What is Willpower?
You know what willpower feels like. A euphoric sense of motivation overcomes you, distractions crumble as they hit the brick wall of your focus, hard work becomes easy. It feels good. But, this feeling, this ability, this state of mind, is impossible to secure. Day in and day out you work. Some days you feel fresh and can easily push your mind and your body to its limits. Other days, you feel sluggish and wonder why you continue to push yourself. The variation seems arbitrary.
But, it isn’t.
A year ago I began tracking eight different activities in my life I wanted to devote myself to. Tracking these activities made them easier, measurable, and fun. Some of these activities, like playing piano, came easily. I enjoyed them, accomplishing them was a breeze. Others were not. Activities like homework and meditating were difficult, often to the point of being unenjoyable. Finally, all these aspects were factored into a productivity score measuring my productivity for the day. My productivity score is on a scale where five is average, ten is double average, and so on (to learn more about this system, check out this article). Now, with almost a year of data, I’ve been able to solve, at least for myself, the problem of willpower.
Will power is momentum. When we exercise willpower, our ability to motivate ourselves only increases. When we fail to use willpower, it’s hard to get back on track. Think of it like a bike. Once you get momentum, balance is easy. But, as you slow down, your balance diminishes as you inevitably wobble and tip.
To gain momentum you must begin. With willpower, this can take many forms. We all know the practice of choosing a goal each year for a New Year’s resolution. Although the position of the earth around the sun has no influence on our ability to set goals, New Year’s resolutions work. The power doesn’t lie in any celestial alignment, but rather in beginnings. Beginnings are clean slate. Beginnings are a starting block. Beginnings are powerful.
This can be seen every week. At the end of each week we’re given a chance to restart, a chance to start off on the right track. The first couple days of the week are essential. After a year of collecting data, I could prove this.
The trend is apparent. If I have a strong Monday and Tuesday, the rest of the week will follow. A R² value of almost .6 is really strong, especially with human behaviour. R² values measure the percentage of the variation that is explained by a linear model. A higher R² value means a higher correlation in the data. So, for every point in productivity, it will translate to two-thirds of a point of productivity in the rest of the week. You might be thinking, sure, but this just shows people are consistent. Obviously if you have two good days in a row, the next five will go well. However, once you get to the end of the week, a high productivity has little effect on the next five or six days.
Saturday’s correlation with the rest of the week only has a R² value of .06, showing almost zero correlation. Even if the correlation holds, a single productivity point on Saturday will only translate to one-fifth of a point for the rest of the week. Saturday isn’t unique. From Thursday to Sunday, the R² scores range from .04 to .24, far less predictive than Monday and Tuesday. Beginnings matter. By working hard at the beginning of the week, I can ensure that the rest of the week will be productive. Once you start biking it becomes easier to stay upright, but to get moving you must work hard.
The momentum of willpower extends beyond the beginning of a week. We know the feeling of being on a productive streak. Unfortunately, we also know what a rut feels like. We can change our productivity, but it constantly wants to stabilize. The graph below represents my productivity from the past year.
My productivity constantly bounces around day to day, but behind the noise, trends appear. I maintain stretches of strong productivity, sometimes lasting months, that are interspersed by defined slumps. Once you establish the habit of productivity, it’s easy to maintain. Getting there is the battle. This consistency is apparent in the span of a year, but exists day-to-day as well. Although the day-to-day productivity fluctuates wildly, there is meaning to the madness.
Although it’s not a single trail of data forming a pretty trendline, the correlation is undeniable. Although each day offers a new start, the previous day’s productivity will influence the next day. Each additional point in productivity will create an additional half a point in productivity the next day. Also, a R² value of .35 is nothing to scoff at, considering the natural variance in our lives.
Willpower in a day
All of the data displayed above is interesting, but not very helpful. If all it’s saying is to be productive so you can be productive. How will that help me with my willpower? The answer lies in what we devote ourselves to. The productivity score is comprised of eight different activities that I want to make a part of my life. So in addition to a productivity score, I know my daily productivity in each of these activities. To check the correlation between an activity and my overall productivity, I used a productivity score that didn’t factor the activity’s productivity. Otherwise a high productivity in one activity would obviously increase my overall productivity because that’s how I built the system. Instead, that given activity is eliminated from the productivity score.
