Positive reinforcement vs. punishment

Hi all,

I’ve been pondering something and I thought it would be interesting to know your point of view.

It’s about positive reinforcement vs. punishment.

Let’s say that, hypothetically, Beeminder had a mode where instead of being charged each time you fail to do a task, you pay a certain amount at the beginning. Then, each time you do the task, you get rewarded by getting part of that amount back. The end result is the same: if you mess up you end up losing money.

For example, I want to go make a difficult but important call once per week. Using Beeminder I have to pay $10 each time I don’t do it. But what if instead of this I paid $40 every month in advance and then I got $10 back every week that I do the call?

Wouldn’t it feel as if I’m getting a prize for getting things done (even if it was originally my own money)? And wouldn’t this work better than the punishment method? Findings on psychology - as far as I know - seem to indicate that positive reinforcement is more effective than punishment.

Any thoughts on this?

6 Likes

Only a sample size of 1, but one of my TaskRatchet users wanted to do it this way so we did this experiment. They didn’t stick with the service, so I took that as evidence that it wasn’t very effective. :man_shrugging:

4 Likes

I started to reply that you’re totally right but… and then had so many buts that I’m changing my answer!

(Btw, @aleix is using these psych terms perfectly but others may find our old blog post, Negative Reinforcement ≠ Punishment, edifying.)

So I think the part that’s right is that positive reinforcement works better than punishment. But punishment also works amazingly. Loss aversion is powerful. And more to the point, even if you prefer positive reinforcement, where do the rewards come from? Paying the money up front and getting it back unless you derail is a trick – it’s equivalent to getting stung.

At least for me personally, the equivalency would always be at the back of my mind and bug me. Also I like having scary high pledges on some goals and then it would feel especially unreasonable to pay up front. Also-also, most goals are like “get 10k steps (or work 40 hours, or practice piano for half an hour or whatever) per day forever”. It feels inefficient to have money always flowing back and forth for such goals and really muddies the mental accounting as well in terms of how much you’re really paying Beeminder for the motivation it’s giving you.

(Not to mention the laws and accounting involved. We’d be kind of a bank and have revenue that wouldn’t count as revenue. I assume this part would be perfectly overcomeable if we were convinced the psychology / behavioral economics were right though.)

Tangentially related to all this, I have blog post draft called “Paying Is Not Punishment” (see Point-Counterpoint: a possible new blog series) so I need to think about all this more.

4 Likes

I’m personally fine with playing tricks with your “lesser half” of you mind.
In fact, Beeminder’s loss aversion concept is another “trick” anyway. All my goals are normally at 5$ max. Will my financial status change if I derail on a single (or ten) goals during a month? Not at all. But still, $5 is enough for me to keep me on track :smiley:

5 Likes

(Finally found made time to weigh in on this, which is a topic I have active thinkings around.)

I wrestle with the punishment aspect of Beeminder all the time—or, rather, with finding a way to harness it that works for my particular brain. I joined Beeminder in late 2014 and have been more active and less active (to the point of non-active) in fairly predictable cycles:

  1. I create a spate of sensible goals that I truly want to achieve.
  2. I spend several weeks to a few months dispatching beemergencies and building buffer to lower the number of beemergencies.
  3. I start feeling like life is one big scheduling time bomb bearing down on me.
  4. I archive or flat-road most or all of my goals and go off to regroup.

It isn’t that Beeminder doesn’t work—it works perfectly. It gets my butt going on things I want to do but would never do without Beeminder. It’s just that for my personality, which tends to overdelivery, unnecessary perfectionism, and harshing on myself for failure, the very fact of Beeminder working turns into a heavy burden.

Then I have another set of data, from using a tool called StepBet in 2019. StepBet is a one-trick pony: you can make yourself walk more steps with it, and that’s it. But it’s a great trick, and my experience was radically different than my experience using Beeminder. Let me explain.

