
The Fundamental Choice
You name is Miss Ann E. Price. It’s the last Friday of the month. You’ve finished work for the day. And you’ve just been paid. So it’s time to grab your coat and…do what exactly?
Well, in the old days, you would have gone on a splurge before closing time. You would have shopped till you dropped, been drastic with the plastic. But now things are different. Times are hard. The credit crunch has taken its toll. So you have got to start cutting back. Indeed, you’ve just learned today that your pension plan is seriously under-funded. So you’ve got to start putting more money aside.
Accordingly, you’ve made the sensible decision: you are going directly to the bank after work to deposit your wages into a savings account.
But on the way fate intervenes. You happen to pass by your favourite shoe store. In the window, you espy—OMG!—the sexiest pair of high-heeled shoes you have ever seen. They are just begging you to take them home. There is just one small problem: the £400 price tag! That makes you hesitate. Then another passer-by stops to gaze at them too. Danger signal: she going for her wallet! Well, it’s now or never….
So what should you do? Should you save for a rainy day, or shell out for the shoes?
Time Bias and Future Discounting
The fundamental choice that the story illustrates—whether to save or to spend—is one that constantly confronts people. Should I get something I desire now (e.g., a pair of sexy shoes) or should I wait for something even more desirable later (e.g., solvency in my golden years)?
Ultimately, it’s up to each of us to decide. But how we decide reflects our level of time bias—that is, how much we value our present satisfaction over our future satisfaction.
In one sense, we all have a time bias for the present. Specifically, we always prefer something now over something later if it’s the same something. So, if I offered you the choice between receiving £100 now, or £100 a year from now, you would surely take the £100 now. After all, a lot can happen in a year. I might not be around to give you the £100, and you might not be around
to enjoy it!
Nonetheless, people differ in their degree of time bias. That is, some people discount the future more compared to the present (i.e., they prioritize getting the sexy shoes right now), whereas others discount the future less compared to the present (i.e., they prioritize drawing on the retirement fund later on.) Why should we care about this?
Discounting the Future More Can Be Bad For You
One reason is that scientific research finds that discounting the future more can be bad for you. In one classic study, young children were given a choice: eat one yummy marshmallow right away, or eat two yummy marshmallows after a delay. The choice they made was recorded by the researchers. Many years later, the children were followed up. How were they getting on? Amazingly, those children who chose one marshmallow right away were more likely to have done worse in school and to have had adjustment problems. More recent research finds much the same in adults. People who report having problems resisting temptation—a clear sign of discounting the future more in favour of the present—also have poorer quality relationships and worse mental health. So the lesson is clear: prudence is better than indulgence. Knowing someone’s degree of time bias can enable you to predict how their lives turn out.
Measuring Time Bias
Now, suppose I wanted to measure your degree of time bias: how would I go about it? Well, for convenience, I might present you with a choice between receiving a smaller amount of money now or a larger amount of money later. (This is a realistic scenario. When money is saved a bank, it typically earns interest, so that more money will be available subsequently if the principal is not spent immediately.)
For example, I might ask you which you would prefer: to receive £45 in three days, or £70 in three months? Finding it hard to choose? Don’t be surprised: people split about 50:50 on this one. But if you are the sort of person who discounts the future more, you’ll likely choose the $45 in three days, whereas if you are the sort of person who discounts it less, you’ll likely choose the £70 in three months.
A more complete way to measure your degree of time bias would be to ask a whole series of questions trading off a smaller amount of money now for a larger amount of money in the future. Your personal “tipping point”–that is, where you switched from one option to the other—could then be determined. Indeed, there exists a questionnaire designed for just this purpose called the Monetary Choice Questionnaire.
Predicting Time bias
Okay, so let’s suppose we can accurately measure your time bias. How, then, might we go about predicting it?
One way would simply ask you whether you regard spending as preferable to saving, or saving as preferable to spending. If you picked the first option, you would probably be inclined to discount the future more. If you picked the second, you would probably be inclined to discount the future less. After all, it’s likely you would possess some insight into your own preference and would be able to report it accurately. To provide the most precise information, a researcher might ask to indicate on a rating scale your degree of explicit preference for spending or saving.
However, your self-report might be misleading for a number of reasons. Maybe you would resent the inquiry, and misreport your preference (i.e., tell a lie). Or maybe you would tell a researcher what you thought he wanted to hear (i.e., succumb to social desirability pressure). Or maybe you would tell yourself what you thought you wanted to hear (i.e., fall prey to self-deception). Or maybe you wouldn’t really have a good sense of your preferences (i.e., suffer from a lack of self-knowledge). For all these reasons, your self-report of your preference for saving versus spending might be misleading. If so, it would fail to predict your time bias.
