What do people want: more money or more time? Who is happier: those who want money, or those who want time? Do these questions mean the same things to different people? Do they mean the same thing to the same person at different times? Do we know what we’re doing when we rate our own happiness?
A few weeks ago I commented on a study by Hal E. Hershfield, Cassie Mogilner, and Uri Barnea, “People Who Choose Time Over Money Are Happier” (Social Psychological and Personality Science, vol. 7, no. 7 , 697-706; see also the authors’ NYT article). I saw possible problems with it but did not have time to read it closely. My criticism was a bit caustic and uninformed; I ended up disliking and deleting the post. I regret the tone but not the critical impulse.
Now looking at the actual study again, I find it both stronger and weaker than I previously thought.
It is stronger in its versatility. The authors considered many possibilities; they were continually revising and refining their hypotheses and tests.
But that’s also a problem. The paper’s seven studies go in somewhat different directions; in my reading, they don’t point together to a conclusion.
Here they are:
Study 1a: 1,301 participants (1,226 in the final sample) were recruited through Mechanical Turk and asked about their preference for time or money. They were also asked to rate their happiness and life satisfaction. The order of these questions was balanced among the participants (I missed this point the first time around).
More people chose money than time, but those who chose time reported greater happiness than those who chose money. The difference does not seem great to me, regardless of statistical significance (M = 4.65, SD = 1.32 vs. M = 4.18, SD = 1.38), but I may be wrong here.
Study 1b: The authors do not describe this in detail, but they claim to have replicated the results of 1a while controlling for materialism. Participants (N = 1,021) were again recruited through Mechanical Turk.
Study 2: This time, 535 participants were recruited in the train station of a major East Coast city and offered a Granola bar to complete the survey. 429 actually did complete it. They reported substantially higher income than the participants in 1a and 1b; also, a majority (55%) chose time over money, unlike the MTurk participants, who tended to choose money over time. (Did the train station setting affect this in any way, I wonder?) Those who chose time were again happier, by their own rating, than those who chose money (M = 5.28, SD = 0.93 vs. M = 4.91, SD = 1.10).
Study 3a: This time, the researchers sought to find out why people preferred what they did. So they recruited participants through MTurk, asked them which they preferred (time or money), asked them to explain why, and then asked them to rate their happiness. This time, the order of the questions was fixed. They saw a split between using the resource to cover needs and using it to cover wants, as well as a split between using the resource for others and using it for oneself. Something curious appears here: participants indicated whether they wanted more time in their days or in their lives. While the desire for more time (generally) correlated with happiness, the desire for more time in one’s day did not, nor did the desire for more time in one’s life. I wonder what this means.
Study 3b: This time, 1,000 participants were recruited through Qualtrics for a nationally representative sample. 943 ended up participating. As in most of the previous studies, the majority indicated a preference for more money over more time, but those who chose time rated themselves as happier. In addition, the ones who indicated that they would spend the resource on wants were happier , by their own rating, than those who said they would spend it on needs; those who said they would spend it on others were happier than those who said they would spend it on themselves. There were some additional findings. (One interesting detail: The Qualtrics participants were on average 15-2o years older than the MTurk and train station participants; also, a much lower percentage were employed.)
Study 4a: This was the first of two manipulation checks. Participants were recruited through MTurk and assigned randomly to one of three conditions: a “wanting time” condition, in which they were instructed to write about why they wanted more time, a “wanting money” condition (likewise with a writing task), and a control condition, for which they had to write down 10 facts. Then they were asked to rate their happiness. Finally, they were to indicate which they would rather have, more time or more money.
Those in the “want time” condition (randomly assigned) tended to indicate a preference for more time; those in the “want money” condition, for more money. The difference in happiness was marginal across the groups, but those in the “want time” condition were slightly happier by their own rating than those in the “want money” condition.
Study 4b: This was the last of the studies and the second manipulation check. This time, participants (again recruited through MTurk) were assigned randomly to a happy condition (instructed to write about why they were happy), an unhappy condition (instructed to write about why they were unhappy), and a control condition (without a writing task). They were then asked to rate their happiness. Finally, they were asked questions about their resource preference. Those in the happy condition reported greater happiness (and a greater preference for time) than those in the unhappy condition.
There are some details I have left out for brevity’s sake: for instance, the researchers included some questions about subjective and objective income and controlled for these. But this is the gist.
Now for some thoughts:
First of all, these seem like pre-study experiments rather than complete studies, in that they deal with different populations, questions, and methodologies. It is good that the researchers were refining their questions and analyses along the way, but in the process they may have come up with explanations that they did not rigorously test. For instance, the relation between an emphasis on wants (rather than needs) and happiness seems hypothetical, even if it makes intuitive sense. There’s a flipside: people can drive themselves into a tizzy by thinking about things they want but don’t have.
Second—and this concerns me more—studies 4a and 4b suggest that participants’ preferences and happiness ratings can be manipulated by something as simple as a writing task. It’s possible that most people want more money and more time; what they think they want at a given moment may have a lot to do with what’s going on around them.
Also, I suspect that the MTurk participants, especially those completing surveys for the money, might be a financially stressed bunch. That could influence the findings considerably.
In addition, money and time are not easily separable. That is my greatest qualm. I wonder how many participants thought: “Well, I’d like to have both, but I think the money would allow me to buy more time, so I’ll choose money.”
Who, then, would choose time? Maybe people who have something important in their lives. People may desire money for all sorts of things—leisure, power, luxury, relief from debt, etc.—but those who wish for more time probably have something in the works that they enjoy or value. That in itself could explain why they rate their happiness a little higher than the others do.
But then, how accurate is my assessment of my happiness? How accurate is it ever? It can fluctuate throughout the day; moreover, it can grow (or shrink) in retrospect. Forsan et haec olim meminisse iuvabit (Virgil, Aeneid); in the translation of Robert Fagles, “A joy it will be one day, perhaps, to remember even this.”