Professional Documents
Culture Documents
net/publication/331348766
CITATIONS READS
0 381
2 authors:
All content following this page was uploaded by Robert J Klein on 28 February 2019.
Note: Correspondence can be sent to Robert J. Klein, Psychology, NDSU Dept. 2765, PO
Abstract
neuroticism may predispose people to a sort of “mental noise”, or cognitive instability, that
creates problems for ongoing efforts after control. If this is the case, cognitive views of
currently exist.
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1111/jopy.12469
This article is protected by copyright. All rights reserved.
Method: In a four study program of research (total N = 541), momentary forms of
monitoring and control were assessed using variants of a continuous tracking task.
Accepted Article
Results: As hypothesized, the dimension of neuroticism was consistently linked to
performance deficits, regardless of whether aversive sounds were present or not (Studies 1-3),
and the relevant deficits also predicted daily levels of negative affect (Study 4).
Conclusions: The results support the idea that momentary self-regulation is noisier in the
The personality dimension of neuroticism has both a long and short history. Greek
and Roman philosophers recognized anxiety as well as unhappiness and developed systems
called “anxietas” and linked it to mental disturbances, fear, and worry (Crocq, 2015). That
anxiety, depression, and unhappiness are linked was recognized subsequently, but
terminology varied and the literature was less consistent (Crocq, 2015). The modern view of
neuroticism began with Freud, continued with ideas about war trauma and the susceptibility
of certain soldiers to it, and reached an early culmination in the work of Hans Eysenck
(1947). Eysenck (1947) named the relevant features (e.g., distress, poor coping) in terms of
“neuroticism” and suggested that they comprise a dimension of personality rather than an
handful of personality dimensions that routinely emerges from factor analyses, whether of
Accepted Article
items or scales and whether the relevant attributes are reported on by selves or others
(Goldberg, 1993). The item content of neuroticism scales suggests moodiness, vulnerability,
and tendencies toward emotional upset, both within and outside of the context of stressors
(Watson, 2000, chap. 6). In turn, neuroticism has been shown to be a robust predictor of well-
being as well as psychopathology and people who are higher in neuroticism are prone to a
variety of serious issues such as self-harm and addictive behavior (Lahey, 2009). In daily
process terms, neuroticism predicts the occurrence of stressful events, negative emotional
reactions to them, and forms of coping that seem more problematic than helpful (Suls &
Martin, 2005).
Although we know a fair amount about the trait of neuroticism and what it predicts,
we know far less about the processes that underlie it or how it operates. This is a problem
because good, process-oriented theories are somewhat necessary when making new
predictions and because the absence of process-oriented theories could result in a somewhat
tautological area of inquiry (Ormel, Rosmalen, & Farmer, 2004). Along these lines, the most
systems responsible for identifying threats or punishments and responding to them (Watson,
Wiese, Vaidya, & Tellegen, 1999). Through mechanisms of this type, we could potentially
explain why neuroticism varies positively with emotional experiences like fear and why
reactivity to stressors is often more pronounced at higher levels of neuroticism (Suls &
Martin, 2005). Such theories might have a harder time accounting for other neuroticism-
seems to be emerging. This alternative can be expressed in terms of the “mental noise”
Accepted Article
hypothesis (Robinson & Tamir, 2005), which posits that neurotic people have noisier, more
chaotic mental control systems. In part because of a vigilant/avoidant mode of dealing with
the world (Mathews, 1990), individuals with high levels of neuroticism will often find
themselves off-task and disconnected from what they are currently supposed to be doing.
Vigilance with respect to the past can result in ruminative thoughts, which attempt to undo
what cannot be undone (Watkins, 2008). Vigilance with respect to the future can result in
worry, which somewhat similarly attempts to control something that cannot (at least
currently) be controlled (Borkovec, 2002). And even when neurotic individuals intend to
attend to the present, they can often find themselves caught up in intrusive thoughts (e.g., of
possible failure or the inadequacies of the self) that can interfere with focusing on the task at
The upshot of these considerations is the possibility that the mental control system
may “drift away” from present-moment reality more frequently at higher levels of
neuroticism (Perkins, Arnone, Smallwood, & Mobbs, 2015; Robinson, in press), potentially
affecting performance. There is evidence in favor of these ideas in the form of the
terms of attention and awareness concerning present reality (Brown & Ryan, 2003), seems to
be in shorter supply at higher levels of neuroticism. Indeed, the correlation between trait
measures of neuroticism and mindfulness can be as strong as r = -.5, despite minimal item
overlap (Hanley, 2016). Furthermore, there is evidence that mindfulness mediates some of
the pernicious consequences of neuroticism and does so in sensible ways. For example,
Fetterman, Robinson, Ode, and Gordon (2010) found that dispositional mindfulness mediated
relationship between neuroticism and self-control, at least when these outcomes were self-
Accepted Article
reported.
performance, as mental noise should tend to undermine it (Sanders, 1998, chap. 9). In fact,
there are fairly complicated relationships in this realm. Meta-analytic findings suggest that
neuroticism tends to be negatively related to job performance, but the magnitude of this
relationship is small (Judge, Klinger, Simon, & Yang, 2008). Moreover, there are cases in
which people with higher levels of neuroticism can outperform those with lower levels, at
least when the job requires a high degree of deliberation or concern (Perkins & Corr, 2005).
