Dynamic workload measurement and modeling: Driving and conversing

Castro, SC; Heathcote, A; Cooper, JM; Strayer, DL · 2023 · publications_jsonl

DOI: 10.1037/xap0000431

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the dynamic cognitive workload associated with driving while engaging in natural conversations, addressing a limitation in prior research that collapsed speaking and listening phases. Previous work by Tillman et al. (2017) suggested that conversing while driving primarily increased response caution (threshold) without altering the rate of evidence accumulation. However, because speech production and comprehension impose different cognitive demands, averaging these phases may have masked critical fluctuations. The current research aims to disentangle these effects by measuring workload dynamically for both drivers and non-drivers during in-person and hands-free cell phone conversations. The experiment involved 44 participants organized into 22 dyads who engaged in 15-minute driving simulations while conversing. Participants were randomly assigned to either an in-person passenger condition or a remote cell phone condition. Workload was measured using a Detection Response Task (DRT), where vibrotactile stimuli were presented every 3–5 seconds, requiring participants to press a button. Microphones recorded audio to code whether the driver or non-driver was speaking, listening, or silent at the moment of each stimulus. Driving performance was also monitored via steering deviation and speed variability. The researchers employed the Linear Ballistic Accumulator with Omissions (LBAO) model to analyze response times and omission rates, allowing for the simultaneous estimation of response threshold (caution), drift rate (evidence accumulation speed), and non-decision time. Behavioral results indicated that conversing significantly increased response times and decreased hit rates compared to silence, with drivers exhibiting higher workload than non-drivers. Crucially, response times were longer when participants were speaking than when listening. A significant interaction revealed that drivers found it more difficult to listen to cell phone conversations than in-person ones. Modeling results demonstrated that the dynamic ebb and flow of conversation altered both the rate of evidence accumulation and the response threshold. Specifically, speaking increased response caution and decreased the rate of evidence accumulation for both drivers and non-drivers. The contribution analysis showed that changes in response threshold accounted for the majority of the variance in response time differences between speaking and listening conditions, particularly for drivers with passengers. The findings signify that cognitive workload during driving and conversing is not static but fluctuates rapidly with task demands. The study confirms that driving and conversing compete for limited attentional resources, with speech production imposing a heavier cognitive load than comprehension. The increased workload is driven primarily by a strategic increase in response caution and a secondary decrease in processing speed. These results highlight the importance of considering the dynamic nature of dual-task performance, suggesting that the impairment in driving caused by conversation is most pronounced during speech production and is exacerbated in cell phone conditions due to the lack of shared contextual cues available in in-person interactions.

Key finding

Drivers showed elevated DRT RT relative to non-drivers, and within both members of the dyad RT was higher when speaking than when listening, with conversational turn-taking producing reciprocal workload trade-offs. Unlike Tillman et al. (2017), who attributed dual-task cost solely to increased response threshold, LBAO modelling here showed that conversing while driving altered both the response threshold and the rate of evidence accumulation, consistent with a combined response-caution and capacity-sharing account; aggregating across speaking and listening in earlier work likely masked the drift-rate effect.

Methodology

simulator

Sample size: N=44 (22 dyads, 23 female, M age=21.1, SD=3.4), University of Utah undergraduates

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via tag_papers on 2026-05-30 (4 acquisition events logged).

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• dual task performance
• workload measurement
• mental demand
• temporal
• auditory
• cognitive

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: behavioral performance data
• Theoretical Contribution: theory or model, computational model