Open Source and Overjustification: Turbo-Charging Your Contributions for Maximum Open Source Project Viability

How many Likes do you have? How many Upvotes? How many Stars? This seems to be the measured karma of 21st century life. If you run a GitHub project, you know about Stars. And there’s nothing wrong with Stars, but a star only takes a moment to click. What shows more community engagement is making a contribution to a project, even a small one. If you want to find a great open source project, look for one with a deep and wide contributor base. But how do you attract contributors to your project?

Open source contributors, for the most part, are volunteers. While they may contribute as part of a paying job, they usually contribute for no direct compensation. So what motivates them to contribute? We’re accustomed to analyzing behavior in a world where people do work in exchange for money. But open source contribution challenges this model. While that motivation may seem like a mystery on its face, it turns out that the social science of 50 years ago helps explain it.

Back then, seminal work was done about the efficacy of various rewards to control behavior. Researchers in this area distinguish between intrinsic motivation and extrinsic motivation. Intrinsic motivation is what we like to do when no one is keeping tabs: for example, I like doing crossword puzzles, just for the fun of it. Extrinsic motivation is provided by others, and can include wages, or even symbolic rewards like the ability to post about my streak to social media.

Rewards can Backfire

Seminal work in this area was done in the early 1970s by Edward Deci and Richard Ryan. Deci and Ryan developed the Self-Determination Theory of motivation, which questioned previously assumed effects of rewards on behavior.

Their core experiment was performed on a group of undergraduate students who already had a baseline interest in solving puzzles, which means they had intrinsic motivation. The experiment involved solving a Soma cube puzzle, which is like a 3-D Tetris. The reward in the experiment was a small amount of money.

In the third phase, all participants were again given free time to solve the puzzles, and researchers once again observed and recorded how much time each participant spent on the puzzles.

Session 1–Baseline: the experimenter asked all the students to complete the puzzles with no reward.
Session 2:
- Control Group: the experimenter asked the students to complete the puzzles with no reward.
- Experimental Group: the experimenter told the group they would receive $1 for each puzzle they completed.
Session 3–Evaluation: The experimenter then told the subjects he was leaving for a while and that they were free to do whatever they wanted while he was gone.

Deci and Ryan concluded that offering the external reward reduced the subjects’ intrinsic motivation once the external reward was removed. Extrinsic motivation could actually sour people’s feelings about the worth of an activity and undermine their intrinsic motivation.

“Good Players”

A later study examined whether this overjustification effect would be similar for symbolic rewards. A team from Stanford and University of Michigan designed a study involving 3-5 year old children drawing with felt-tip pens.

Experimental Group: The experimenter told the children they would receive a “Good Player Award,” colored 3×5 inch cards with the words “Good Player Award” and spaces for the child’s name and school engraved on the front next to a large gold star and a red ribbon.
Experimental Group 2: In this group, the experimenter did not tell the children about the reward until they had completed the drawing session.
Control Group. The experimenter did not tell the children about or give any reward.
Final: The children had a “free-play” session seven and 14 days after the experiment, and were offered the opportunity to draw with the markers.

The experimental group, in which the children were promised a reward for drawing, played significantly less during free play. Children who were surprised with a reward showed more interest during free play than this experimental group. The control group, which was promised no reward, and received none, showed more interest during free play than either experimental group. The experimenters concluded that a bargain for the activity decreased motivation, but an unexpected reward did not.

Deci and Ryan’s studies spawned many others over the years. The ones described above are two of many. Behavioral scientists use the term overjustification effect–to describe this phenomenon: rewarding people for doing things they enjoy makes them enjoy it less. Over time, most studies have confirmed that the overjustification effect exists in various settings. In our own lives, we have all experienced the motivation to seek intrinsic rewards. The feeling of satisfaction from solving a puzzle or helping a friend is something we all know.

More SDT

Deci and Ryan’s SDT theory of motivation, which they explored in later studies in 1985 and 2000, posits that humans (including contributors!) have three primary psychological needs :

Autonomy
Competence
Relatedness

These are fairly self-explanatory–probably because they have been part of our collective social awareness in the intervening years. Open source participation dovetails with all of these. Open source is a quintessential setting for autonomy; the free software movement was premised on decentralized collaboration and the marketplace of ideas. Having PRs accepted is a key indicator of competence for contributors–it is literally resume material. The area where most open source projects struggle the most, among these three, is relatedness. The way you communicate with your developer community is key to fulfilling this element of motivation.

Open Source Contributions

These basic findings should ring true with open source contributors and project managers. Although some open source contributions carry monetary rewards, most do not. But all of them carry symbolic rewards, because open source development takes place publicly, where everyone can see who participates and roughly how much. An extrinsic reward could include accepting a PR, getting added to the contributor list, or even a special shout-out in a README.MD.

Many people, over the years, have tried to solve the problem of open source robustness by compensating contributors.We all know that lack of contributor engagement leads to security problems and technical glitches. In classical economic theory, compensating contributors with money should work. Over the years, many people have devised all sorts of ways to keep contributors motivated: hiring maintainers via legal contracts, compensating them automatically in DAO structures, and creating foundations to pool resources, to name a few. And these have all worked, to a greater or lesser degree, in different cases. There is probably no one solution to this problem, because each project is different. But we would be well served to remember the lessons of overjustification when we create these paradigms.

So, here are a few ideas for compensating contributors in ways other than money–or even direct rewards of any kind. Behavioral science tells us that the efficacy of direct rewards will wither over time, and that symbolic rewards may be more effective than monetary ones.

License. This may seem obvious, but many contributors choose to contribute to projects under licenses like GPL because they want others to share their source code. The license for the project matters, but in my experience, it does not matter to all contributors; the new generation of developers is seldom dogmatic about licensing and often considers all open source licenses to be similar. Nevertheless, some contributors will count the source code sharing aspects of copyleft licenses to be a major intrinsic motivation. This bears on discussions of using a CLA for your project. (That’s a big topic, but for more, see here: Contributions and the Latest Generation)
Respectful Treatment. While Codes of Conduct have been controversial in open source, treating contributors well is essential to encourage contribution, especially among new or inexperienced contributors. This can be an important boost to intrinsic motivation.
Timely Engagement. Most of us are accustomed to prioritizing our “paying” to-do tasks over non-paying tasks. But if you are an open source project manager, you need to prioritize contributor communications very high. If you don’t have time to do this, you need to hire someone to help you. Of course, you don’t need to fix every issue and merge every PR right away. But responding quickly shows contributors you value their input. You might have forgotten how much it can mean to a novice contributor to get tagged with a task.
Easy Engagement. Contributors are most attracted to projects that make potential contributions digestible. Project managers should clearly identify what needs help, and not require contributors to wade into the entire codebase to lend their help.
International Outlook. Enabling contributors to do translations and localization shows respect for the project’s broader community. It also gives non-technical participants or novice coders a good way to get involved.
Contributor Thank Yous. This is nothing new, but a “shout out” to contributors can mean a lot to them–whether as a matter of personal gratification or resume building.

Do What Works, But Don’t Assume it’s All About Money

All this doesn’t mean monetary or symbolic compensation is not desirable or important. Contributors need to pay their bills, and street cred on GitHub is important. But for startup projects with no money to fund payment for contributions, non-monetary and indirect rewards can be the difference between a healthy project and a stagnating one.

Like this:

Leave a ReplyCancel reply