I was home for a few weeks, and amongst the old mail that I rifled through, there was a book mailed from my undergraduate economics department. And so I read the book, The Why Axis, by Uri Gneezy and John List.
The book is about using field experiments to generate insights. The authors described field experiments that they performed, and the results were fascinating.
Structurally, the book could have been tighter, more concise. Oftentimes, the author would repeat exactly what he said in the previous paragraph. And there were awkward transitions that made it more difficult to follow the train of thought. Also, the chapters were written inconsistently. Some chapters had detailed storytelling or tried to focus on impacted individuals, while other chapters were descriptions of the experiment and data. There were fluff sentences with vague pronouncements that lacked any insight.
The authors state that by observing the world, you can come up with many correlations. But determining causality requires an experiment, whether the experiment was accidentally or purposefully orchestrated.
One issue the authors look at is why men get paid more. They find that in patrilineal societies, men are more competitive and aggressive. The opposite is true in matrilineal societies. Also, they find men are more likely to negotiate salary unprompted. When a job listing states that salary is negotiable and the ambiguity is removed, men and women are equally likely to negotiate salary.
The authors say much of the bias today does not stem from hate, but the desire to make money. In a field experiment, they found that disabled people were given a 30% higher quote for a car repair because it is assumed they won’t want to go through the effort of collecting multiple quotes. If they tell the car repairman that they are receiving 3 quotes, then the repairman gives them the same quote as able people. In another experiment, they found that young black males wearing hoodies were the least likely to receive help when they asked for directions. In order to receive equal treatment, they had to dress better and wear business clothes.
The authors discuss experiments they performed for Chicago public schools. In one experiment, they confirmed that loss is more motivating than gain. One effective strategy to motivate test-taking students is to give them $20 before the test and have them write what they want to spend the money on, then let them keep the money if their test scores improve. That is better than telling students that they will receive $20 if their test scores improve. In another experiment, they wanted to see what would improve scores the most, giving financial rewards to students, parents, teachers, or a mix. They found that financial rewards improved performance in all cases, and rewards improved scores the most when they were given to any one group (such as just parents, or just teachers).
The authors discuss experiments that they performed for various charities. One question they had was what would result in more giving, saying that the goal was 0% reached, 33% reached, or 66% reached? They found saying the goal was already 66% reached was the most effective. Even though people would have to give less for the charity to meet its goal, they actually gave more, because having the goal partially met provided validation that the charitable cause was valid. Another interesting finding was that when there is a matching gift promotion, all matching gifts perform equally well, whether the match is $1 for each dollar you donate, or $2 for each dollar you donate. Again, the matching gift provides validation for the charitable cause, and the match multiple does not matter.
The authors urge individuals and businesses to experiment more. They find a few reasons why businesses do not experiment as much as they should. One is that managers want to validate their positions, and they fear using data-driven methods will invalidate their own methods or compromise the appearance of their expertise. A second reason is inertia, some businesses are slow to act. A third reason is managers are scared of uncertainty and change. They want to use familiar methods that have been satisfactory in the past.
For incentives to be effective, they must speak to people’s underlying motivations, lest unintended consequences occur. Field experiments are an excellent way of discovering these underlying motivations.