homework intervention designed to ease some homework tensions between students and families.” Rather than an adult sitting at a child’s shoulder and playing the role of teacher, TIPS asks students to proactively involve adults, much like a reporter interviewing an adult subject (who comes to expect, check up on, and sign the completion slip for this weekly project).

It’s “autonomy support” and Steinberg’s third point – composed of thoughtfully designed assignments that compartmentalize a student’s autonomy and a parent’s involvement. It was also the third strike against involved, teacher-like helping.

I still want to help. I still want Leif to succeed. And after this weekend, as hard as it is, I know what that has to look like. For daily assignments, I need to lovingly provide support for autonomy and then get the heck out of the room. And next science fair, I need to provide the time, cardboard, scissors and duct tape – and see what happens.