So, if your self-report might not always be reliable, is there any other way of predicting your time bias? Thanks to the wonders of scientific psychology and modern computing there is. It goes by the name of the Implicit Association Test, or IAT.
The Implicit Association Test
What is the IAT? Basically, it’s a rapid-fire classification task run on computer. Your job is to classify items into categories as quickly as you can without making errors. To do this, you press one key if the items presented fall into one pair of categories, and another key if they fall into another pair of categories. Halfway through the task, the assignment of keys to categories gets switched around. Depending on which half of the task you find easier to do, it is possible to infer how much you automatically associate one pair of categories with another. For example, it can be inferred from how strongly you link the categories Spending and Saving with the categories Good and Bad respectively.
The nitty-gritty details of how the IAT works, and how its data get processed, are complex and need not detain us here. For current purposes it is enough to know that the IAT can gauge your degree of implicit preference for spending versus saving.
Now, implicit preferences have two potential advantages. First, they resist efforts to disguise them. Hence, they will be less vulnerable to problems of outright deception and social desirability. Second, implicit preferences originate at a more primitive level of mental operation—one more unconscious than conscious, more emotional than rational, and more impulsive than reflective. Hence, they can reflect information that would be concealed by motivated self-deception or insufficient self-knowledge.
The Proof of the Pudding
Do implicit preferences really have these advantages? How can we know the claims above are true? The answer is that they have been proven in scientific research. Dozens of studies have tested whether or not the IAT can predict important outcomes. The verdict is that it can—in cases ranging from what people eat, to how people vote. Importantly, the IAT often predicts important outcomes independently of self-report. This means that it can tap into unique aspects of people’s mental and behavioural tendencies. For example, when the IAT and self-report are used to measure people’s social prejudices—ones they might not always wish to reveal or admit—the IAT predicts related behavioural discrimination better than self-report does. This means that, in cases where self-reports might be suspect, the IAT comes into its own.
Another Proof of Concept
One of Strata’s unique selling points is its capacity to use the IAT as an investigative tool. Strata’s resident IT guru, James Klymowsky, has built a powerful, efficient, and user-friendly online system that can deploy customized IATs to large samples. This state-of-the-art system also optimally processes the data collected with confidential algorithms, and can provide individual summary feedback to IAT respondents on request. This system has enabled Strata to apply the IAT profitably to questions raised in healthcare market research. Indeed, one of our recent research projects where the IAT played a key role is in the running for a prestigious EphMRA award.
But we at Strata are always keen to double-check the validity of our tools. We wanted to verify for ourselves that the IAT could reveal something unique about people. Accordingly, we arranged for our consultant psychologist from the University of Southampton, Dr. Aiden P. Gregg, and a psychology undergraduate from the University of Essex, Samir Soormally, to run a little proof-of-concept study for us. This study featured over 250 participants, recruited from around the globe. All participated via our online IAT system.
This study had to do—yes, you’ve guessed it—with predicting people’s degree of time bias. Again, this is a crucial determinant of the financial choices people’s make—whether about shoes, pension plans, or anything else. Specifically, we tested whether and how well people’s degree of time bias would be predicted by two indices: (a) a standard self-report measure of people’s preference for spending versus saving (i.e., their explicit preference); and (b) our special IAT measure of people’s preference for spending versus saving (i.e., their implicit preference).
The results could have come out in many different ways. Naturally, we expected people’s explicit preferences to predict their degree of time bias. But would people’s implicit preferences do so too? Furthermore, would implicit preferences do so as effectively as explicit preferences? And would implicit preferences predict time bias independently of explicit preferences? If this last result worked out, it would once again confirm that the IAT could reveal information above and beyond self-report. It would show that, here, the IAT could uniquely predict whether people are liable to be either indulgent (e.g., take one marshmallow; shell out for shoes) or prudent (e.g., wait for two marshmallows; put money in the bank). So what happened?
The Results
Before we tell you—we like to build suspense—let us note some preliminary findings. First—and reassuringly—participants showed an overall preference for saving. Moreover, they showed this preference both in self-reports and on the IAT. The fact that explicit and implicit preferences here concurred means that we can have extra confidence about the accuracy and significance of the result. This is a “two thumbs up” situation. We can conclude that our sample were definitely more prudent than indulgent overall.
