Portal de Periódicos da CAPES

Brain Drain: The Mere Presence of One’s Own Smartphone Reduces Available Cognitive Capacity

2017; University of Chicago Press; Volume: 2; Issue: 2 Linguagem: Inglês

10.1086/691462

2378-1823

Adrian F. Ward, Kristen Duke, Ayelet Gneezy, Maarten W. Bos,

Media Influence and Health

Previous articleNext article FreeBrain Drain: The Mere Presence of One’s Own Smartphone Reduces Available Cognitive CapacityAdrian F. Ward, Kristen Duke, Ayelet Gneezy, and Maarten W. BosAdrian F. Ward, Kristen Duke, Ayelet Gneezy, and Maarten W. BosPDFPDF PLUSAbstractFull TextSupplemental Material Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinked InRedditEmailQR Code SectionsMoreAbstractOur smartphones enable—and encourage—constant connection to information, entertainment, and each other. They put the world at our fingertips, and rarely leave our sides. Although these devices have immense potential to improve welfare, their persistent presence may come at a cognitive cost. In this research, we test the “brain drain” hypothesis that the mere presence of one’s own smartphone may occupy limited-capacity cognitive resources, thereby leaving fewer resources available for other tasks and undercutting cognitive performance. Results from two experiments indicate that even when people are successful at maintaining sustained attention—as when avoiding the temptation to check their phones—the mere presence of these devices reduces available cognitive capacity. Moreover, these cognitive costs are highest for those highest in smartphone dependence. We conclude by discussing the practical implications of this smartphone-induced brain drain for consumer decision-making and consumer welfare.We all understand the joys of our always-wired world—the connections, the validations, the laughs … the info. … But we are only beginning to get our minds around the costs.Andrew Sullivan (2016)The proliferation of smartphones has ushered in an era of unprecedented connectivity. Consumers around the globe are now constantly connected to faraway friends, endless entertainment, and virtually unlimited information. With smartphones in hand, they check the weather from bed, trade stocks—and gossip—while stuck in traffic, browse potential romantic partners between appointments, make online purchases while standing in-store, and live-stream each others’ experiences, in real time, from opposite sides of the globe. Just a decade ago, this state of constant connection would have been inconceivable; today, it is seemingly indispensable.1 Smartphone owners interact with their phones an average of 85 times a day, including immediately upon waking up, just before going to sleep, and even in the middle of the night (Perlow 2012; Andrews et al. 2015; dscout 2016). Ninety-one percent report that they never leave home without their phones (Deutsche Telekom 2012), and 46% say that they couldn’t live without them (Pew Research Center 2015). These revolutionary devices enable on-demand access to friends, family, colleagues, companies, brands, retailers, cat videos, and much more. They represent all that the connected world has to offer, condensed into a device that fits in the palm of one’s hand—and almost never leaves one’s side.The sharp penetration of smartphones, both across global markets and into consumers’ everyday lives, represents a phenomenon high in “meaning and mattering” (e.g., Kernan 1979; Mick 2006)—one that has the potential to affect the welfare of billions of consumers worldwide. As individuals increasingly turn to smartphone screens for managing and enhancing their daily lives, we must ask how dependence on these devices affects the ability to think and function in the world off-screen. Smartphones promise to create a surplus of resources, productivity, and time (e.g., Turkle 2011; Lee 2016); however, they may also create unexpected deficits. Prior research on the costs and benefits associated with smartphones has focused on how consumers’ interactions with their smartphones can both facilitate and interrupt off-screen performance (e.g., Isikman et al. 2016; Sciandra and Inman 2016). In the present research, we focus on a previously unexplored (but common) situation: when smartphones are not in use, but are merely present.We propose that the mere presence of one’s own smartphone may induce “brain drain” by occupying limited-capacity cognitive resources for purposes of attentional control. Because the same finite pool of attentional resources supports both attentional control and other cognitive processes, resources recruited to inhibit automatic attention to one’s phone are made unavailable for other tasks, and performance on these tasks will suffer. We differentiate between the orientation and allocation of attention and argue that the mere presence of smartphones may reduce the availability of attentional resources even when consumers are successful at controlling the conscious orientation of attention.Cognitive Capacity and Consumer BehaviorConsumers’ finite