Findings of both types can be viewed in terms of attentional control theory (Eysenck,
Derakshan, Santos, & Calvo, 2007), which is an updated version of processing efficiency
theory (Eysenck & Calvo, 1992). According to Michael Eysenck and colleagues, people with
higher levels of trait anxiety (which is closely related to neuroticism) can compensate for
their lower levels of processing efficiency with effort, in effect canceling the limitations of
their noisier mental control systems. Indeed, there is evidence that anxious or neurotic people
typically want to do well and will devote extra effort in support of this possibility (Bresin,
useful to study tasks in which high levels of effort cannot compensate for deficiencies in
mental control. Robinson and Tamir (2005) used choice reaction time tasks in their studies,
with the reasoning that the choice-accuracy component of such tasks does not easily lend
itself to simple cognitive effort (Sanders, 1998, chap. 2). What Robinson and Tamir (2005)
found was that neurotic individuals displayed greater variability in their categorization times
from trial to trial than people low in neuroticism did. The authors interpreted this finding in
in semantic processing could also be involved (Sanders, 1998, chap. 5). Puzzlingly, though,
Accepted Article
there was no relationship between neuroticism and the speed with which the tasks could be
performed. Thus, these tasks were not able to capture neuroticism-related processes with the
precision necessary to demonstrate performance deficits. Or, put another way, the variability
that occurred at higher levels of neuroticism could have reflected stylistic variables rather
continuously track moving targets presented on the computer screen. Owing to the
continuous nature of the task, we could sample joystick or mouse position with a very high
degree of temporal precision, one that should be capable of revealing even short-duration
lapses of attention or control (Slifkin & Newell, 1998). To the extent that neurotic individuals
have noisier control systems (Robinson & Tamir, 2005), they should exhibit worse
performance on the task, defined in terms of greater discrepancies between target and cursor
position. Of note, the task is a very simple one in which one merely needs to attend to the
target and move toward it. Because the target is always moving, however, this process must
be repeated throughout the entire trial. Vagaries in attention and control should therefore
While investigating these basic questions about neuroticism and control, we sought to
contrast punishment-sensitivity ideas about neuroticism versus the mental noise hypothesis.
To the extent that punishment-sensitivity processes are involved, they should be more evident
under conditions of stress or threat (Perkins et al., 2015). To examine contingencies of this
type, we provoked participants with white noise blasts on some trials but not others.
By contrast, more endemic difficulties in mental control, which are emphasized by the mental
Accepted Article
noise perspective (Robinson & Tamir, 2005), as well as by Perkins et al. (2015), should be
evident in terms of main effects for neuroticism within both aversive and non-aversive
conditions.
We conducted four studies, both for the sake of replication and to explore the
substantive and programmatic. In Study 1, we instituted relatively long trials, with the
reasoning that long trials might be especially conducive to attentional lapses (Cheyne,
Carriere, & Smilek, 2006). Even so, it is our contention that mental noise should be evident
on shorter trials as well and we therefore switched to shorter trials in Studies 2-4. Doing so
also allowed us to blast individuals with shorter-duration noise blasts, which we reasoned
explored the consistency of the findings across different effectors (joystick versus mouse) and
Given the findings from the first three studies, we conducted a fourth in which we
thought it likely that neuroticism would manifest itself as poorer tracking performance.
Finally, because we think that mental noise follows from vigilance-related systems as well as
intrusive, unpleasant thoughts (Robinson & Tamir, 2011), we explored the possibility that
Wilkowski, 2015). Specifically, we explored the possibility that individuals displaying poorer
performance in the tracking task would be more prone to emotional distress in daily life.
Altogether, then, we performed multiple analyses and studies to examine the generalizability
of the findings.
Study 1 constituted an initial test of the idea that neuroticism disrupts the control
Accepted Article
functions necessary for continuous performance. The trials were relatively long (30 seconds),
which can contribute to lapses of attention (Steinborn, Langner, Flehmig, & Huestegge,
2016). In addition, we manipulated stress levels through the provision of loud white noise,
which has been linked to emotional and behavioral reactivity in previous research (Bushman
& Anderson, 1998; Westman & Walters, 1981). We hypothesized that neurotic individuals
would exhibit poorer control than non-neurotic individuals, potentially irrespective of trial
type.
Method
Power calculation software revealed that sample sizes of 113 would give us ample
(.90) power to detect correlations as well as interactions in the medium effect size (.30) range.
Because we sought this degree of power, we ran all of the studies for lengths of time that had
yielded similar sample sizes in previous research within the lab. All studies were approved by
who received psychology course credit. Demographics were not collected in this study, but
the participant pool tends to be 90% Caucasian, 60% female, and of traditional college age.
