How long till today’s cigarette will make me ill? Time estimation revisited

Today’s cigarette may yield considerable satisfaction to a smoker, although that satisfaction will likely pale in comparison to the disutility arising from that person’s lung cancer in the future. Likewise, today’s Big Mac and extra-large bag of potato chips may gratify the consumer but the long-term consequences—from obesity to diabetes to heart attacks—are unlikely to be welcome.

So why do people smoke? Why do they indulge in junk foods, heroin, excessive alcohol intake, and all those other behaviors that are known to be damaging to one’s future health and well-being. The answer is complex and nuanced, but one factor that has frequently been cited in explaining unhealthy behaviors is known as future discounting. Unhealthy behaviors typically have a delayed effect on health: no one drops dead after smoking their first cigarette or eating their first bag of potato chips. And even binge drinkers get over their hangovers quite quickly, especially at an early age.

And the delay in the consequences relative to the immediacy of the reward (some people find a Big Mac, cigarettes, or heroin rewarding) implies that one’s decision about whether or not to smoke or drink needs to weigh today’s benefits against tomorrow’s costs. (There are, of course, other factors that contribute to addictive behaviors, although we can ignore those here for simplicity.)

Whenever people have to weigh costs and benefits across time, they exhibit discounting. As we noted earlier, very few people would prefer to wait a month to receive $51 if the alternative were to receive $50 today, even though the accrual during this delay would correspond to a gratifying annual interest rate of nearly 27%.

This tendency to discount is so strong, that our strong preference for immediate rewards has been linked to the proverbial “reptilian brain,” which competes with our “rational brain” that is telling us to consider and plan for the future.

So what determines our tendency to discount the future? Is it that we just don’t care much about what happens a year down the road? Or might it be that we just don’t have a concept of how far (or near) a year from now really is?

After all, if July 2018 seems to be a decade away in psychological units of time perception, then perhaps we do not really discount the future quite as steeply as it may appear when expressed in calendar time. Conversely, if July 2018 is barely further away in psychological terms than next month, then we may discount even more steeply than it appears when expressed in calendar time.

A recent article in the Psychonomic Society’s journal Attention, Perception, and Psychophysics examined the relationship between physical (calendar) time and our subjective perception of that time.

Researchers Camila Agostino, Marcelo Caetano, Fuat Balc, Peter Claessens, and Yossi Zana asked participants to express their perceived duration of time intervals as the lengths of lines. In one condition, participants would respond using stimuli of the type shown in the figure below:

Across trials, the durations would vary from 3 to 36 months, and participants indicated their perception of the duration by clicking the corresponding position along the line. Participants were made aware of the range of times before the experiment commenced, so they had the ability mentally to divide the line between “very short” and “very long” into segments of the appropriate length.

There were 3 other conditions that used the same basic stimulus but with different labels for the end points—including one in which the extreme end was labeled “10 years”, thus being far beyond the maximum interval actually used and forcing participants to use only parts of the time line provided.

In a final condition, participants were given 4 seconds to “imagine how long the interval lasts” in the absence of any response line, before recording their response on a line without end markers. The figure below illustrates this condition:

The results were straightforward and are shown at the aggregate level (i.e., averaged across participants) in the figure below:

It can be seen that the subjective estimates are a linear function of actual calendar time, irrespective of how the judgments were elicited. (The Roman numerals in the legend correspond to the various conditions in the experiment: because the data did not differ much between conditions we do not have to be concerned with the details.) Formally speaking, 98% of the variance in responses was explained by a linear function of the actual calendar time.

These average data are, however, only part of the story because Agostino and colleagues found notable individual differences in their study. Specifically, slightly less than half the sample (14 out of 35 participants) exhibited the linear relationship between calendar time and subjective estimates. A further third of the sample (12) showed an acceleration in their estimates such that as calendar time increased, their estimates increased faster than the physical times. The remaining 9 participants exhibited a compression in their estimates, such that the estimates grew more slowly as calendar time continued to increase.

What do these data show us?

As a first approximation, at an aggregate level, they suggest that people’s perception of time is well calibrated to calendar time. That is, on average, 3 months are perceived to be 3 months and 30 months are 30 months and so on. It follows that the steep temporal discounting that is observed in numerous experiments is unlikely to be merely a consequence of people’s non-linear perception of future time.

That said, the presence of individual differences in the data of Agostino and colleagues meshes well with existing findings that reveal individual differences in people’s discounting behavior. Although discounting is affected by contextual variables, such that a person may exhibit different inter-temporal preferences in different circumstances, there is also considerable stability across multiple tests:  The test-retest reliability of discounting is substantial at intervals of up to 1 year and across different methods of testing discounting.  Moreover, the extent of discounting a person shows for monetary rewards is also related to other forms of discounting, such as the enjoyment of cigarettes for smokers and of heroin for addicted patients, and so on.

Those existing findings raise the intriguing possibility that the individual differences that are observed in discounting behavior are the same individual differences that arose in the study by Agostino and colleagues in the context of temporal estimation.

Perhaps smokers just have a different sense of time than people who do not smoke because they are worried about their future health.

Article focused on in this post:

Agostino, C. S., Caetano, M. S., Balci, F., Claessens, P. M., & Zana, Y. (2017). Individual differences in long-range time representation. Attention, Perception and Psychophysics79, 833-840. DOI: 10.3758/s13414-017-1286-9.

You may also like