Magic Dots: Quasi-Reinforcement Helps Get Things Done

This photo illustrates a method I have used for many years to get work done, usually writing. Every six minutes of work, I make a dot or line. One hour = 10 marks = a box (counting method from Exploratory Data Analysis). I use a stopwatch. I make a mark when I am more than halfway to the goal. If I glance at the clock and it says 4 minutes (more than halfway to 6 minutes), I make a mark. If I glance at the clock and it says 10 minutes (more than halfway to 12 minutes from 6 minutes), I make a mark. I only zero the clock when I take a break. I use one piece of paper per day.

I devised this. It is based on an effect discovered by Allen Neuringer and Shin-Ho Chung called quasi-reinforcement. Neuringer and Chung studied pigeons. They found that if you give a pigeon food every 500 times it pecks a key, it will peck the key slowly (say, 2 pecks/minute). If you give the pigeon a brief flash of light every 20 pecks — a marker that shows it is doing the right thing to get food — it will peck much faster (say, 4 pecks/minute). The flashes of light are quasi-reinforcement, said Neuringer and Chung — they have some but not all of the properties of ordinary reinforcement, such as food. By themselves, the flashes of light don’t interest the pigeon. It won’t peck a key to get them. The amazing thing about this effect is that it doubles how hard the pigeon works without raising its salary.

I noticed improvement — it was easier to write — within about 20 minutes the first time I tried this. I chose six minutes as the unit because shorter times were more distracting and longer times less effective.

I told Gary Wolf about the dots method two years ago and he’s been using it ever since. He says it is good for getting started on something he needs to write. After he gets going, he stops doing it. He uses it as an example of the value of self-tracking. I too find that after I get going on something, I need it less. If I stop, however, I drift backwards toward doing less productive stuff or nothing.

Gary asked me about this a month ago and I started doing it again (instead of percentile feedback). I noticed something I had never noticed before, which was that the system lifted my whole energy level and gave me a “can’t wait to get started” feeling in the morning. This too made it easier to get stuff done. It reminded me of some rat research I’d done. Put a rat in a Skinner box and it will explore for a while. If it doesn’t get any food,  after a while (10 minutes?) it will stop exploring and curl up in the middle of the box. However, if I give the rat a pellet of food at random times (at the rate of one pellet/minute), it will keep exploring the box indefinitely. Learning psychologists have emphasized that when you reward an action, you make it more likely. The rat experiment I just described suggests a second effect: when you give reward — at least, when reward is rare — you make all actions more likely. You increase exploration, not just the rewarded response. When I was a young professor I went to a two-week neuroscience program at Dartmouth. It was all lectures. The other attendees were graduate students. I had little in common with them. There was little to do in the town, besides eat Ben & Jerry’s. The next town was 8 miles away. I couldn’t find anything I enjoyed doing. After a week, I had trouble getting out of bed, like the rat curled up in the middle of the Skinner box. A psychiatrist might have said I had major depression. I flew home and was fine.




17 Replies to “Magic Dots: Quasi-Reinforcement Helps Get Things Done”

  1. A couple years ago I came up with a similar timer-based strategy to handle household chores. I wrote a simple Windows app that, every 9 minutes, grabs a random item from my long to-do list, displays it, and beeps loudly (in case I’m not in the same room as the computer). I settled on 9 minutes because that was usually enough time to do the kind of task on my list (start a load of laundry, wash dishes, pay bills, order vitamins, brainstorm ways to improve my diet, etc.). I find it to be a very effective motivator – it makes even the most tedious tasks almost enjoyable when I know in 9 minutes I can drop it. The randomness is important, too. If I did things in the order I’d prefer to do them, the cat’s litter box would never get changed.

    I noticed the higher energy effect, too. But I always thought it was physiological, not psychological. Most of my tasks involve getting up and physically doing something, so I assumed the energy came from deeper breathing and getting some light exercise.

  2. I don’t see how these methods let you benefit from procrastination.

    (I was due to move office. I delayed (ill health). Finally I’ve been told they’ve found another solution and I needn’t move at all. Result!)

    Don’t do today what can be put off beyond tomorrow.

  3. I guess both systems are good. (the box, and percentile feedback) Your mind just became inured to the percentile feedback system and so it is not having the same exciting effect on your productivity. Maybe in a few months you will switch back to percentile feedback. Do you think?