capacity for cognitive processing is one of the most fundamental influences on “real world” consumer behavior (e.g., Bettman 1979; Bettman, Johnson, and Payne 1991). Individuals are constantly surrounded by potentially meaningful information; however, their ability to use this information is consistently constrained by cognitive systems that are capable of attending to and processing only a small amount of the information available at any given time (e.g., Craik and Lockhart 1972; Newell and Simon 1972). This capacity limit shapes a wide range of behaviors, from in-the-moment decision-making strategies and performance (e.g., Lane 1982; Lynch and Srull 1982) to long-term goal pursuit and self-regulation (e.g., Hofmann, Strack, and Deutsch 2008; Benjamin, Brown, and Shapiro 2013).Consumers’ cognitive capabilities—and constraints—are largely determined by the availability of domain-general, limited-capacity attentional resources associated with both working memory and fluid intelligence (e.g., Halford, Cowan, and Andrews 2007; Jaeggi et al. 2008). “Working memory” (WM) refers to the theoretical cognitive system that supports complex cognition by actively selecting, maintaining, and processing information relevant to current tasks and/or goals. “Working memory capacity” (WMC) reflects the availability of attentional resources, which serve the “central executive” function of controlling and regulating cognitive processes across domains (Baddeley and Hitch 1974; Miyake and Shah 1999; Engle 2002; Baddeley 2003). “Fluid intelligence” (Gf) represents the ability to reason and solve novel problems, independent of any contributions from acquired skills and knowledge stored in “crystallized intelligence” (Cattell 1987). Similar to WM, Gf stresses the ability to select, store, and manipulate information in a goal-directed manner. Also similar to WM, Gf is constrained by the availability of attentional resources (e.g., Engle et al. 1999; Halford et al. 2007). Crucially, the limited capacity of these domain-general resources dictates that using attentional resources for one cognitive process or task leaves fewer available for other tasks; in other words, occupying cognitive resources reduces available cognitive capacity.Given the chronic mismatch between the abundance of environmental information and the limited ability to process that information, individuals need to be selective in their allocation of attentional resources (e.g., Kahneman 1973; Johnston and Dark 1986). The priority of a stimulus—that is, the likelihood that it will attract attention—is determined by both its physical “salience” (e.g., location, perceptual contrast) and its goal “relevance” (i.e., potential importance for goal-directed behavior) (e.g., Corbetta and Shulman 2002; Fecteau and Munoz 2006). Preferential attention to temporarily relevant stimuli, such as those associated with a current task or decision, is supported by WM; when a goal is active in WM, stimuli relevant to that goal are more likely to attract attention (e.g., Moskowitz 2002; Soto et al. 2005; Vogt et al. 2010). Frequently relevant stimuli, such as those associated with long-term and/or self-relevant goals, may automatically attract attention even when the goals associated with these stimuli are not active in WM (Shiffrin and Schneider 1977; Johnston and Dark 1986); for example, individuals automatically orient to the sounds of their own names in ignored audio channels (Moray 1959), and mothers, more so than nonmothers, automatically attend to infants’ emotional expressions (Thompson-Booth et al. 2014). Automatic attention generally helps individuals make the most of their limited cognitive capacity by directing attention to frequently goal-relevant stimuli without requiring these goals to be constantly kept in mind. However, automatic attention may undermine performance when an environmental stimulus is frequently relevant to an individual’s goals but currently irrelevant to the task at hand; inhibiting automatic attention—keeping attractive but task-irrelevant stimuli from interfering with the contents of consciousness—occupies attentional resources (e.g., Engle 2002).Smartphones serve as consumers’ personal access points to all the connected world has to offer. We suggest that the increasing integration of these devices into the minutiae of daily life both reflects and creates a sense that they are frequently relevant to their owners’ goals; it lays the foundation for automatic attention. Consistent with this position, research indicates that signals from one’s own phone (but not someone else’s) activate the same involuntary attention system that responds to the sound of one’s own name (Roye, Jacobsen, and Schröger 2007). When these devices are salient in the environment, their status as high-priority (relevant and salient) stimuli suggests that they will exert a gravitational pull on the orientation of attention. And when consumers are engaged in tasks for which their smartphones are task-irrelevant, the ability of these