Behaviorist Problems in Measurement

I was wondering if there was a way to measure data to possibly eliminate negative feedback loops that it could cause. I.e., I’m trying to incorporate behavioral design ideas into how one logs. Tl;dr at the bottom.

Most of the time, when someone tries to measure data relative to some goal they want to achieve, two feedback loops can occur. One is a positive one, where the data reflected is good and this gives positive reinforcement to the person gathering the data to continue doing so and to improve his/her own performance; the other is a negative loop where the data reflected indicates poor or decreasing performance and this punishes the person gathering the data and discourages them from continuing doing so.

Anecdotally, this has happened to myself. I do time logging (just to see where I spend my time in a day), and after some time I eventually labeled a certain group of items as ‘good’ and wanted to maximize them. So, whenever the time I spent on those things decreased or stabilized (which is inevitable, each day only has 24 hours, so there’s already a theoretical maximum at which I would plateau) the reinforcement and thus encouragement stopped and I lost interest in logging until a few weeks or months later.

I hear the same thing occurs with others and keeping track of their weight on the scale.

From the mind of an engineer, the negative aspects of the data are crucial. They too provide useful feedback. But from a behaviorist perspective, it would be beneficial to mitigate the punishment aspects in logging (c.f. Science and Human Behavior by Skinner, Chapter 12 ( http://www.bfskinner.org/BFSkinner/Society_files/Science_and_Human_Behavior.pdf )).

I suppose people’s first thought might be to simply “change one’s frame of reference”. In other words, to change how the person views the system of logging rather than the system of logging itself. However, if it is possible to change the system of logging itself to make it so such a therapeutic and pseudo-scientific wager would have to be taken, then I think that would simply be the easier way to go.

One simple way of overcoming the psychological effects of logging is to simply have some impartial third party log for you. E.g., I hire a guy to follow me around with a notepad and log what I’m spending my time on. However, this is typically somewhat impractical. Another way is to arrange things so that it’s impossible to go through the day without having logged. E.g., my friend attaches a scale to his bed (he has to go to sleep). However, this doesn’t get around the punishing effect the logging system is having. I.e., I would still feel incredibly discouraged towards accomplishing my goals, and my friend similarly. This would only solve the problem of being able to continue taking data.

I would think the most obvious way is to subtly change what is being measured. For example, your goal is to become a prolific writer, and so you start to take simple day-by-day word count data. Tallying these numbers has a different psychological effect than keeping a total overall word count level. Which has a different psychological effect than displaying the amount of words left to write to accomplish a daily writing goal. This is what I am getting at with respect to behaviorist design principles in taking data; and is what I’m hoping to get some suggestions on here.

Tl;dr:
-How can we design data acquisition to be less punishing?
-I think the punishing aspect is a crucial one to eliminate.
-I don’t think changing the person’s ‘internal outlook’ is the easiest possible way of solving the problem.
-Is it possible to solve this by finding ways to change what is being measured?

1 Like

http://quantifiedself.com/2012/04/thinking-about-side-effects-of-personal-informatics-systems/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+QuantifiedSelf+(The+Quantified+Self)

And now there’s a post about it…

As an update, one easy solution–although not complete–is to only take positively reinforcing data and ignore the rest. Or situate the data so that it only comes in a positively reinforcing format–i.e., like an EXP bar. For example, measuring total steps walked is better than measuring steps walked/day in terms of reinforcement.

http://quantifiedself.com/2012/04/thinking-about-side-effects-of-personal-informatics-systems/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+QuantifiedSelf+(The+Quantified+Self)

And now there’s a post about it…

As an update, one easy solution–although not complete–is to only take positively reinforcing data and ignore the rest. Or situate the data so that it only comes in a positively reinforcing format–i.e., like an EXP bar. For example, measuring total steps walked is better than measuring steps walked/day in terms of reinforcement.

I can imagine several strategies: set standards low (working for five minutes) so that early experiences are likely to be reinforcing; provide motivational frames prior to data entry that reduce punishment (normalizing difficulties); automate data collection. of course, self-monitoring is a reactive and effective behavior change tool, so that when people monitor, they tend to shift their behavior in the desired direction. but this would be provided that their expectations/goals are not set too high.
Another key strategy would be to embed them in some kind of interpersonal support, that can sustain them in the face of initial difficulty. as when people see a therapist, or more generally, joining some kind of mutual support organization/team wherein individuals keep one another motivated and problem solve obstacles.
Joe

To go for a mutual support sounds good because that can very well help in finding solutions in an easy way!