Are routes well that end well? Experienced utility of impeded simulator routes

https://doi.org/10.1016/j.trd.2021.102719Get rights and content

Highlights

  • Experimental manipulation of route segmentation related to impedance in a simulator.

  • Examining effects on remembered utility (i.e., global affect, speed, time, flow).

  • Capturing also moment affect through touch-sensitive pad that served as accelerator.

  • Despite equality in mean features, segmentations yielded different subjective experiences.

  • Notably the order of more or less impeding route sections affected global judgments.

  • Suggestive evidence for taking route segmentation into account in route assignment.

Abstract

A remaining challenge for optimizing car travel is that route assignments still contain temporary impedance. The question arises as to how the degree and the order of more and less impeding route sections influence experienced utility of routes. On driving simulator routes, participants were assigned one out of two mean speeds and one out of three segmentations (i.e., “Fast Slow,” “Constant,” and “Slow Fast”). Independent of assigned mean speed, routes with the segmentation “Slow Fast” were rated less arousing, more positive, faster and more flowing than routes with the segmentation “Fast Slow.” Further differences were revealed by an interaction between mean speed and route segmentation. In summary, optimized segmentation can increase the utility of impeded routes. This may increase the acceptance of traveler information systems and in turn subjective well-being due to more efficient travel. The study warrants further investigations of varying driving cycles under real-world conditions.

Introduction

Traveling by car is usually goal-directed. Drivers want to reach a destination guided by a desired arrival time. However, goal attainment of drivers is most often restrained by aversive travel conditions, which constitute so-called physical impedance (Novaco, Stokols, Campbell, & Stokols, 1979). Being physically impeded can be best described as being subject to an objective deviation from free-flow and direct traveling. It is best indexed by how rapid a driver moves through an environment (Schaeffer, Street, Singer, & Baum, 1988). In brief, higher impedance means slower speed. Exposure to travel-related impedance negatively affects subjective well-being and even health (Novaco et al., 1990, Novaco et al., 1991).

Traveler information systems (TIS) aim at supporting drivers with their goal attainment. Several strategies are used to minimize impedance on individual routes (e.g., user-optimal routing). Advances in intelligent transportation systems (i.e., facilitation of cooperative route choice based on coordination) even allow the social optimization of entire road networks in order to improve traffic flows (Peeta and Mahmassani, 1995, Cagara et al., 2015, Chopra et al., 2018). Especially this strategy prompted to reinvestigate the following remaining challenge of optimized route assignments.

Unavoidably, optimized route assignments and therefore route proposals still contain temporary impedance (i.e., low-speed route sections). For example, in the optimal solution for solving the Braess-Paradox (Braess, 1968), route assignments contain either an immediate or a later slow route section in order to achieve minimum average travel times. Socially optimal routing even requires a minority of drivers to comply with route proposals that contain more than average impedance (cf. Papageorgiou, Diakaki, Dinopoulou, Kotsialos, & Wang, 2003). Following decision-utility theory, it is unlikely that drivers comply with these route proposals as speed (i.e., mean speed referring to reaching a destination within a specific time) is highly weighted in a route decision situation and drivers are dragged by higher decision-utility on less impeding routes (cf. Bovy & Stern, 1990).

However, drivers also evaluate their trip while driving and upon reaching their destination. Thus, the utility of a route also encompasses the experience of the route in question. Experienced utility implies the evaluation of experiences at given times during an episode (i.e., moment-utility) and a global evaluation of the entire episode (i.e., remembered utility), which is aggregated from moment-utilities (Kahneman, 2000). Indeed, moment-utility is reduced through temporary impedance, similarly to decision-utility, and may deteriorate aggregated remembered utility. Deteriorated remembered utility may in turn negatively affect the acceptance of route assignments, the compliance with future route proposals, and the usage of TIS as it constitutes a decisive part in the evolution of decision-utility (cf. Berridge & O’Doherty, 2014; see also Garbinsky, Morewedge, & Shiv, 2014). However, remembered utility is also literally remembered (i.e., involves memory processes), thus, prone to biases. The aggregation of moment-utilities may yield, at goal attainment and depending on the specific nature of impedance, global judgments that are more or less conducive to future choices. This is one of the reasons why transport researchers have argued to place greater emphasis on this type of travel-related utility (Ettema, Gärling, Olsson, & Friman, 2010), which extends to judging criteria, such as affect, perceived speed, time, or flow (Mokhtarian and Salomon, 2001, Redmond and Mokhtarian, 2001, Ettema et al., 2013, Susilo and Cats, 2014).