In StepBet, you put $40 into a collective pot with umptyhundred other people and promise to make your personal target steps every single day for six weeks. If you miss even one day, your money’s gone. But if you make it, you get your $40 back, plus a cut of the money lost by everyone who didn’t make it. This makes it a combination of positive punishment (do the behavior or you’ll pay money—just as with Beeminder) and positive reinforcement (do the behavior and you’ll be rewarded with extra money).

Turns out that this method makes all the difference in my experience of having to do the thing I said I wanted to do. It is super, super motivating in a way that “do it or you’ll pay” eventually ends up feeling oppressive or suffocating to me. I think there are two aspects to this:

  • I’ve already paid the money up front, so it doesn’t feel like punishment; it feels like reward. (This is directly related to your point, @aleix.)
  • The prospect of earning a real reward if I stick it out the whole six weeks makes it feel like a game I can win.

There’s a net feeling of having accomplished something that I don’t get from just avoiding paying money at the moment I slip up. In my brain, every StepBet day completed feels like progress (I’m closer to getting $40+ dollars!); every Beeminder day I complete feels like maintaining (still alive, but it’ll probably kill me in the morning [0]).

Now that I’m thinking out loud: perhaps the presence of a short-term end date is a factor—StepBet always ends in a few weeks, while all my Beeminder goals are forever goals. Perhaps if I set six-week Beeminder goals, the sense of treading water would become a sense of forward progress. Worth more thought! Because I love Beeminder, and have no plan to abandon it ever. But consistently feeling more like I’m gaining altitude and less like I’m putting out fires would be a welcome change. [1]


[0] That’s just a Princess Bride reference, for those who are now alarmed at how seriously I’m taking my Beeminder goals…

[1] The QS aspect of long-term graphs does feel like a reward, though, and I’m already trying to use that to max effect.

7 Likes

Maybe try never-ending goals but with the slope set to zero after six weeks until you feel ready to reset it for another six weeks?

:100: :smiley:

3 Likes

Is there any info about typical payout available on their page or somewhere else? I looked at it in past and I classified it as scam as there was no info anywhere about how much they actually transfer. And therefore I expected 20% to fairly lose your 40$, 70% chance to win back 40$ and 0.05$ prize. And 10% of “sorry, we transfer back money only in USA and therefore your money is now stolen”.

What I guess is yet another problem with such more complicated schemas, Beeminder is clear with “we will take your money, you can buy premium plan to configure this in a more elaborate way”.

BTW, it seems that they are multi-trick pony - DietBet | Motivation to get healthy WayBetter WayBetter are linked in their footer

2 Likes

It’s definitely not a scam. I ended up earning more than € 300 in the 12 months I participated (that’s true profit, after deducting my pay-ins and the $50 membership fee that lets you do three games at a time). Not a day job, but a nice bit of extra money, for sure.

On Reddit there are some detailed spreadsheets people have put together showing payout information, such as this single-person one and this one that several people contributed to.

There’s no chance involved in whether you lose your money: if you make all your steps, you get all your money back, so you’re in full control of that. You have no control over whether you earn anything above that, because that depends on whether other people make all their steps.

Happily, no! I live in Europe and there was never a problem.

Yep, WayBetter is the company behind StepBet and they do have other apps, but each one is a standalone one-trick pony. I was going to say that would make WayBetter a three-trick pony, but actually I guess they’re a one-trick-pony breeder… :slight_smile:

5 Likes

@alys, thank you for this! You’ve set off a chain of thinking that I will report back on once it’s matured.

4 Likes

Hey Aleix,

That’s definately an interesting idea worth considering, but have no thoughts to add for now.

Mainly wanted to reply as I saw your username and discovered that another Catalan (or at least someone that lives in Barcelona) uses Beeminder. So cool, benvingut! :slight_smile:

1 Like

I’ve used StepBet and DietBet a couple of years ago, and got paid when I won.

3 Likes

Ha, hola Marc! Thank you for the warm welcome :slight_smile:
I’m from Catalonia but I haven’t been much there during the last years. And I’m currently living in Prague.

