What really drives employees’ happiness?

For those who who prefer pictures to words, check out the infographic below this post.

In a culture dominated by money, we often forget that a job can provide many different kinds of benefits to people, from friendship to influence to intellectual stimulation.  Money is only one of many motivators, and some would argue it’s one that does a poorer job of buying happiness than we realize.  (For example, see Aknin, Norton & Dunn, 2009.)

Recently we ran a simple study to see how different motivators stack up: Embedded among other questions, we asked 584 employed Americans from a variety of jobs and industries how satisfied they were in their jobs and how happy they were in their lives.  We also asked these same folks the degree to which 7 different things were important to them, and how much they received each of these from their jobs.  In my last post I showed how important employees believed each of these to be.

The contenders, organized by how important employees said they were:

  1. Security (3.76/5)
  2. Work-Life Balance (3.72/5)
  3. Intellectual Stimulation (3.54/5)
  4. Money (3.22/5)
  5. Achievement/Prestige (3.21/5)
  6. Affiliation/Friendship (2.94/5)
  7. Power/Influence (2.58/5)

Today we’ll answer a related but different question: How important do these actually turn out to be for Job Satisfaction and Life Happiness?  For this, we can look at the statistical relationship between these key outcomes and how much of each motivator people receive from their jobs.  In the chart below, these relationships are plotted, based on “squared correlations,” which essentially tell us the percent of one variable accounted for by the other.  In other words, what percent of an employee’s happiness and satisfaction are accounted for by money, intellectual stimulation, etc.?

On the Job Motivator How much of a person’s Job Satisfaction does it account for? How much of a person’s Overall Happiness does it account for?
Intellectual Stimulation 18.5%*** (WINNER) 8.6%*** (WINNER)
Achievement/Prestige 18.4%*** (2nd) 4.3%*** (3rd)
Power/Influence 14.9%*** (3rd) 6.9%*** (2nd)
Security 8.8%*** (4th) 4.0%*** (4th)
Work-Life Balance 8.0%*** (5th) 3.8%*** (5th)
Affiliation/Friendship 5.9%*** (6th) 3.8%*** (6th)
Money 5.4%*** (LAST) 2.6%*** (LAST)

These numbers look quite different from what employees said was important!

A few things to notice:

  • Some of these things mattered more than others, but all of them mattered.
  • The “winners” aren’t necessarily what you might have guessed.  Things that make jobs more stimulating, such as task variety, tended to be the most powerful drivers of one’s job satisfaction, and even happiness away from work.
  • Security and Work-Life Balance—the things employees reported were most important to them—are right in the center (4th and 5th) in terms of their actual prediction of happiness and satisfaction.  These things matter, but maybe not quite as much as people think.
  • As much as we may think getting that raise will make us happy, money was the single weakest predictor of both job satisfaction and overall happiness.
  • The amount of power one holds at work is something employees don’t think matters very much. (It was rated, far and away, as the least important motivator.)  But your employees may not know themselves as well as they think!  In actuality the amount of power and influence one holds at work was second only to intellectual stimulation in terms of predicting one’s overall happiness in life.

If you’re anything like me, when you want to know what matters to someone, you probably just ask them.  From that perspective, these results seem a little bit disheartening.  Do employees even really know what makes them happy?  And if they don’t, does it even help to ask them?

But before concluding that it doesn’t, we should look at the numbers a bit more closely.  Averages aside, if a person reports that a factor is more important to her, will she also respond more strongly to it?

To help visualize this, we divided the sample into two groups (high and low) for each of the “importance” questions.  If employees are able to accurately assess how much a particular motivator matters to them, then for those that say it matters more, the effect of it on their Job Satisfaction should be higher.  Is this the case?

