West Point Cadets Applaud President Obama's Speech on Afghanistan

Yesterday, a friend of mine sent me a link to Wrzesniewski et al. (2014). I was excited to see what might be the largest longitudinal study on motivation to date, but as I read further, I began to notice some problems. In the 1970s, two brilliant psychology researchers noticed something pretty neat in their experiments: people did stuff for lots of reasons. So Ed Deci and Richard Ryan split these motivations into two basic categories “intrinsic” and “extrinsic.” 40 years on and thousands of peer-reviewed studies later covering the gamut from qualitative, to longitudinal, to randomized control trials, their “Self-Determination Theory (SDT)” is the best framework for understanding why people do stuff, and persist in doing it, that modern psychology has to offer. It’s a damn good theory with a lot of validated measurements and common vocabulary... that I wish the authors of Wrzesniewski et al. (2014) had used.

The problem I have with this study is that it makes a bold claim: that people with more than one reason to persist in an action are less likely to persist over time. And this claim runs counter to 40 years of motivation research and common sense. Because it violates the very foundational observation of self-determination theory, that people do things for lots of different simultaneous reasons, aka “motivation is multivariate.” It’s not that we get to choose to have more or less, we have millions of “whys'” whether we want to or not. Trying to have less is like choosing not to think about pink elephants right now. (No really. Try as hard as you can not to think about pink elephants. How’s that working out for ya?).

Making a bold claim is fine, but in the words of Carl Sagan, “extraordinary claims require extraordinary evidence.” Wrzesniewski et al. (2014) certainly has a large dataset, but the authors of the study did not choose to create that data with standard questionnaires. SDT has used a lot of measures of the years and a few have emerged as robust test of what researchers are generally looking for. They even made a website so we baby-researchers can easily find them (http://www.selfdeterminationtheory.org/questionnaires/). I can’t intelligently comment on the validity of the items in Wrzesniewski et al. (2014), but I looked at them and didn’t recognize any of them from the standard measures (I might be wrong though; most of my experience has been with a handful of health-focused questionnaires like the BREQ-2R). That doesn’t mean the authors’ analysis are wrong, per se, but it doesn’t give me a lot of confidence in their conclusions. It increases the risk of “junk in, junk out.”

Furthermore, Wrzesniewski et al. (2014) does not reference or measure the cadets’ perception of their Basic Psychological Needs, the mechanism, according to SDT, by which motivation is created. Which is fine if the study is making purely correlative conclusions, but Wrzesniewski et al. (2014) draw causal conclusions. “We show that holding both internal and instrumental motives for attending West Point harms outcomes” (Wrzesniewski et al., 2014, pg 1). The authors could have drawn this causal chain by measuring BPN, but they did not.

Finally, the authors of Wrzesniewski et al. (2014) have created their own “motivation types” for this study. That’s a much bigger deal, in my opinion, and more worrying. After 30 years of research, Deci and Ryan (2000) proposed SDT and to capture what they were seeing proposed a continuum with 6 different motivation types (amotivation, external, introjected, identified, integral, and intrinsic). Many studies since have grouped these 6 into a “Relative Autonomy Index (RAI) of “self-determined” (identified, integral, and intrinsic) and “non-self-determined” (amotivation, external, and introjected), though no studies to my knowledge have created new motivation types without foundational analysis that gave them credence. “We prefer the terms ‘internal’ and ‘instrumental'” doesn’t give me a lot of confidence in how or why the authors established those terms. And even the studies that have used the RAI may need to be revisited thanks to the profound and rigorously researched meta-analysis of Chemolli and Gagné that came out in March of 2014. Those researchers used two different analytical methods to show that RAI was not as good a fit to the data as previously indicated using two of the popular SDT measures. “Using the concept of a continuum and the practice of using the RAI dilutes this richness and also dilutes the richness of the results that are drawn from research” (Chemolli & Gagné, 2014, pg. 9). So this is two further strikes against making up your own motivational types AND questionnaires.

Here’s what we know pretty well from the bulk of previous SDT studies that Wrzesniewski et al. (2014) seems to validate: 1) motivation is multivariate. We do stuff for lots of reasons. 2) Some of those “whys” result in greater persistence. 3) some of those “whys” resulted in less persistence. 4) focusing on the “whys” that promote Basic Psychological Needs increase persistence. 5) the best “whys” reward our feelings of autonomy, competence, and relatedness to others.

The fact is, we will always have lots of reason to do stuff, but what matters most to persistence is identifying with those reasons that “feel” the best in the moment to get shit done. Over time, it is best to focus on the reasons that promote our sense of autonomy, competence, and relatedness to others, and it takes being aware of as many reasons as possible to know which ones are going to work best for you.