This study is concerned with the perception of impedance and the aggregation of objectively reduced moment-utility into subjective remembered utility of assigned routes. The question remains as to how routes, which vary in the specific nature of impedance, induce varying subjective experiences.

Lower perceived velocity during goal attainment is generally associated with a more negative affect (Hsee and Abelson, 1991, Hsee et al., 1994). In particular for the driving context, temporal goal-oriented participants rated singular impedance situations (i.e., caused by traffic lights or stationary vehicles) more arousing, more negatively and less under control than preceding parts of videos depicting car travel (Cœugnet et al., 2013).

However, for the aggregation of such aversive moments into global affective judgments related to valence and arousal, empirical evidence from related domains (i.e., research on aversiveness or stress inducing stimuli, and consumer research) indicates that judgment heuristics are used. These heuristics take the relation between and especially the order of more or less aversive events during an episode into account. In particular, moment-utility during end moments (i.e., end-rule), or end together with peak pleasant or unpleasant moments (i.e., peak-end-rule) were shown to dominate the aggregation into global judgments (for a review, see Fredrickson, 2000). Thus, episodes that contain an end with increased (decreased) moment-utility or a (less) conducive peak-end ratio are evaluated better (worse) than objectively expected.

Recent findings suggest indeed that the explanatory power of these judgment heuristics might be limited to the evaluation of continuous (versus discrete) events (Ariely & Zauberman, 2000), discrepant, i.e., salient and distinctive (versus similarly) negative or positive events (Seta et al., 2008a, Seta et al., 2008b), episodes that have clear ends with an explicit meaning (Miron-Shatz, 2009), episodes that consist of a large number of moment-utilities (Suzuki, Fujii, Gärling, Ettema, Olsson, & Friman, 2014), and repeated episodes in which the variation of the sequence of succeeding events is the only manipulated factor (Tully & Meyvis, 2016). However, car travel seem to generally fulfill these criteria. Although it was argued before that travelers might consider travel related events discrete, especially if salient changes appear only occasionally, car travel might be perceived as rather continuously with a high cohesiveness among a large number of single events. Further, car travel contains most likely discrepancy across singular route sections (i.e., varying speed), clear ends (i.e., arriving at a destination), and is repeated within trip categories (e.g., commutes). This might prevent drivers from a normatively correct aggregation of affect at singular moments.

For affect in terms of control (i.e., feeling of power, dominance, locus of control) and related to car travel, Koslowsky (1997) argued that it is mostly affected by the confirmation or disconfirmation of an expected control level. With regard to varying levels of assigned impedance, control expectations might be formed according to the ability of adjusting speed towards an appropriate level. These expectations might be adapted, namely confirmed, or positively (i.e., more appropriate succeeding less appropriate speed) or negatively (i.e., less appropriate succeeding more appropriate speed) disconfirmed throughout a route (cf. Oliver, 1977). This discrete evaluation with respect to the order of conducive control levels might in turn dominate global control judgments. For example, a negative disconfirmation might entail an excessively negative global judgment, whereas a positive disconfirmation reveals the opposite. In addition, Koslowsky (1997, referring to Murphy, 1988) suggested that it might be perceived more negatively never to have had control over a situation than to loose perceived control throughout an episode. Thus, constant speed and therefore confirmation of missing perceived speed control might be perceived more passive than a negative disconfirmation.

