دانلود مقاله ISI انگلیسی شماره 4911
ترجمه فارسی عنوان مقاله

شگفت انگیزی کم قدرت انگیزشی پاداش های آینده : مقایسه اندازه گیری مرسوم مبتنی بر پول تنزیل با اندازه گیری مبتنی بر انگیزش

عنوان انگلیسی
The surprisingly low motivational power of future rewards: Comparing conventional money-based measures of discounting with motivation-based measures
کد مقاله سال انتشار تعداد صفحات مقاله انگلیسی
4911 2010 22 صفحه PDF
منبع

Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)

Journal : Organizational Behavior and Human Decision Processes, Volume 111, Issue 2, March 2010, Pages 71–92

ترجمه کلمات کلیدی
تنزیل زمانی - آینده - ارزش فعلی - انگیزش - انگیزه مالی - نرخ تنزیل - اندازه گیری - استراتژی گذاری - تاثیر ارزیابی - تاثیراکتشافی
کلمات کلیدی انگلیسی
پیش نمایش مقاله
پیش نمایش مقاله  شگفت انگیزی کم قدرت انگیزشی پاداش های آینده : مقایسه اندازه گیری مرسوم مبتنی بر پول تنزیل با اندازه گیری مبتنی بر انگیزش

چکیده انگلیسی

Temporal discount rates are often poor predictors of behaviors that we expect will be motivated by the future. The current research suggests this may be because conventional discounting measures are poor measures of the motivational value of future rewards. In six studies, I develop motivation-based measures of the present value (PV) of future rewards and compare the PVs obtained with those obtained using conventional money-based discounting measures. Conventional money-based PVs consistently overestimate motivation-based PVs and are discriminable from them. I explore explanations for this mismatch, including timing of effort exertion (Study 2) and loss aversion (Study 3), both features of the motivation-based measures. In Study 5, I use self-reports of valuation strategies and a time pressure manipulation to demonstrate that participants use different valuation strategies in the conventional money-based and the motivation-based measures that, in part, determine the difference in PVs obtained and the relatively low correspondence between them.

مقدمه انگلیسی

A wide range of people’s behaviors in the present are motivated by future concerns: employees work for future bonuses, students study to obtain a degree, people exercise and follow healthy diets for future health, and they save for retirement. People are often myopic in these behaviors: procrastinating in their work, failing to exercise, eating unhealthy foods, and saving insufficiently. The value that people place on the future relative to the present is the focus of research on temporal discounting, a topic that is studied extensively in psychology and economics and in some applied fields, such as health behavior and financial investment. Temporal discount rates are broadly applied, often used in the construction of public policies that concern the future in domains such as finance, health, the environment, and urban planning, where they can have a large impact on whether or not a policy or project is implemented. While there is debate in these domains on the appropriate rate to use, individuals’ discount rates are considered one important source (see Baron, 2000 and Gyrd-Hansen and Søgaard, 1998, for a related discussion; Schelling, 2000). Much research in temporal discounting involves measuring these discounting rates for individuals, to examine how rates vary for different individuals (e.g., Green et al., 1994 and Kirby et al., 1999), in different decision domains (e.g., Chapman, 1996), for different magnitudes of rewards (e.g., Kirby, 1997), for different delays into the future (e.g., Thaler, 1981), and to examine the relationship between individuals’ discount rates and relevant real-life behaviors (e.g., Bickel et al., 1999 and Chapman and Coups, 1999). The motivation for much of this research, either explicitly stated or implied, is to predict individuals’ present behavior based on individuals’ discount rates for behaviors that researchers expect will be influenced by future concerns. These behaviors, such as saving for retirement, exercising, or eating healthily are typically multi-determined, motivated by future concerns (“I would love to be in shape for the beach this summer and I want to be healthy when I am older”), present concerns (“I am too tired to go to the gym today”), and factors intrinsic to the present behavior (“but once I get running on that treadmill I usually enjoy it”). Researchers are interested in all of these influences on present behavior. However, I suggest that research on temporal discount rates has been focused on understanding the first of these, the motivational power of future concerns for present behavior, but that it commonly uses measures that do a poor job of assessing motivational power. Good measures of constructs should satisfy a number of criteria. In particular they should be valid, reliable, and sensitive (Cook and Campbell, 1979 and Cronbach and Meehl, 1955). I suggest that the methods conventionally used to measure individuals’ discount rates mean that the discount rates obtained are likely to be inaccurate, and therefore invalid, measures of the motivational power of the future in the present. Specifically, I suspect that conventional discount rate measures are likely: (1) to overestimate the motivational power of the future for individuals and (2) to be discriminable from measures of the motivational power of the future, showing relatively low correlations with those measures. These are the hypotheses examined in the present research.

