A Sound Decision? The Impact of Audio Presentations on Economic Rationality

Participants are randomly assigned to one of three treatment conditions, which differ in the speed at which auditory presentations of decision problems are delivered: slow, moderate, or fast.

Participants can click on an option in all treatment conditions to access its auditory information. Upon clicking, they hear two numbers [x,y], representing a lottery that pays either x or y with equal probability. Each digit is articulated sequentially, with a 0.5-second pause between x and y. For example, for the option [0,194], the digits are spoken as: "0" (zero), followed by "1" (one hundred), "9" (ninety), and "4" (four), with each digit enunciated individually at a fixed speed. Participants can perceive only one option at a time by clicking on it, as often as needed. All auditory content is delivered in Mandarin Chinese.

The key difference across treatments is the speed at which the auditory information is presented, corresponding to approximately 90 (slow), 120 (moderate), and 150 (fast) words per minute, respectively.

2025-05-28

2026-05-28

Choice consistency

Choice consistency is a binary variable equal to 1 if a participant’s choices satisfy the Generalized Axiom of Revealed Preference (GARP). This indicates consistency with preference maximization as provided by revealed preference theory (Afriat, 1967; Varian, 1982). We hypothesize that faster presentation impairs decision-making, leading to lower choice consistency. Specifically:

H1: Choice consistency is higher in the slow-speed condition than in the moderate-speed condition.

H2: Choice consistency is higher in the slow-speed condition than in the fast-speed condition.

H3: Choice consistency is higher in the moderate-speed condition than in the fast-speed condition.

References:
Afriat, Sidney N. 1967. “The Construction of Utility Functions from Expenditure Data.” International Economic Review 8 (1): 67–77.
Varian, Hal R. 1982. “The Nonparametric Approach to Demand Analysis.” Econometrica 50 (4): 945–973.

The Houtman–Maks Index (HMI; Houtman and Maks, 1985); the number of GARP violations; more stringent variants of the above measures that additionally satisfy first-order stochastic dominance (FOSD); participants’ response times; and search behavior patterns, including the number of options considered.

References:
Houtman, Martijn, and Julian Maks. 1985. “Determining All Maximal Data Subsets Consistent with Revealed Preference.” Kwantitatieve Methoden 19 (1): 89–104.

We include several robustness metrics. The HMI is the minimum number of choices that must be removed from a dataset to render the remaining set GARP-consistent; these removed observations are typically interpreted as choice mistakes. The number of GARP violations similarly quantifies deviations from economic rationality. Higher HMI scores or more GARP violations indicate lower levels of rationality. As further robustness checks, we consider FOSD-compliant variants, which impose a stricter normative benchmark for evaluating decision-making under risk. For exploratory analysis, we also examine response times and search patterns (e.g., the number of options clicked) as proxies for decision effort.

The experiment comprises three parts. In the first, participants make economic decisions based on their assigned treatment. In the second, they complete a series of cognitive tasks. In the third, they complete a set of questionnaires assessing attitudes toward inconsistency and demographic characteristics.

The experiment will be conducted in the Zhejiang University laboratory via the Qualtrics platform, with participants recruited from the laboratory’s participant pool. In the first part, participants face twenty risky decision problems. Each problem presents eleven distinct options in random order, with each option representing a lottery that yields one of two monetary prizes with equal probability. In the second part, cognitive ability is measured using the ICAR test (Condon and Revelle, 2014), selective attention with the Stroop task (Stroop, 1935), and working memory with the Sternberg task (Sternberg, 1966). In the third part, participants rate their comfort with a choice inconsistency scenario on a scale from 0 (least at ease) to 10 (most at ease), where higher scores indicate a less negative attitude. The scenario illustrates the attraction effect, a well-documented form of choice inconsistency (Huber, Payne, and Puto, 1982).

References:
Condon, David, and William Revelle. 2014. “The International Cognitive Ability Resource: Development and Initial Validation of a Public-Domain Measure.” Intelligence 43: 52–64.
Sternberg, Saul. 1966. “High-Speed Scanning in Human Memory.” Science 153 (3736): 652–654.
Stroop, J. Ridley. 1935. “Studies of Interference in Serial Verbal Reactions.” Journal of Experimental Psychology 18 (6): 643–662.
Huber, Joel, John W. Payne, and Christopher Puto. 1982. “Adding Asymmetrically Dominated Alternatives: Violations of Regularity and the Similarity Hypothesis.” Journal of Consumer Research 9 (1): 90–98.

Randomization will be implemented via Qualtrics.

Individual

No

360 participants

360 participants

120 individuals in each treatment condition

Power is assessed via simulation using data from previous trials. We apply logistic regression to 10,000 simulated samples across a range of effect sizes for the primary outcome, choice consistency. Assuming 120 participants per treatment condition, 80% power, and a 5% significance level, each hypothesis's minimum detectable effect size is 0.83 log-odds, corresponding to an odds ratio of 2.3, or a 130% increase in the odds of choice consistency.