Global speed and temporal judgments of route assignments that include differently impeding route sections were previously investigated by Leiser, Stern, and Meyer (1991). Participants estimated mean speed and the duration of assigned routes that were identical in total time and distance, but contained driving cycles with varied order of slow and fast route sections (i.e., three sections in total). Speed judgments were dependent on mean speed: the higher the number of fast route sections, the higher mean speed estimates. More importantly, similarly to global affective judgments, mean speed estimates were subject to order effects. Participants estimated speeds for routes with slow route sections at the end or at the beginning significantly lower than for routes with fast route sections at the end or at the beginning. In support of their expectation that perception of time is rather cumulative than selective (i.e., accumulation of time is not disproportionate in specific moments), the authors found that prospective duration estimations were not affected by order effects.

However, in the experiment of Leiser et al. (1991) participants first did not control their own speed, they only observed speed, and second lacked a temporal goal. As such, it remains unclear how increased engagement in driving (i.e., speed control) and likely increased task difficulty influence duration estimates and, moreover, a distinct (not observed by Leiser et al. (1991), but important) feeling of the passage of time. In fact, prospective duration judgments as well as judgments of how lengthy passaged time felt were found to shorten if time periods become more content-rich, imply increased non-temporal task engagement and task difficulty (i.e., difficulty of processing during these periods increases), and consequently less attention is allocated to time (Zakay and Block, 1997, Wearden et al., 2014, Wearden, 2015). For appraisals of more goal-directed travelers, Li (2003) stated that travelers might disproportionately weigh impeding moments in global temporal judgments due to relatively strongly negative affective states (i.e., peaks). Negative affective states, however, revealed inconsistent effects on temporal judgments. In particular, findings in relation to negative affect ranged from a small tendency towards overestimation (Hornik, 1992) to overestimation (Droit-Volet, Fayolle, & Gil, 2011), and, in the driving context, even underestimation of durations (Cœugnet et al., 2013). Rates of time passage were shown to be lowered under negative affective states (Droit-Volet & Wearden, 2016). Positive mood states, on the other hand, were shown to be associated with underestimation of episode durations (Hornik, 1992) and likewise higher rates of time passage (Droit-Volet & Wearden, 2016). Li (2003) additionally argued that especially affect during end moments should bias temporal judgments due to travelers’ predisposition towards evaluating travels particularly towards the end. Thus, it remains to be shown how temporal judgments of particularly temporal goal-directed drivers are influenced by differently impeding route sections.

Finally, discrepancies in speed (i.e., routes with discrepant route sections) per se imply an interference with the continuous moving towards a goal (i.e., flow). According to Koslowsky (1997), this particular experience of a sense of motion during a journey adds a distinct quality to a route profile (see also Kluger, 1998). For example, drivers sometimes adapt their route choice throughout a journey and even accept traveling longer distances simply in order to keep that sense of moving. Thus, driving flow is considered here a seamless succession of events (Novak et al., 1996), which includes unambiguous feedback in terms of goal pursuit and speed control (cf. Csikszentmihalyi, 2014). In contrast to feeling the power of controlling speed (i.e., affect in terms of control), it pertains to exercising driving control effortlessly with a high predictability, in particular, “lacking the sense of worry about losing control” (Csikszentmihalyi & Csikzentmihaly, 1990). Yet, similarly to control, any change in speed might disconfirm flow and discrete confirmation or disconfirmation might be excessively weighed in global flow judgments. In particular, disconfirmation of initially experienced speed might impair global flow judgments.

The question arises as to how the degree and the order of more and less impeding route sections (i.e., discrepant, thus, slower or faster sections, or similar speed) within objectively similar routes (i.e., in terms of assigned mean speed, distance and duration) influence global affective (i.e., valence, arousal, and control), speed, temporal, and flow judgments of these routes.