نتیجه گیری انگلیسی

The goals of this research were: (1) to examine whether conventional money-based measures of discounting or PV are accurate measures of the motivational power of the future for present behavior and (2) to explore potential explanations for any observed inconsistency between conventional money- and motivation-based measures of PV. This involved developing several motivation-based measures of PV for comparison with money-based PV measures. Using the financial incentive literature, I identified tasks on which individuals’ exerted effort, a marker of motivated behavior (Martin & Tesser, 2009), would vary closely with variation in monetary rewards. For these tasks, I calibrated the value of one or more future monetary rewards against the value of a range of present monetary rewards to estimate a motivation-based PV for the future monetary rewards. Value was measured in precisely the same way for present and future rewards. These motivation-based PVs were compared with money-based PVs for the same future rewards. The results were consistent across six studies. Money-based PVs were larger than the motivation-based PVs for the same monetary rewards for the same participants. This was found for three different measures of motivation-based PV: actual behavior on a detection task (Studies 1a and 1b), estimated behavior on the same detection task (Studies 2–4), and estimated behavior on an endurance task (Study 5). It was found for money-based PV measures based either on participants’ stated equivalent-value (Studies 1a, 1b and 2) or on their stated willingness to invest (Study 3). It was found with varied response methods for valuation, including providing a numerical value (Studies 1a, 1b and 2) and clicking on a visual-analogue scale (Studies 3–5) both with and without markings (Study 5). It was found when rewards were manipulated within-subjects (Studies 2–5), where participants presumably had more opportunity to “correct” this inconsistency in value. And it was found when the presentation and response methods used were identical for both money- and motivation-based measures of PV (Studies 3–5), that is: when the money-based PV measure used separate assessment of present and future rewards, mimicking the structure of the motivation-based measures of PV, and when the response mode for valuation was identical for both measures. Money- and motivation-based measures of PV for the same rewards were, in general, discriminant from each other. In Studies 1a, 1b and 2, where there were differences in the methods used for reward presentation and for valuation response across the two measures, the correlations were very low: all .22 or less, often close to zero. In studies, 3, 4 and 5, where the methods used for reward presentation and valuation response were matched across the two measures, this inter-correlation was higher, but the two measures of PV still showed some discriminant validity. Different money-based measures showed higher correlations with each other than with a motivation-based measure (Study 3) and motivation-based measures showed higher correlations with each other than with a money-based measure (Study 4) providing support for the convergent and discriminant validity of both money- and motivation-based measures (Campbell & Fiske, 1959). Beyond demonstrating the persistence of this difference in PV between money- and motivation-based measures across contexts and after removing differences in presentation and response methods between the measures, I also tested several potential explanations for the difference between PVs. In Study 2, I examined whether the proximity of effort exertion and valuation for the motivation-based measure might account for the difference between money- and motivation-based PVs. Separating in time participants’ effort exertion and their valuations had an effect on participants’ valuations on the motivation-based measure, but it changed valuations of present and future rewards to a similar extent, resulting in no overall change in PV. Therefore the difference in PV between the money- and motivation-based measures occurred and was of similar magnitude for both effort timing conditions suggesting that the exertion of effort proximal to valuation in the motivation-based measure is unlikely to account for this difference in PV. Similarly, using an invest version of the money-based measure had an effect on participants’ valuations (Study 3), but it changed valuations of present and future rewards to a similar extent, resulting in no overall change in PV. So the difference in PV between the money- and motivation-based measures occurred and was of similar magnitude for both the equivalent-value version and the invest version of the money-based measure suggesting that it is unlikely to be the loss aversion from exerting effort in the motivation-based measure of PV for a future reward that accounts for the difference in PV between money- and motivation-based measures. These conclusions, that the difference in PV is unlikely to be caused by proximity of effort exertion or loss aversion in the motivation-based measures, both depend on null effects: the lack of an interaction effect on valuation between measure (money- and motivation-based) and time of reward (present and future). So these results do not conclusively prove that these two factors do not contribute to the difference in PV between money- and motivation-based measures. However, the