Papers we’re reading (week 1)

We’re trying something new. We’ll pick 3 papers our team has read and do a quick summary + commentary on them. Let us know if you like this via twitter: @Irrationallabs

— — — —

Summaries this week by Kristen Berman

‘Working’ Remotely? Selection, Treatment, and Market Provision of Remote Work (JMP).

This paper tells a story about remote work and productivity.

In this study, researchers look at results from a Fortune 500 retailer’s experience with remote work. For this retailer, they find that in pre-covid conditions, remote work had an adverse selection problem. Less productive workers opt for remote work.

How did they find this out? They looked at workers who self-selected into remote work prior to COVID and compare their productivity with the productivity of all workers in COVID times. They summarize: “formerly on-site workers productivity rose by 8–10% relative to their already remote peers.”

Takeaways? First, they provide evidence that remote work can improve productivity. Both groups increased productivity when moving to remote work. Second, (and this is our interpretation, not theirs) it’s possible one reason remote work wasn’t more popular before was its signaling power. People may not have opted into remote work because it was seen as — and actually was — a signal that you’re less productive. It probably wouldn’t serve a star performer to associate with remote working, as one may incorrectly associate their performance potential with their choice to do remote work. Now that remote work has removed its stigma, we may be more likely to scale it.

Emma Harrington and Natalia Emanuel. Working Paper. “‘Working’ Remotely? Selection, Treatment, and Market Provision of Remote Work (JMP)”.

— — — —

Effectiveness of Financial Incentives in a Worksite Diabetes Prevention Program.

If you start paying people, you should keep paying people — otherwise behavior change effects likely won’t sustain. This study succeeded in helping people maintain weight loss, because they kept paying people.

Many diabetes studies pay people for three to four months, stop paying people, and then do a follow up two months later. People regain weight after the money stops. (Hot tip: skip to their discussion section to read a nice overview.)

This study put all participants into the Diabetes Prevention Program. Some people were paid up to $160 ($10 per pound) for 4 months, with a weight loss goal of 11 pounds. They were then eligible for a bonus of $100 to keep the weight off for the following 3 months. And it worked. People who received financial incentives lose an average of 5 pounds more than people who didn’t get paid. This is an impressive result. However, the researchers didn’t complete the punchline — what happens after they took away the incentives. Likewise, the takeaway is still: If you pay people, keep paying them.

Of note, 50% of people voluntarily dropped out of the study. This could be because they were fired or re-located — the employer went through a big re-org during the study. But it should also raise some flags. One, going through the employer to drive health outcomes can be tough if jobs are not stable. Two, the diabetes prevention program and/or the financial incentives, may not be appealing enough to maintain engagement.

Faghri PD, Li R. Effectiveness of Financial Incentives in a Worksite Diabetes Prevention Program. Open Obes J. 2014;6:1–12.

— — —

The Effect of Patient Reminders and Gas Station Gift Cards on Patient Adherence to Testing Guidelines for Diabetes.

Financial incentives tend to work when you’re incentivizing one-time behaviors. Credit card companies know this — they pay us to switch. Chase pays $300 to get your debt.

In 2005, the University of Wisconsin (UW) Health tried this. They implemented a program targeted to people with diabetes aimed at improving rates of receiving HbA1c and LDL-C screenings.

Researchers sent half the people a reminder to do a HbA1c or LDL-C screening (you’re overdue!) with a $6 gas card coupon. And it worked — on average, the target patient population received about two-thirds more HbA1c screenings than the matched comparison group, and a smaller proportion had no screenings at all (14.9% vs 23.5%). Note that this study uses quasi-experimental methods (propensity score matching) and it doesn’t isolate the difference between a reminder and a $6 incentive.

Austin, S., Wolfe, B., PhD., The Effect of Patient Reminders and Gas Station Gift Cards on Patient Adherence to Testing Guidelines for Diabetes. Wisconsin Medical Journal 110no1, Feb 2011

Like behavioral science?

Want to learn more about applying behavioral economics to product design? Irrational labs has an immersive online 8 week bootcamp. We teach you how to design and build products that change behavior. bebootcamp.com