In order to examine this question, the author and research colleagues conducted a driving simulation experiment. Participants judged differently segmented routes with driving cycles that were constrained to a maximum speed (i.e., bisection with slow route section succeeding fast route section referred to as “Fast Slow,” constant speed across both sections referred to as “Constant,” or fast route section succeeding slow route section referred to as “Slow Fast”). Expanding on work by Leiser et al., 1991, Cœugnet et al., 2013, judgments were based on self-produced movement of a simulated vehicle. In particular, temporal goal-directed participants were in control of longitudinal speed, however, without knowing about the maximum speed constraint. Acceleration behavior was assessed on-line. Further, we investigated several judging criteria (i.e., arousal, valence, control, perceived speed, passage of time and duration, and flow) related to the routes. Moreover, two mean speed assignments were tested. One that was similar to permitted free-flow speed on the type of simulated road and one that was considerably lower. Thus, distinct degrees of impedance were induced by lower versus higher mean speed and, throughout a route, by assigning more or less temporary impedance.

First, for impedance related to mean speed, we hypothesized that routes with considerably lower than expectable mean speed yield worse ratings than routes with mean speed closer to permitted free-flow speed on the type of simulated road. This means, higher arousal, worse valence, less actively control, lengthier passage of time and longer duration, slower speed, and higher flow ratings (H1).

Second, we expected that participants weight impeding moments disproportionately with respect to the order of more and less temporary impedance (H2). This means that the relation between global judgments differs from normatively correct Fast Slow = Constant = Slow Fast (proposition P0), which is expected if participants weight more and less impeding moments discretely and equally, and no order effects are revealed (cf. Ariely & Zauberman, 2000). Table 1 summarizes expected relations between global judgments resulting from order effects revealed by the proposed aggregation rules for each judgment criterion.

Section snippets

Participants

The participants were 45 persons that had been acquired through a dedicated recruiting system run by Humboldt-Universität zu Berlin. All participants had a valid driving license. Their age ranged between 20 and 73 years (M = 33.89, SD = 15.05). Twenty-nine participants (64%) were female. Thirty-four participants (76%) owned a car. Median car usage for all participants was three days a week. Almost all (42) finished high school. Half of the participants (22) were students, the other regularly

Preliminary analyses

Several measurements indicate the validity of the simulated scenario of driving and accelerating on an empty German two-lane expressway. The mean expected speed in km/h (M = 101.82, SD = 24.04) equaled the permitted free-flow speed on German two-lane expressways (i.e., 100 km/h). Participants would correspondingly choose their own speed (M = 98.78, SD = 15.45), whereas they stated safe speed to be lower (M = 83.89, SD = 15.52). Indeed, the mean comfortable speed at which participants

Discussion

In order to investigate the effects of the degree and the order of more and less impeding route sections on global affective, speed, temporal, and flow judgments of routes, a driving simulation experiment was conducted. Participants were assigned two mean speeds and differently segmented routes on a two-lane expressway (i.e., slow route section succeeding fast route section, constant speed, and fast route section succeeding slow route section).

First, significant main effects of mean speed were

Conclusion

The remaining challenge of assigning and proposing routes that contain temporary impedance for drivers prompted to investigate subjective experiences induced by different route segmentations. We found suggestive evidence that optimized segmentation (i.e., relatively less impeding ends and partly also constant flow) yielded less negative (or more positive) route evaluations in terms of affect, perceived speed, experienced time, and flow. This may be conducive to decision-utility of and

Acknowledgements

We wish to thank Dr. John E. Anderson for proofreading and the referees for their valuable comments on the manuscript. We thank our fellows at the BIGS2 research school and the Department of Cognitive Psychology at Humboldt-Universität zu Berlin for their constructive criticism, and especially Olaf Menzel and Roman Wolter for helping with the experiments. The author is grateful for the support by the BIGS2 research school at Humboldt-Universität zu Berlin. The author has no conflict of interest

References (82)

  • C.K. Hsee et al.

    The quasi-acceleration relation: Satisfaction as a function of the change of velocity of outcome over time

    J. Exp. Soc. Psychol.

    (1994)
  • D. Leiser et al.

    Mean velocity and total time estimation effects of order and proportions

    J. Environmental Psychology

    (1991)
  • B. Lewis-Evans et al.

    Task difficulty, risk, effort and comfort in a simulated driving task —Implications for Risk Allostasis Theory

    Accid. Anal. Prev.

