Measurement Precision versus User Fatigue: Temporal Discounting in Politics

• In Event: Contemporary Issues in Public Opinion Research Methods

Abstract

Many important policy choices involve intertemporal tradeoffs. For example, reducing greenhouse gas emissions entails raising energy costs sooner for the benefit of limiting catastrophic climate change decades in the future. Intertemporal choice focuses attention on how people discount future costs and benefits. Discounting describes a change in the relative valuation of an outcome that occurs because the timing of the outcome changes.

Measuring discounting is challenging for survey respondents because it involves a long series of questions asking about intertemporal allocations (of money or collective benefits or costs). Typical measurement approaches used in behavioral economics can involve dozens of such user choices. These questions vary in ways important for calibration of the underlying measurement model, but the variation may appear minor, confusing, or repetitive to respondents.

This paper examines the tradeoff between using a large number of questions (to measure discounting with high precision) and avoiding respondent fatigue or confusion (to optimize data quality).

We separately measure discounting and smoothing preferences. We measure both the broader phenomenon of impatience (the directly observed choice of a (desirable) outcome sooner rather than later) and more specific causes of such impatient intertemporal tradeoffs, discounting and smoothing. (Smoothing costs or benefits over time is one important alternative consideration guiding intertemporal tradeoffs.) Measurement of impatience imposes lower cognitive burden and requires fewer questions; measurement of discounting captures a more precise concept.

To develop self-reported and behavioral measurement instruments of these concepts, we combine approaches from political science, behavioral economics, and psychology. Measuring both discounting and impatience involves asking people about outcomes of the same kind that occur at different times. Our simplest measure asks respondents how they prefer to allocate quantities of the same outcome (e.g., economic growth) between two timepoints (e.g., this year and next year). It is easy to administer and straightforward for the respondent to comprehend. But it does not distinguish discounting and smoothing as reasons for their answers. Our simple measures thus capture impatience broadly, not discounting specifically.

To isolate discounting, we then introduce a more complex measurement approach that controls for other processes that might affect intertemporal comparisons, including smoothing preference. Aiming to understand choice between present and future alternatives, behavioral economists have devoted a lot of attention to describing the extent to which people discount future benefits. One measurement strategy is the Convex Time Budget (CTB) approach developed by Andreoni and Sprenger (2012). It controls for the diminishing appeal of more and more of the same thing at the same time (which reveals the smoothing preference) in order to isolate the extent of discounting in intertemporal choice. We adapt the CTB approach to understand how the timing of economic growth and job creation affects economic evaluations.

We compare the CTB approach to our simpler measurement approaches in terms of response quality (manipulation check failures, amount of missing data, response times) and internal consistency (between-measure correlations). Moreover, we designed a graphical survey interface that represents intertemporal choices visually rather than with numerals. Respondents are randomly assigned to either the graphical or the standard numeral interface, allowing us to determine if the former can improve data quality. We included the instruments in a series of opt-in surveys and a probability sample designed to be representative of the U.S. population provided by NORC at the University of Chicago.

Authors

• Markus Prior, Princeton University

• Abdelaziz Mohammed Alsharawy, University of Texas Health Science Center at Houston