Back to the topic, I’m still pondering about it and considering different ways it could be applied. I’m reorganizing my thoughts and I’ll post soon with some ideas. I love to see all these different points of view from you all.

1 Like

You and Alys are both very good at that. I think every comment either of you make on any forum post increases the quality of the discourse markedly. Hugely grateful to have you two active in here! Even (especially?) if you’re down on some aspect of Beeminder, like in this thread, it feels constructive and even inspiring.

And huge thanks to @aleix for kicking this off, too!

Back to the topic at hand, I’m generally a big fan of open-ended goals but now thinking about how to have the best of both worlds… I definitely like @alys’s idea of frequently scheduling long breaks to avoid burnout and think that ending a goal altogether until you feel ready to restart it is too dangerous. You start out thinking you’ll restart in some number of days or weeks and that’s a slippery slope to months or years to never. Beeminder’s Achilles’ heel is how to get yourself to create the goal in the first place. So when you’ve overcome that I think you should make sure to never expose yourself to it again. Schedule a break for as long as you want – a year even, if you’re ok with that – but always make (eventual) ever-increasing awesomeness the path of least resistance.

3 Likes

It’s great to see such varied opinions on this! Thank you for pouring out your thoughts.

I was already writing a mini-essay to share here with my theories about this… But then I thought “You know what? I’ll rather try it in real life!”

Since my project CheckWise is in “beta” mode (and I’m still accepting new beta tester users), I will propose to a few users to try out these ideas.

Here’s how: What if the users paid a certain amount of money upfront each week, and by the end of the week they get back a % proportional to the % of tasks they have completed?

(For those who don’t know it, in CheckWise the users tell me their to-dos and they have to prove to me (a real person) that they have done it, usually with photos.)

For example, you pay $30 upfront. That’s your “bet” for that week*. During that week you plan a total of 20 tasks, and you prove that you’ve accomplished 15 of them. That’s a 75% completion rate, which means you would get back 75% of the initial bet: $22.50 of “reward”. If you had completed all your tasks, you would get back the full amount, $30 of reward.

This way you get the accountability of someone checking on you plus the illusion of positive reinforcement we’ve been commenting on this thread.

I think it would be best if the user decides the amount of money paid upfront. This way, the more serious the user is about getting things done, the higher the amount it would make sense for he/she to “bet”. $5, $50, $500, or more, it’s up to him/her.

(*I was initially thinking on monthly cycles, where the users would bet/pledge at the beginning of the month and get their “reward” at the end of it. But making it weekly instead means a more immediate sense of reward/punishment, making it all more vivid and effective.)

Any thoughts on this? How do you think this will work?

2 Likes

I guess I feel pessimistic but I don’t want to pooh-pooh it prematurely and it will be very exciting to be proven wrong so I’m on the edge of my seat to hear how the experiment goes, how it feels for people, etc.

I don’t think I’d want to do it that way personally, for reasons I described towards the top of this thread:

Basically, I don’t feel like my own psychology is such that it could feel like positive reinforcement. In my mind it would be the standard “do this or else pay up” but with extra fuzziness and inefficiency with money going back and forth.

But I am very weird / idiosyncratic (especially when it relates to money!) so don’t put much weight on that.

2 Likes

It’s fine, I’m glad to hear your point of view.

I’ll keep you updated on the results :slight_smile:

3 Likes

Hey @grayson - if you’re interested, I’m setting up a Beeminder-themed StepBet for early Feb. @dreev set up a private message thread to keep track of interested participants. I’ll try to add you and the others from this thread (no pressure to join the game tho :wink: ).

PS: love this:

:smile:

It’s live!!!

I just created a StepBet game (starts Mon Feb 8, 2021) and I want you to join! It’s a fun 6-week game where we’ll motivate each other to stay active and win some cash in the process. To join, download the StepBet app and type in game code BEEMIND or follow this link:
https://step.bet/iFM35ggrWcb

Hope to see you in the game! You will be able to join the game the same way as your friends by searching for it with that code on the Find Games page.

Have a great game!