devices to automatically attract attention may undermine performance in two ways (Clapp, Rubens, and Gazzaley 2009; Clapp and Gazzaley 2012). First, smartphones may redirect the orientation of conscious attention away from the focal task and toward thoughts or behaviors associated with one’s phone. Prior research provides ample evidence that individuals spontaneously attend to their phones at inopportune times (e.g., Oulasvirta et al. 2011), and that this digital distraction adversely affects both performance (End et al. 2009) and enjoyment (Isikman et al. 2016). Second, smartphones may redistribute the allocation of attentional resources between engaging with the focal task and inhibiting attention to one’s phone. Because inhibiting automatic attention occupies attentional resources, performance on tasks that rely on these resources may suffer even when consumers do not consciously attend to their phones. We explore this possibility in the current research.Smartphone Use and Conscious Distraction (the Orientation of Attention)Research on the relationship between mobile devices and cognitive functioning has largely focused on downstream consequences of device-related changes in the orientation of attention. For example, research on mobile device use while driving indicates that interacting with one’s phone while behind the wheel causes performance deficits such as delayed reaction times and inattentional blindness (e.g., Strayer and Johnston 2001; Caird et al. 2008); these deficits mirror those associated with distracting “live” conversations (Recarte and Nunes 2003). Similarly, research in the educational sphere demonstrates that using mobile devices and social media while learning new material reduces comprehension and impairs academic performance (e.g., Froese et al. 2012). However, mobile device use does not affect performance on self-paced tasks, which allow individuals to compensate for device-related distractions by picking up where they left off (e.g., Fox, Rosen, and Crawford 2009; Bowman et al. 2010). Taken together, these findings suggest that many of the cognitive impairments associated with mobile device use may simply represent the general deleterious effects of diverting conscious attention away from a focal task. What may be special about smartphones, however, is the frequency with which they seem to create these diversions; their omnipresence and personal relevance may combine to create a particularly potent draw on the orientation of attention.A more limited body of work explores the cognitive consequences of smartphone-related distractions in the absence of behavioral interaction (i.e., when consumers consciously think about phone-related stimuli, but do not actually use their phones). Research on the attentional cost of receiving cellphone notifications indicates that awareness of a missed text message or call impairs performance on tasks requiring sustained attention, arguably because unaddressed notifications prompt message-related (and task-unrelated) thoughts (Stothart, Mitchum, and Yehnert 2015). Related research shows that individuals who hear their phones ring while being separated from them report decreased enjoyment of focal tasks as a consequence of increased attention to phone-related thoughts (Isikman et al. 2016). Forced separation from one’s ringing phone can also increase heart rate and anxiety and decrease cognitive performance (Clayton, Leshner, and Almond 2015). To our knowledge, only one prior study has investigated the cognitive effects of the mere presence of a mobile device—one that is not ringing, buzzing, or otherwise actively interfering with a focal task. Thornton et al. (2014, 485–86) found that a visually salient cellphone can impair performance on tasks requiring sustained attention by eliciting awareness of the “broad social and informational network … that one is not part of at the moment.” Together, these investigations of phone-related distractions provide evidence that mobile devices can adversely affect cognitive performance even when consumers are not actively using them. Similar to earlier research on distracted driving and learning while multitasking, however, these studies connect the cognitive costs of smartphones to their (remarkable) ability to attract the conscious orientation of attention. When individuals interact with or think about their phones rather than attend to the task at hand, their performance suffers.Smartphone Presence and Cognitive Capacity (the Allocation of Attentional Resources)We suggest that smartphones may also impair cognitive performance by affecting the allocation of attentional resources, even when consumers successfully resist the urge to multitask, mind-wander, or otherwise (consciously) attend to their phones—that is, when their phones are merely present. Despite the frequency with which individuals use their smartphones, we note that these devices are quite often present but not in use—and that the attractiveness of these high-priority stimuli should predict not