There’s this pervasive conception that willpower is limited. Willpower is a bowl. Every time we do something that requires willpower, it fills up a little bit. If we keep exerting willpower, the bowl will overflow and we cannot will ourselves to act until the bowl empties. This view is both aggressively pessimistic and untrue, at least for me. If this were true, using willpower on an activity would lead to less willpower for other activities. We would see that when a certain activity’s productivity score is high, the day’s productivity score without the activity would decrease. This doesn’t happen. Not only is there never a negative correlation, the correlation is either positive or non-existent.
Practicing Piano, Completing Homework, Reading, and Reducing Distractions Online
These four activities, although different, share similarities. They either take very little willpower, or are things I’m forced to do and have little choice in. For reading and practicing piano it’s the former, I enjoy doing it, so the willpower involved is minimal. With homework and my “Online Productivity” things are different. Although they take more willpower, I am forced to do homework. To measure online productivity I use the app RescueTime, but since I use an internet blocker, StayFocusd, it also doesn’t require much choice. With all four of these aspects, they don’t force me to make hard decisions or completely exert my willpower. So, how do they influence my day’s productivity when you take the activity itself out of the calculation.
Note: On the graphs, the X-Axis represents my completion of the given activity. A score of 1 is average, 2 is double average, and so on. The Y-Axis is the productivity score when the activities score is taken out of the calculation.
Although there might be some correlation between these activities and being productive, it is minimal. All of the R² values are below .1 which is abysmal, even for human behavior. All of the activities I complete throughout the day, so it’s also hard to show whether me being productive makes me read more, or reading more makes me more productive. It’s impossible to know. Although these activities might contribute to being more productive, they have little effect on a given day. This makes sense, practicing piano and reading require little willpower. Although completing homework and staying focused online might require willpower, on any given day they are out of my control.
Working Out, Meditation, Completing to-do-list, and Simple Self Care
Unlike the four activities above, these activities boast closer correlation to a productive day. The first is working out. It has modest correlation and likely less causation. Here the R² value is just under .2, significant enough to show semblance of correlation, but relatively minimal. More importantly, since I exclusively workout at night, the amount working out is helping me on a given day is limited to the late night. Although there’s correlation, productivity might force me to workout rather than working out forcing me to be productive
Next is meditation. With an R² value of over .3, it poses a more serious claim of correlation. Since most of my time spent meditating is before breakfast, there is likely a causal link as well. This makes some sense, meditation is often heralded as an activity to increase concentration and reduce stress. If the correlation holds, completing an average amount of meditation will increase an average day’s productivity by one point, or 20%. It might not make a huge difference, but it’s significant
The completion of my To Do list offers strong, but unsurprising correlation. With an R² value of .44 there is a clear correlation between completing a To Do list and being productive. Furthermore, completing an average amount of my To Do list will increase a day’s productivity by over two and a half points, 50% of an average day. This isn’t surprising. For one, the to do list requires that I read 20 pages every day, so some of this category bleeds into the reading category. Also, by completing things on my list I am forced to exert willpower, which gives me momentum.
This miscellaneous category was the most surprising. If you didn’t read the other article, this category measured simple stuff like brushing my teeth, wearing my retainer, showering, and checking my excel sheet. This posed an R² value of .4, and increased a day by over 2 points, or 40% of an average day. The correlation is strong and poses a serious benefit. Furthermore, since most of these tasks are done in the morning there isn’t any reverse causation. This surprised me. Brushing teeth and showering take no ability and little willpower. I can’t say why these small things correlate to an effective day, but they do.
All of these aspects has some level of positive correlation with a productive day. Some offer more correlation, and some have a more significant effect, but none of them show a negative correlation. Below is a table of the various activities’ R² values and slopes. The slope category represents how much doing an average amount of the given activity will increase overall productivity (where five is average).
Long Term Effects
But what if the effects of doing these activities have effects that last longer than a day. To measure this I looked at how a month’s worth of activity will influence productivity (once again factoring out the activity).