manipulations of these two factors did not fail: each had a main effect on valuation. So, the failure to detect an interaction suggests that any interaction effect is likely to be small and so is unlikely to account for more than a small part of the difference between PVs. A third explanation tested in Study 5 – that the valuation strategy that participants use when valuing future rewards differs for the conventional money-based and motivation-based measures – does account for at least part of the difference between money- and motivation-based PV. In Study 5, most participants relied on an anchor-and-adjustment strategy for the conventional money-based measure and most participants relied on an affect-based strategy for the motivation-based measure. This latter strategy, where feelings associated with present behavior on a task are directly compared with feelings associated with future rewards, results in lower PVs for future rewards. The valuation strategy participants used also changed when under time pressure, resulting in greater relative use of the affect-based strategy for valuing future rewards on the conventional money-based measure and a corresponding drop in the PVs. This finding is consistent with previous research looking at valuation of present and future rewards (Ebert, 2001 and Ebert and Prelec, 2007), which suggested that people may use an anchor-and-adjustment strategy for the valuation of future rewards. In that research, application of time pressure increased PV for distant future rewards, probably by restricting adjustment for delay or increasing neglect of the future time dimension. Here, the effect of time pressure was to change the valuation strategy that participants used on the conventional money-based measure, so they were more likely to rely on an affect-based strategy, resulting in a lower PV. This was perhaps the natural response in this situation, where many participants had already used an affect-based strategy at least once when completing the measures with no time pressure. The finding that money- and motivation-based measures dissociate – that is, they result in different PVs for the same rewards and show discriminant validity – fits with previous research that has found dissociation between monetary assessments of value and affect-based assessments of value (Amir et al., 2008 and Hsee and Rottenstreich, 2004). In the current studies, correlations between PVs obtained from money- and motivation-based measures increased with identical methods of presentation and response for both measures, but the measures still showed some discriminant validity and the difference in PV remained. These results suggest that money- and motivation-based measures may assess different separable dimensions or constructs of future rewards, their perceived monetary value and their motivational value, respectively. Consistent with this idea, correlations between the money- and motivation-based measures were considerably higher for those individuals who used the same valuation strategy for both the conventional money-based or motivation-based measures compared to those individuals who used different strategies on the measures (Study 5). That is, those individuals who use the same strategy on both measures are perhaps assessing the same dimension or construct of the future rewards on both measures, i.e., the perceived monetary value or motivational value on both measures. Certainly the present research establishes that conventional money-based measures provide poor assessments of the motivational value of the future for the present. Crucially, this was the case even though the valuation of present rewards on each type of measure was used to calibrate the money- and motivation-based measures to enable comparison across the measures of PVs for future rewards. And the difference seen consistently across all six studies for the conventional money-based and the motivation-based measures was for PV, i.e., for the relative value of future to present rewards. There was not simply a difference across measures on absolute values of rewards. Future rewards are surprisingly low in motivational value for people, such that people exert effort, and estimate they will exert effort, considerably more for present rewards that they claim have exactly the same present value to them as the future rewards have. This result is perhaps especially interesting given that discount rates for future money that are obtained using conventional money-based measures tend to be quite high (and PVs low) relative to market interest rates, a benchmark sometimes used by intertemporal researchers (Frederick, Loewenstein, & O’Donoghue, 2002). Because of the care taken in the development and comparison of these different measures of PV, there are a number of explanations used to explain other inconsistencies in the literature on discounting that cannot account for the inconsistency seen here between the conventional money-based and the motivation-based PVs. For example, past research has shown that people use different discount rates for different domains, for different products, for different magnitudes of reward, and for gains versus losses. However, the current results do not simply reflect the use of different discount rates for different quantities of any kind, e.g., different discounting for effort versus money, for different magnitudes, etc. Both the conventional money-based and the motivation-based