    I notice the same thing with hobbies. No matter how much I love something, after I’ve done a lot of it it gets stale.

    This is too bad. I could create a menu of hobbies for myself and rotate among them to keep my excitement high every day, but it’s difficult to make progress at something when one must drop it and pick it back up frequently. Not to mention finding enough things that I actually really enjoy.

    Seth: Percentile reinforcement becomes hard for me when for whatever reason productivity goes down. It is hard to look at low scores all day.

  4. Dr. Roberts,

    How many days after starting up the dots method again, was it before you first had the can’t wait to get started feeling?

    Seth: about 2.

  5. MikeW, so you’ve been using the 9 minute random timer for a couple of years now, and it’s still as useful as when you started?

  6. by: yes, I’d say the random task timer is just as useful as ever, but I’ve never used it on a regular schedule. It goes in spurts. Lately I’ve had it running daily because I felt I wasn’t making enough headway on my to-do list.

  7. Andrea: I’ve seen a similar app available for Android and iPhone, called Job Jar. It has the random selection I want, but not the cracking-the-whip-every-9-minutes functionality. I suppose you could download that, then use a stopwatch app (or a real stopwatch) in conjunction with it.

    I’m a programmer, so I often cook up apps for my own use with no desire to distribute them commercially. I’ve thought about rewriting this Windows app to work on my Android phone, but haven’t put that on my to-do list yet, ha-ha. It would take more than 9 minutes!

  8. Seth, I’ve been using this type of secondary reinforcement for some of my pigeon and rat studies, too. It’s very useful when the subject must complete a set of responses before receiving its primary reinforcement.

    I’d say the reason why the rats were exploring the box throughout the session when you gave them random food deliveries is that you kept activating the general search mode (exploration) of the food reward system. This idea ties into the behavioral systems theory work of Staddon & Simlehaug, Shettleworth, and Timberlake. I doubt that ANY activity would increase (in fact, you state that sleeping curled up in a ball decreased), but certainly feeding-related activities (which includes exploring for more food) would increase. Self grooming might be another activity that would decrease (cf. Sara Shettleworth’s studies of hamsters in the 1970s).

  9. I gave the dots a first try. They didn’t motivate me enough that I could keep at the task for a full hour. There was confounding from the effects of coffee, though. I kept track of my motivation while using the dots, and it went down the whole time, but that was because caffeine was wearing off.

    I guess my point is the effect isn’t stronger than caffeine, but I’ll give it another go later.

    Seth: My first question is: do they motivate you at all? Meaning: do they make work easier?

  10. This first experiment didn’t give me any reason to think so. All I noticed was that I was looking at the clock to see if it were time to make another dot. I will try the system again tomorrow.

    The percentile feedback system I downloaded here though has increased my productivity about 20% since I started using it three weeks ago.

    Seth: Glad to hear it about the percentile feedback. About the dots: Their obvious value is to get started and gain momentum on difficult stuff. I don’t think you will notice any effect on stuff that isn’t hard to do.

  11. I tried again, and this time my enthusiasm for the work increased over the hour while doing the dots. I think it was just due to getting into the task, but the dots did give me an extra push to keep going.

    The stream of ratings of my enthusiasm was (18% before starting), 22%, 14%, 22%, 19%, 20%, 26%, 22%, 33%, 30%, 26%.

    Thanks for discussing this and the other productivity techniques. They’re fun to try.

    Seth: That’s a good idea, measuring enthusiasm on a numerical scale. I think I would use a 0-100 scale with 50 = neutral, >50 = want to do the task, < 50 = don't want to do the task. Maybe 60 = barely want to, 70 = somewhat want to, 75 = want to, 80 = quite want to, 90 = very much want to.

  12. Video game designers have been using this strategy for years now. I first noticed this playing the online multiplayer game “Call of Duty”. Even if you ended up with the worst score at the end of the game, you always received an accolade of some kind. This positive feedback always had me playing the game well past a reasonable bed-time.

  13. (re: Dr. Roberts’ numerical rating system) I’m very glad to hear how you interpret percentages and their corresponding feelings. To me, this is the hardest part of self-experimentation, coming up with a rigorous, reproducible scale.

    The percentages I used were “what popped into my head from my gut” when asking myself, so, how enthusiastic do I feel right now? I am chronically sleep deprived, so I think that’s why my numbers were so low.

    I see your system provides guideposts for keeping the ratings consistent.

Comments are closed.