On the Job Motivator Influence on Job Satisfaction among those who rated the motivator as LESS important. Influence on Job Satisfaction among those who rated the motivator as MORE important.
Intellectual Stimulation 7.6% 23.0%
Power/Influence 8.7% 19.7%
Achievement/Prestige 12.7% 27.8%
Security 3.3% 12.0%
Work-Life Balance 7.8% 9.0%
Friendship/Affiliation 0.4% 8.4%
Money 2.8% 11.6%

A cursory glance at the table above shows us that employees do know what drives their happiness, at least with regards to 6 of our 7 motivational elements.  Those employees who reported that a particular motivator was more important to them also showed significantly higher relationships between the relevant factor and Job Satisfaction.  The only exception was Work-Life Balance, which influenced Job Satisfaction to about the same degree regardless of whether one reported that it was important.  Though many people (likely workaholics like myself) reported that having time for other things (vacations, hobbies, etc.) was not particularly important to them, actually having a bit of time outside of work was just as beneficial for them.

However, a closer look at the 2nd table also suggests that 3 of the other 6 motivators are so beneficial that they’re strongly driving happiness even for those that don’t care about them.  These are intellectual stimulation, power and prestige. (Does this make anyone else think of J. Richard Hackman’s work?)  So basically if you want to have really happy employees try the following formula:

  • Offer employees extra money, security and social opportunities…but only to the extent they say these things matter to them. (But of course, don’t forget that people always need a bit of these!)
  • Try to ensure that jobs provide intellectual stimulation and task variety.  Give your employees some autonomy and influence in their work.  And provide folks with opportunities to acquire prestige and recognition.  Give people these things if they don’t say they need them.  Give them even more if they say they do.
  • Give all your employees a break now and again, even the consummate workaholics who will tell you they don’t want or need it.

I for one am putting these ideas into practice.  In my own job I’ve got plenty of intellectual stimulation and autonomy, and also a reasonable dose of prestige.  But now might be a good time to ask my boss for that vacation I’ve been meaning to take for the last year!


Cangrade_Employee_Hapiness Infographics

Info-graphic by Vadik Bakman


  1. Avatar Ryan Connors   •  

    Great insights Steve. It’s interesting to see the underlying motivating factors in workplace happiness.

  2. Avatar Steve Phelan   •  

    How much multicollinearity was there in the data and how did you control for it?

    • Steve Lehr Steve Lehr   •     Author

      Hi Steve- Good question (and one that almost sounds like it comes from a professor). For the sake of simplicity, this blog reports single variate results, which of course are not truly independent. In general the 7 motivators described are modestly related, with correlations ranging from r=.017 (between amount of money and amount of affiliation/friendship) to r=.455 (between intellectual stimulation and prestige), with most in the r=.2 range. Of course, before we address multicollinearity (and for extra rigor and clarity), we should look at these numbers in a multivariate fashion, starting with Job Satisfaction as DV. First, note that the general pattern of results is similar if the motivators are entered into multivariate models. In general, each predictor remains positive and retains statistical significance, with the exception money, which sneaks out of the equation and becomes insignificant in most models. In most models, intellectual stimulation retains the top spot. One noteworthy change is that Work-Life Balance, which is only very weakly correlated with other factors, tends to become slightly more important in multivariate models. To be precise, the standardized coefficients in a simplistic model that includes all 7 motivators (but nothing else) as predictors of Job Satisfaction are as follows (R-square = .337):

      Power/Influence: S-Beta = .132, p=.001
      Intellectual Stimulation: S-Beta = .210, p<.001
      Money: S-Beta = .030, p=.427
      Security: S-Beta = .116, p= .002
      Achievement/Prestige: S-Beta = .194, p<.001
      Work-Life Balance: S-Beta = .177, p<.001
      Affiliation/Friendship: S-Beta = .088, p=.014

      For overall life happiness, an additional and interesting difference is that achievement/prestige also loses significance. The part of achievement/prestige that predicts life happiness (as opposed to happiness specifically in one's job) may be somewhat conflated with intellectual stimulation and power/prestige (its two highest correlates).

      All that said, to answer your question, in these models, multicollinearity does not appear to be a serious problem in these models, with VIFs ranging from 1.06 (Work-Life Balance) to 1.50 (Achievement/Prestige) and averaging 1.29.

  3. Avatar Dragomir   •  

    Shouldn’t the current level of a respondent’s access to any of the above mentioned motivators be uniform among the respondents? What I mean is, if we are asking people how important work-life balance is to them, for example, shouldn’t we ask only people who work in places with the same work-balance policies?