just their ability to capture the orientation of attention, but also the cognitive costs associated with inhibiting this automatic attention response.We propose that the mere presence of one’s smartphone may impose a “brain drain” as limited-capacity attentional resources are recruited to inhibit automatic attention to one’s phone, and are thus unavailable for engaging with the task at hand. Research on controlled versus automatic processing provides evidence that the mere presence of personally relevant stimuli can impair performance on cognitive tasks (e.g., Geller and Shaver 1976; Bargh 1982; Wingenfeld et al. 2006). Importantly, these performance deficits occur without conscious attention to the potentially interfering stimuli and as a function of inhibiting these stimuli from interfering with the contents of consciousness (e.g., Shallice 1972; Lavie et al. 2004). Consistent with this evidence, we posit that the mere presence of consumers’ own smartphones can reduce the availability of attentional resources (i.e., cognitive capacity) even when consumers are successful at controlling the conscious orientation of attention (i.e., resisting overt distraction).If smartphones undermine cognitive performance by occupying attentional resources, the cognitive consequences of smartphone presence should be sensitive to variation in both the salience and the personal relevance of these devices, which together determine their priority in attracting attention (e.g., Fecteau and Munoz 2006). Prior research suggests that smartphones are chronically salient for many individuals, even when they are located out of sight in one’s pocket or bag (e.g., Deb 2015). However, we expect that increasing the salience of one’s smartphone—for example, by placing it nearby and in the field of vision—will amplify the cognitive costs associated with its presence, as more attentional resources are required to inhibit its influence on the orientation of attention. We also expect that these costs will vary according to the personal relevance of one’s smartphone. We operationalize relevance in terms of “smartphone dependence,” or the extent to which individuals rely on their phones in their everyday lives. We posit that individual differences in dependence on one’s smartphone will moderate the effects of smartphone salience on available cognitive capacity, such that individuals who most depend on their phones will suffer the most from their presence—and benefit the most from their absence.Overview of the ExperimentsIn two experiments, we test the hypothesis that the mere presence of one’s own smartphone reduces available cognitive capacity. We manipulate smartphone salience by asking participants to place their devices nearby and in sight (high salience, “desk” condition), nearby and out of sight (medium salience, “pocket/bag” condition), or in a separate room (low salience, “other room” condition).2 Our data indicate that the mere presence of one’s smartphone adversely affects two domain-general measures of cognitive capacity—available working memory capacity (WMC) and functional fluid intelligence (Gf)—even when participants are not using their phones and do not report thinking about them (experiment 1). Data from experiment 2 replicate this effect on available cognitive capacity, show no effect on a behavioral measure of sustained attention, and provide evidence that individual differences in consumers’ dependence on their devices moderate the effects of smartphone salience on available WMC.Experiment 1: Smartphone Salience Affects Available Cognitive CapacityIn experiment 1, we test the proposition that the mere presence of one’s own smartphone reduces available cognitive capacity, as reflected in performance on tests of WMC and Gf. Each of these domain-general cognitive constructs is constrained by the availability of attentional resources, and the moment-to-moment availability of these resources predicts performance on tests of both WMC (Engle, Cantor, and Carullo 1992; Ilkowska and Engle 2010) and Gf (Horn 1972; Mani et al. 2013). If the mere presence of one’s own smartphone taxes the limited-capacity attentional resources that constrain both WMC and Gf, then the salience of this device should predict performance on tasks associated with these constructs. We test this hypothesis in experiment 1.MethodParticipantsFive hundred forty-eight undergraduates (53.3% female; Mage = 21.1 years; SDage = 2.4 years) participated for course credit. Data collection spanned two weeks. Duplicate data from repeat participants were discarded prior to analysis. We applied the same three data selection criteria in experiments 1 and 2; see the appendix for additional detail. In experiment 1, three participants were excluded for indicating they did not own smartphones, eight participants were excluded for failing to follow instructions, and seventeen participants were excluded due to excessive error rates on the OSpan task (less than 85% accuracy; see Unsworth et al. 2005). Our final sample consisted