methods used identical rewards (identical in domain, size, type, time available, and whether they were real or hypothetical) and both measures identified the equivalent present monetary value for a future monetary reward. That is, both the money- and motivation-based measures measured discounting of future money relative to present money. I have argued that discount rates ought to reflect the motivational value of individuals’ enduring future concerns relative to their enduring immediate concerns if they are to be useful for predicting people’s behavior on average or over the long-run, i.e., when a temporary change in the attraction of short-term considerations does not overwhelm people’s longer-term considerations. But this research suggests that discount rates measured using conventional money-based measures do not accurately represent the motivational value of future rewards in the present. Past research has often shown poor correspondence between conventionally-measured discount rates and real behaviors that we expect to be motivated by the future, but are motivation-based measures likely to be more useful in this regard? The motivation-based measures used here recruited a very limited set of present behaviors: actual or estimated effort exerted on simple tasks for which effort was closely tied to variation in financial incentive. This was deliberate: the primary goal of this research was to examine whether money-based measures of discounting or PV are accurate measures of the motivational power of the future for present behavior. Participants were motivated by the future rewards offered, and effort was used as a marker of motivation. Any other reasons for variation in present behavior were minimized or held constant across experiment conditions (by randomization of participants to condition). These constraints help to ensure internal validity in this research, allowing me to conclude that conventional money-based measures of discounting are inaccurate measures of the motivational value of future rewards. However, these constraints are also likely to restrict the predictive power of the motivation-based measures for other behaviors that involve motivation by the future. Real behaviors that involve future motivation are commonly complex, determined by future and present concerns, and any connections between them. For example, whether someone chooses to invest the money they receive from a tax return or to buy themselves a vacation is likely to depend on the importance of future concerns (“I want to be sure I have enough savings for when I want to buy a house”), present concerns (“but I could really do with a vacation right now”), and the perceived connection between the present action and the future reward (“if I invest this $700 in my savings account now, how much difference will that make for my savings in 5 years”?). Also, the present research only used monetary rewards. The considerations involved in other kinds of rewards, such as those in other domains, will be different. For example, in considering future health people may feel less certain of the effect of their actions now on the likelihood of achieving the reward in the future (“if I eat a piece of cake instead of an apple now, will that really make any difference to my health 10 years from now”?). As such, multiple measures that relate to all of these considerations are likely to be needed to improve prediction of real behaviors that involve future motivation (see Frederick, Loewenstein, & O’Donoghue, 2002, for a related discussion). For example, in the area of health such additional measures are common, such as measures of the perceived connection between current behavior and future health goals, and they improve prediction of present behavior (Salovey, Rothman, & Rodin, 1998). Nonetheless, the motivation-based measure may prove to be a more broadly useful measure of temporal discounting than conventional money-based measures, in part because of the strategy people use in valuing future rewards on the motivation-based measure: an affect-based strategy. Affect-based strategies or heuristics are used for making a wide variety of judgments and choices as affective responses to objects are often used as proxies for value (Pham, 2004, Schwarz and Clore, 1996 and Slovic et al., 2002). This fact alone may broaden the predictive power of motivation-based measures, at least for behaviors and other measures that are largely determined by future concerns. Related to this, in the current research the motivation-based measures were somewhat more predictive of other measures of inter-temporal preference and temporal orientation (Studies 4 and 5), perhaps reflecting the use of affect-based heuristics or strategies on some of those measures. In particular, in Study 5, the relationships with participants’ temporal orientation was more positive for those participants who relied predominantly on a affect-based strategy for either the conventional money-based or motivation-based measures. Whether or not motivation-based measures have greater predictive power than conventional money-based discounting measures for a wider range of real behaviors that involve consideration of the future is a question for future research. But understanding what measures of discounting do and do not measure can only help us to understand when and why these measures are likely to be useful predictors of behaviors that involve consideration of the future.