    I am speculating this way: if i were already working at place with perfect work-life balance, then, just because I already have that I might perceive it as less valuable. But if I were bereft of it, I might then value it much higher.

    How does your study take into account such possible discrepancies?

    • Steve Lehr Steve Lehr   •     Author

      Hi Dragomir- Very good question. In general I think that you are right that one source of these results may be that most people tend to enjoy “adequate levels” of some motivators, but have more variation in others. For example, in the United States even those who make very little money still make enough to live. Would these results be the same if we included people well below the starvation line (e.g. say in the Democratic Republic of Congo where adjusted per capita GDP sits at just over $200 a year)? These data can’t speak to this question, but my guess is that were we to take that into account, things would change. The idea is reminiscent of Maslow’s hierarchy of needs: Presumably people will care less about Achievement or Intellectual Stimulation if they do not have even rice to eat.

      That said, while acknowledging that we do have some degree of range restriction (as well as other reliability limitations) in these data, we can introduce some statistical controls to ensure results are driven by differences in levels of the individual motivators, as opposed to by relationships between these variables or other factors such perceived importance levels. For example, look at the multivariate model I just posted in response to Steve Phelan’s question. Essentially multiple regression analysis uses math to artificially hold other things constant when examining a variable’s influence. So for example, in the model posted in my comment the effect of “Power/Influence” on Job Satisfaction represents an estimate of how large this effect would be if we were able to artificially keep the other 6 motivators at the same level for each person. In general (coming back to your question), adding other factors (such as the importance ratings) into this model increases complexity, but does not change the overall pattern of results.

  4. Avatar Dragomir   •  

    I cannot but ask one more question. How do you even out different personality inclinations towards one or another type of work? For example, there are people who seem to enjoy repetitive, mind-numbing work or at least feel nothing against such. The same people might feel uncomfortable and even threatened about keeping their jobs if given intellectually stimulating work. Same holds true for achievement/ prestige pursuits – different personalities have different levels of need for achievement/ prestige.

    • Avatar J. Gentile   •  

      The categories have a familiar ring very closely resembling the David McClelland’s motivational model – nPow, nAch, nAff. It would be interesting to see how the priorities of each of the individuals aligned with their overall motivational profile. Those who favored influence potentially being higher in nPow than those favoring intellectual stimulation (who may favor nAch). Just a thought.

      • Steve Lehr Steve Lehr   •     Author

        Hi Gentile- Yes you are definitely correct. In fact I almost gave David McClelland a shout out in the blog, as I did Richard Hackman. Both occupied offices on the top floor of William James Hall, which (as it happens) is also where I sit and write this comment! Beyond any odd sense of closeness I may feel accordingly, both were masterful researchers whose influence can be found all over modern theories of motivation. While I have not administered the Thematic Apperception Test (and so cannot correlate it with these survey responses), my suspicion is that you are correct: Some of the questions we ask are likely to be strongly related to nPow, nAch and nAff as described by McClelland.

    • Steve Lehr Steve Lehr   •     Author

      A reasonable question, Dragomir. In these data we capture people’s relative inclination by simply asking them how important each thing is to them. The interactions described in the blog (and visualized in the infographic) suggest that these are (with the exception of work-life balance) valid (if not perfect) measures. However, to truly understand a complex individual person, we would want to look at things of this sort in far greater detail. Indeed, this is one of the things we do in our job-fit inventory, which includes a couple of dozen different personality/preference/motivation scales, including (for example) “preference for routine.” Naturally such scales tend to work better, as the multi-item format improves their reliability.

  5. Avatar Dragomir   •  

    On a second note, what about growth as a major motivator? It encompasses different personalities (for one person growth in salary might be most important, for somebody else growth in hierarchy, etc). It would be interesting to see research comparing preferences to growth in salary vs secuity or something like that.

    • Steve Lehr Steve Lehr   •     Author

      This is a great suggestion. Thanks, Dragomir! In our personality inventory, we capture a variant of mastery goal orientation, which is related to this. However, I think this is a great motivator to add to our next study.

  6. Avatar agen nova88   •  

    It’s an awesome post in favor of all the internet viewers; they will obtain advantage from it I am sure.

Leave a Reply

Your email address will not be published. Required fields are marked *