Reading and Piano
Once again reading and practicing piano did little to contribute to my productivity. Both R² values were around .1, and even if correlation existed, the contribution was only one additional point over the course of a month. Neither of these activities require much willpower. Instead, often when I’m feeling unmotivated, I’ll read or practice piano as a substitute for other activities. Although there’s no evidence of these activities hurting my overall productivity, it’d be dubious to claim they help it.
Homework, and Meditation
With these activities, although there was likely some long term correlation, the correlation was either limited, or the predictive factor was minimal.
Homework boasted a high predictive factor, but little correlation. The R² value was .33, not awful, but also not strong enough to be sure there was lots of correlation. However, doing an average amount of productivity would increase a month’s productivity by 4 points overall, a significant amount. The long-term correlation is stronger than the short term correlation mentioned earlier. Long periods spent exerting willpower are hard, so I gain momentum and other tasks become easier.
Meditation was the opposite. The correlation was stronger with an R² value of .58, however an average amount of meditating only increased my overall productivity by 1 point, or 20% of average. It’s hard to accurately judge the effect because there is one huge outlier where I meditated a lot, but was also more productive. Because of the outlier it’s hard to gauge the predictive effect because there are likely diminishing returns with meditating. Either way, it likely increases my productivity to some extent.
Online Productivity, Simple Self Care, and To Do List completion
These graphs look similar because the data is similar. All of these pose strong correlation and a decent predictive effect. With R² values ranging from .55 to .65, and slopes around 4, the effect is clear. Completing miscellaneous tasks and doing my To Do List contribute to productivity on a daily scale, but even more significantly throughout a month. Daily online productivity doesn’t have a huge effect on my overall productivity, however being productive online for prolonged periods of time has a more significant positive effect.
On the scale of a day, working out had little effect on the day’s productivity. Since I workout later in the day, much of the benefit might not be seen that day. However, on the scale of a month, things are different.
The correlation here is impossible to miss. With an R² value of almost .9, working out likely has strong correlation and some aspect of causation with overall productivity. Furthermore, working out an average amount in a month will increase the month’s productivity by 4 points, almost doubling it. This is something we might know anecdotally to be true. Working out is good for your health. But, at least in my case, it’s also good for willpower.
All these activities help our long term productivity, but some more than others. Focusing on what makes us best will take care of the rest. By working out, I increase my ability to read. By being productive online, I make meditating that much easier. Everything is connected.
Below are the R² values and slopes for the various activities’ contributions to productivity.
Why This Matters
Willpower may be limited. But, it’s not as limited as we think. There’s this idea that you can waste willpower on some activity, reducing your ability to complete other tasks. If this were true, doing one activity would limit the time and willpower for other activities, and the other activities would falter. For me, this couldn’t be farther from the truth. No activity I tracked reduced the rest of my productivity. Instead, with varying levels of correlation, they all increased my productivity, sometimes by a lot.
What can we learn from this? For me, I know I can brush my teeth and take showers to have a productive day. Or to increase general productivity throughout a month, I can workout more often. But, the specifics are different for everyone. Beyond these specifics lies a more important truth.
Being productive makes us productive.
On a large scale this is obvious. By exerting willpower one day, our ability to work hard the next day only increases. Having a productive beginning of the week sets us up for the rest of the week. We go through periods of significant productivity where work is effortless. We have slumps where every page, every pushup, every meeting, is a battle. Every day that we work hard makes it that much easier to work hard the next day. Every day we fail to achieve makes it that much harder the next day. We are creatures of habit, consistency is easy.
A single day is the same. Choosing to do things that require willpower, such as working out, meditating, and finishing a To Do List, have effects that spill over into the other activities. Exerting willpower doesn’t diminish willpower, but rather creates more ability to exert willpower. Every moment is a choice between the what is hard and what is easy. Every choice to skip out on brushing teeth, put a little less time into homework, or miss a workout has effects beyond the activity itself. Fortunately, the converse is true as well.
Working hard can make working hard easy.
If you’re curious about any of the data or methods mentioned in this article you should check out the other article I’ve been not so subtly linking constantly. Please email me with any questions, comments, or corrections at firstname.lastname@example.org.