    (2009)
  • Y.W. Li

    Evaluating the Urban Commute Experience: A Time Perception Approach

    J. Public Transportation

    (2003)
  • P.L. Mokhtarian et al.

    How derived is the demand for travel? Some conceptual and measurement considerations

    Transp. Res. Part A

    (2001)
  • D.A. Redelmeier et al.

    Patients' Memories of Painful Medical Treatments: Real-time and Retrospective Evaluations of Two Minimally Invasive Procedures

    Pain

    (1996)
  • M. Ringhand et al.

    Investigating urban route choice as a conflict between waiting at traffic lights and additional travel time

    Transp. Res. Procedia

    (2017)
  • J.A. Russell et al.

    Adaptation level and the affective appraisal of environments

    J. Environmental Psychology

    (1984)
  • J.J. Seta et al.

    Averaging and summation: Positivity and choice as a function of the number and affective intensity of life events

    J. Exp. Soc. Psychol.

    (2008)
  • Y.O. Susilo et al.

    Exploring Key Determinants of Travel Satisfaction for Multi-Modal Trips by Different Traveler Groups

    Transp. Res. Part A

    (2014)
  • J. Törnros

    Driving behaviour in a real and a simulated road tunnel—a validation study

    Accid. Anal. Prev.

    (1998)
  • T. Tozman et al.

    Understanding the psychophysiology of flow: A driving simulator experiment to investigate the relationship between flow and heart rate variability

    Comput. Hum. Behav.

    (2015)
  • J.H. Wearden

    Passage of time judgements

    Conscious. Cogn.

    (2015)
  • Akai professional M.I. K.K.

    MPD32 USB/MIDI Pad Control Unit: Operator’s Manual

    (2008)
  • D. Ariely et al.

    On the making of an experience: The effects of breaking and combining experiences on their overall evaluation

    J. Behavioral Decision Making

    (2000)
  • J.R. Averill

    Personal control over aversive stimuli and its relationship to stress

    Psychol. Bull.

    (1973)
  • D. Bates et al.

    Fitting Linear Mixed-Effects Models Using lme4

    J. Stat. Softw.

    (2015)
  • A. Bhattacherjee

    Understanding Information Systems Continuance: An Expectation-Confirmation Model

    MIS Quarterly

    (2001)
  • P.H. Bovy et al.

    Route Choice: Wayfinding in Transport Networks

    (1990)
  • D. Braess

    Über ein Paradoxon aus der Verkehrsplanung

    Unternehmensforschung

    (1968)
  • P.C. Burns et al.

    Measuring distraction: Task duration and the lane-change test (LCT)

    Proc. Human Factors Ergonomics Society Annual Meeting

    (2005)
  • D. Cagara et al.

    Traffic optimization on Islands

    2015 IEEE Vehicular Networking Conference (VNC

    (2015)
  • Chopra, A., Kalhan, D. S., Bedi, A. S., Gupta, A. K., & Rajawat, K. (2018). On Socially Optimal Traffic Flow in the...
  • Y. Croissant

    mlogit: Multinomial Logit Models

    R package version

    (2020)
  • M. Csikszentmihalyi

    Flow and the Foundations of Positive Psychology: The Collected Works of Mihaly Csikszentmihalyi

    (2014)
  • M. Csikszentmihalyi et al.

    Flow: The psychology of optimal experience

    (1990)
  • S. Droit-Volet et al.

    Emotion and time perception: effects of film-induced mood

    Front. Integr. Neurosci.

    (2011)
  • S. Droit-Volet et al.

    Passage of time judgments are not duration judgments: Evidence from a study using experience sampling methodology

    Front. Psychol.

    (2016)
  • L. Evans

    Speed estimation from a moving automobile

    Ergonomics

    (1970)
  • Exposito, M., Hernandez, J., & Picard, R. W. (2018). Affective keys: towards unobtrusive stress sensing of smartphone...
  • B.L. Fredrickson

    Extracting meaning from past affective experiences: The importance of peaks, ends, and specific emotions

    Cogn. Emot.

    (2000)
  • Cited by (1)

    View full text