Judge a coin toss

It is difficult to analyze a person’s thoughts through surveys.

In order to have enough curious visitors fill out the whole survey, there need to be very few questions asked.

Each question needs to be simple, yet very specific.

With such restrictions, I’ve failed multiple times, to get any useful data.
Until now.. An analysis on how people make choices with chance.

I gave this survey for people to fill out on Reddit: https://docs.google.com/forms/d/e/1FAIpQLSetGo6za0c0keg5n6mibPJmQn0WFOy3t5vHqhYi5QXzCdFfJA/viewform

The questions here were carefully engineered to look very simple, yet to ask a complicated question: How do people judge chance?

These series of questions go over various scenarios of people accepting or rejecting a bet for a coin toss.
In the first two questions, both Heads and Tails are portrayed as wins.
The last two questions, are meant to be portrayed as an obvious losses.

It was built this way for people to get used to and understand the series of questions, how they relate to each other.
It prepares them for the less simple questions, where there is no obvious answer.

An important factor for these questions is to assume there is only one toss. No retries.
$5 was chosen as the reward, due to it being a small enough that it didn’t significantly impact most people’s finances, yet its sufficient enough to buy a small snack.

I got back 125 responses.

As expected, the first two questions were answered Yes:



Next question, there is still no loss, most people still answered yes:


Now here come the more interesting scenarios and responses:


From an economics standpoint, although there exists a $1 loss, the $5 win should be sufficient for you to always say Yes.


Once again, the $5 win out-ways the the $2 loss, so this should also always be a “Yes”.


The $5 win continues to be better than a $3 loss, so from a rational economics standpoint, this should have still been a 100% Yes.


$5 win is still better than $4 loss, so this still should have been a 100% yes.


Here, $5 win and $5 loss is even. From a rational economics perspective, this should have been an even 50% Yes, 50% No. Even though you just get one try.

As expected, the last two questions were answered No:



This just gives a glimpse into how people think about chance.

The results however show me two things:

  1. Loss aversion
  2. Priming: By showing the previous available bets, people seem less likely to accept relatively worse bets, even though the bets are still good in absolute terms.

Unfortunately these results are inconsistent.. They vary depending on the audience answering the question. People more versed in economics provide answers closer to what is considered “rational”.

Although there is probably no way to predict how a specific individual would answer such questions, I wish there were a way to predict how people will answer as a whole.