of 520 smartphone users.ProcedureWe manipulated smartphone salience by randomly assigning participants to one of three phone location conditions: desk, pocket/bag, or other room. Participants in the “other room” condition left all of their belongings in the lobby before entering the testing room (as per typical lab protocol). Participants in the “desk” condition left most of their belongings in the lobby but took their phones into the testing room “for use in a later study;” once in the testing room, they were instructed to place their phones face down in a designated location on their desks. Participants in the “pocket/bag” condition carried all of their belongings into the testing room with them and kept their phones wherever they “naturally” would. Of the 174 participants in this condition, 91 (52.3%) reported keeping their phones in their pockets, and 83 (47.7%) reported keeping their phones in their bags; a planned contrast revealed no difference between these groups on our key dependent variable (p = .17), and they were pooled for all subsequent analyses. Participants in all conditions were instructed to “turn your phones completely on silent; this means turn off the ring and vibrate so that your phone won’t make any sounds.”After participants entered the testing room, they completed two tasks intended to measure available cognitive capacity: the Automated Operation Span task (OSpan; Unsworth et al. 2005) and a 10-item subset of Raven’s Standard Progressive Matrices (RSPM; Raven, Raven, and Court 1998). The OSpan task, a prominent measure of WMC, assesses the ability to keep track of task-relevant information while engaging in complex cognitive tasks. This particular measure was designed to stress the domain-general nature of the attentional resources at the heart of the WM system (Turner and Engle 1989); in each trial set, participants complete a series of math problems (information processing) while simultaneously updating and remembering a randomly generated letter sequence (information maintenance). Performance on the OSpan assesses the domain-general attentional resources “available to the individual on a moment-to-moment basis” (Engle et al. 1992). The RSPM test, a nonverbal measure of Gf, was developed to isolate individuals’ capacity for understanding and solving novel problems (fluid intelligence), independent of any influence of accumulated knowledge or domain-specific skill (crystallized intelligence). In each trial, participants are shown an incomplete pattern matrix and asked to select the element that best completes the pattern. Much like the OSpan task, performance on the RSPM test is sensitive to the current availability of attentional resources (e.g., Mani et al. 2013). Complete details of the tasks and measures used in experiments 1 and 2 are provided in the appendix.Participants also completed an exploratory test of the “ending-digit drop-off” effect, modeled after the procedure of Bizer and Schindler (2005). In this task, participants are shown a series of products with .99-ending and .00-ending prices and asked to report the quantity they would be able to purchase for $73. Overestimating purchasing power for a .99-ending price relative to a matched .00-ending price (e.g., $3.99 vs. $4.00) constitutes evidence of the drop-off effect. We thought this effect might be more pronounced for those whose phones were made salient. However, we failed to replicate the basic effect and did not find any evidence of ending-digit drop-off in any condition (F(1, 514) = .20, p = .65). See the appendix for detailed analyses and results.Next, participants completed a questionnaire that included items related to their experiences in the lab and their lay beliefs about the connection between smartphones and performance. These questions assessed how often they thought about their phones during the experiment, to what extent they thought the locations of their phones affected their performance in the lab, how they thought phone location might have affected their performance, and to what extent they believed their phones affected their performance and attention spans more generally; all responses were measured using 7-point Likert scales. Finally, participants answered a series of demographic questions (gender, age, ethnicity, nationality) and provided information about their cellphone make/model and data plan.Results and DiscussionAll analyses in experiment 1 include a “Week” factor to account for variation across research assistants; this factor does not interact with Phone Location in any analysis (all F < 1.27, all p > .28).Cognitive CapacityWe assessed the effects of smartphone salience on available cognitive capacity using two measures of domain-general cognitive function: OSpan task performance and RSPM test score. Because both tasks rely on limited-capacity attentional resources, both should be sensitive to fluctuations in the availability of these resources.A multivariate analysis of variance (MANOVA) testing the effects of Phone Location (desk, pocket/bag, other room) on the optimal linear combination of these measures revealed a significant effect of Phone Location on cognitive capacity (Pillai’s Trace = .027, F(4, 1028) = 3.51, p = .007, partial η2 = .014). Paired comparisons revealed that participants in the “other room” condition performed better than those in the “desk” condition (p = .002). Participants in the “pocket/bag” condition did not perform significantly differently from those in either the “desk” (p = .09) or “other room” (p = .11) conditions. However, planned contrasts revealed a significant desk → pocket/bag → other room linear trend (Pillai’s Trace = .023, F(2, 513) = 6.07, p = .002, partial η2 = .023) and no quadratic trend (Pillai’s Trace = .004, F(2, 513) = .96, p = .39), suggesting that as smartphone salience increases, available cognitive capacity decreases.Follow-up univariate ANOVAs separately testing the effect of Phone Location on OSpan performance and RSPM score were consistent with our focal multivariate analysis. Phone Location significantly affected both OSpan performance (F(2, 514) = 3.74, p = .02, partial η2 = .014) and RSPM score (F(2, 514) = 3.96, p = .02, partial η2 = .015). See figure 1 for means, and the appendix for detailed analyses and results.Figure 1. Experiment 1: effect of randomly assigned phone location condition on available WMC (OSpan Score, panel A) and functional Gf (Correctly Solved Raven’s Matrices, panel B). Participants in the “desk” condition (high salience) displayed the lowest available cognitive capacity; those in the “other room” condition (low salience) displayed the highest available cognitive capacity. Error bars represent standard errors of the means. Asterisks indicate significant differences between conditions, with *p < .05 and **p < .01.View Large ImageDownload PowerPointConscious ThoughtA one-way ANOVA on participants’ responses to the question “While completing today’s tasks, how often were you thinking about your cellphone?” (1 = not at all to 7 = constantly/the whole time) revealed no effect of Phone Location on phone-related thoughts (F(2, 514) = .84, p = .43). Notably, the modal self-reported frequency of thinking about one’s phone in each condition was “not at all.” Combined with the significant effect of Phone Location on available cognitive capacity, these results support our proposition that the mere presence of one’s smartphone may impair cognitive functioning even when it does not occupy the contents of consciousness.Perceived Influence of Smartphone PresenceThere were no differences between conditions on any measures related to the perceived effects of smartphones on performance (“How much / in what way do you think the position of your cellphone affected your performance on today’s tasks?”; “In general, how much do you think your cellphone usually affects your performance and attention span?”), either in the context of the experiment (all F < 1.58, all p > .21) or in general (F(2, 494) = 2.26, p = .11). Across conditions, a majority of participants indicated that the location of their phones during the experiment did not affect their performance (“not at all”; 75.9%) and “neither helped nor hurt [their] performance” (85.6%). This contrast between perceived influence and actual performance suggests that participants failed to anticipate or acknowledge the cognitive consequences associated with the mere presence of their phones.DiscussionThe results of experiment 1 indicate that the mere presence of participants’ own smartphones impaired their performance on tasks that are sensitive to the availability of limited-capacity attentional resources. In contrast to prior research, participants in our experiment did not interact with or receive notifications from their phones. In addition, self-reported frequency of thoughts about these devices did not differ across conditions. Taken together, these results suggest that the mere presence of one’s smartphone may reduce available cognitive capacity and impair cognitive functioning, even when consumers are successful at remaining focused on the task at hand.Experiment 2: Smartphone Dependence Moderates the Effect of Smartphone Salience on Cognitive CapacityThe results of experiment 1 support the proposition that the mere presence of one’s smartphone reduces available cognitive capacity, even when it is not in use. In experiment 2, we replicate the basic design of experiment 1, with the following exceptions. First, we conduct a stronger test of the proposed impairment-without-interruption effect by examining the effects of smartphone salience on both cognitive capacity (WMC) and a behavioral measure of sustained attention. Consistent with both the proposed theoretical framework and participants’ self-reports in experiment 1, we predict that increasing smartphone salience will adversely affect the availability of attentional resources without interrupting sustained attention. Second, one could argue that participants who had access to their phones in experiment 1 surr