Nonetheless, when we looked at participants individually, their explicit preference was only a weak guide to their implicit preference (and vice versa). But this is a good thing. This means that, despite pointing in the same direction overall, participants’ explicit and implicit preferences were nonetheless relatively independent—possibly because they were tapping into different aspects of participants’ minds. If so, then maybe they would show different patterns of prediction as regards time bias.
And they did. Explicit preferences for saving versus spending predicted time bias; but implicit preferences also did. Moreover, both types of preference predicted time bias to the same extent. Most interestingly, however, both types of preference predicted it largely independently. That is to say, what self-report predicted about time bias, the IAT did not predict; and what the IAT predicted about time bias, self-report did not predict.
Consequently, this result is a validation of the usefulness of both self-report and the IAT. Each can provide useful predictive information that the other does not. Hence, the IAT should be seen, not as replacing conventional self-report measures but as supplementing them. Our findings illustrate how, with both the self-report measures and the IAT together, one can maximize one’s ability to predict an outcome of interest.
Apply the IAT to Healthcare Market Research
Our successful proof-of-concept illustrates the potential utility of IAT in research of all sorts. This includes healthcare market research.
For example, suppose a healthcare market researcher wanted two things: (a) to determine whether physicians preferred a newly developed drug Avantgardium to an old traditional stand-by Nostalgium; and (b) to predict whether physicians would be more likely to prescribe Avantgardium over Nostalgium.
Extrapolating from our proof-of-concept study, the IAT could perform two valuable services here.
First, the IAT’s results could be compared to those from interviewers or questionnaires. Let’s suppose the latter indicated an explicit preference for Avantgardium. That would be a good
initial sign.
But still, can we be certain that physicians were being fully frank? Or could they have just been pleasing the researcher with their answers? Could they have been flattering themselves about their openness to innovation? Were they coming up with plausible answers on the spot rather than identifying their true inclinations?
Sometimes there is no need to worry. After all, self-reports are often accurate. But where there is money on the line, greater certainty and additional insight may be desired.
Suppose that IAT results also showed that Avantgardium was more associated with the concept Good (or Effective, Safe, Cheap etc.) and that Nostalgium was more associated with the concept Bad (or Ueless, Dangerous, Expensive etc.). Such an implicit preference for Avantgardium over Nostalgium, in conjunction with the explicit preference, would serve to certify physicians’ receptivity towards Avantgardium. Everything would suggest that it was genuine. But suppose, on the other hand, that the IAT results came out the other way around. Suppose they showed that Avantgardium was more associated with negative concepts and Nostalgium with positive concepts. This would be a red flag. Something somewhere would be amiss. Further investigation would be warranted.
The second service that the IAT could provide would be in helping to predict overall levels of Avantgardium versus Nostalgium prescription among physicians, or in helping to predict which physicians would be more likely to prescribe one or the other. As our proof-of-concept study illustrated, and as much prior research proves, the IAT can predict important outcomes above and beyond self-report measures. Such an extra degree of predictive accuracy is useful for anyone who cares about knowing more than the competition does, and about telling as much as possible what the bottom line is going to be.
Summary
People differ in the degree to which they prefer a smaller benefit now over a larger benefit later: their time bias. Self-reports of people’s explicit preference for spending to saving can predict time bias. But, above and beyond such self-reports, an IAT-based index of people’s implicit preference for spending and saving predicts time bias too. This proof-of-concept illustrates what valuable information the IAT can provide, and other scientific research proves it. By extension, the IAT has a useful role to play in healthcare market research too.
References
Frederick, S., & Loewenstein, G., & O'Donoghue, T. (2002). Time discounting and time preference: A critical review. Journal of Economic Literature, 40, 351-401.
Greenwald, A. G., Poehlman, T. A., Uhlmann, E., & Banaji, M. R. (2009). Understanding and using the Implicit Association Test: III. Meta-analysis of predictive validity. Journal of Personality and Social Psychology, 97, 17–41.
Shoda, Y., Mischel, W., Peake, P. K. (1990). Predicting adolescent cognitive and self-regulatory competencies from preschool delay of gratification: Identifying diagnostic conditions. Developmental Psychology, 26, 978–986.
Tangney, J.P., Baumeister, R.F., & Boone, A.L. (2004). High self-control predicts good adjustment, less pathology, better grades, and interpersonal success. Journal of Personality, 72, 271-324.
Online link to 50 studies showing real-world validity of the IAT:http://faculty.washington.edu/agg/pdf/Real-world_samples.pdf