Participants arrived to the laboratory in groups of 6 or fewer and completed the remainder of
Neuroticism Assessment
Neuroticism was assessed with a broad-bandwidth scale from Goldberg (1999) that
converges highly with BFI and NEO versions of neuroticism (John & Srivastava, 1999) and
maximizes correlations with neuroticism factor markers (Goldberg, 1999). Participants were
disturbed”, “get stressed out easily”) describe the self (1 = very inaccurate; 5 = very accurate)
Accepted Article
and an average score was computed (M = 2.80; SD = 0.81; α = .89). For comparison
purposes, we also assessed the emotion-related (Watson, 2000, chap. 6) Big 5 traits of
.86).
Participants were asked to use a Saitek Aviator model joystick to track a visual target as
carefully and accurately as possible, using their dominant hands. The task was programmed
with E-Prime software, which used a 32-bit version of Windows XP, and the monitor, which
had a screen height of 13.65 inches, was set to a resolution of 1600 by 1200.
In the first study, we programmed 8 long, 30-second trials. Each trial began with a
visible joystick cursor (a white + sign), which was displayed at center screen. 500 ms later, a
target (also a white + sign) appeared, displaced to the left or right by 25%. Participants
moved the joystick cursor to the target (average latency = 2211 ms), which turned yellow
when the cursor was within 5 pixels both horizontally and vertically. At this point, the target
started moving further away from center screen, in a strait horizontal line, at the constant rate
of 300 pixels per second. Participants were instructed to track the target as well as possible
throughout the trial. Targets turned around when they reached the edge of the screen (which
position every 100 ms. To study the possible impact of emotional stimulation, we presented
80 dB white noise sounds, continuously, during the course of 4 of the trials, selected at
random. These sounds were both unpleasant and arousing (Bushman & Anderson, 1998).
For every sample of every trial, the Pythagorean theorem was used to calculate the
Accepted Article
distance between target position and joystick position. Because these distances were
positively skewed, we replaced distances longer than 200 pixels with this value, which
replaced 1.38% of the observations (Robinson, 2007). We then computed 2 means per
participant – one for silent trials and one for trials on which aversive noise had been present.
A repeated-measures ANOVA revealed that performance was somewhat better when noise
was present (M = 41.96 pixels) than not (M = 43.45), F(1, 149) = 6.27, p = .013, ηp2 = .04,
possibly because the constant noise helped participants maintain alertness during the long
trials.
a procedure that should be used when looking at potential interactions between continuous
personality variables and experimental manipulations (Robinson, 2007). The results revealed
main effects for Neuroticism, F(1, 148) = 6.73, p = .010, ηp2 = .04, and Noise Condition, F(1,
148) = 6.23, p = .010, ηp2 = .04, but there was no Neuroticism by Noise Condition interaction,
F < 1. Estimated means for the neuroticism main effect (Aiken & West, 1991) revealed that
performance was markedly worse when neuroticism was high (+1 SD: 45.60) relative to low
(-1 SD: 39.72). Also, additional analyses revealed that there were no comparable effects for
extraversion or agreeableness, all personality-related ps > .20. Thus, neuroticism was unique
performance means for the first two trials of each type (e.g., noise) versus the second two
trials of each type, and added this trial variable as an additional predictor. If the neuroticism-
related results are due to performance anxiety, they might be more apparent during earlier
trials. If they are due to fatigue, they might be more apparent during later trials. If, by
matter. There was a main effect for the Trial variable (Ms = 44.58 & 41.57 for earlier & later
Accepted Article
trials), F(1, 148) = 18.53, p < .001, ηp2 = .11, but interactions between neuroticism and this
trial variable were not significant, ps > .20. Accordingly, high-neuroticism individuals were
consistently worse on the task, regardless of trial order or aversive stimulation. The
consistency of these individual differences can be underlined in one final way: Cronbach’s α
across the 8 trials of the task was .96, indicating robust tendencies toward noisy performance
versus not.
Discussion
performance. Through the use of such procedures, we could catch micro-momentary lapses of
control, which would show up in terms of greater distance from a tracked target. At this level
of temporal resolution, we were able to show that neurotic people display difficulties
consistent with the mental noise hypothesis (Robinson & Tamir, 2005). The phenomenon did
not interact with aversive stimulation and it was neither more nor less apparent on earlier
trials, arguing against factors like evaluation apprehension (which would be more prevalent
on early trials) or fatigue (which would be more prevalent on later trials). The results
therefore highlight somewhat chronic difficulties in control that are apparent with a fine
temporal resolution.
Study 2
Despite the success of the first study, there were some ambiguities that led to an
altered Study 2 paradigm. In Study 1, participants used a joystick to track targets, but the
vagaries of this device (e.g., springs that sought to return to center screen), as well as
differing degrees of familiarity with joysticks, could have affected performance. Study 2
therefore switched to a regular computer mouse, which everyone should be familiar with. In
greater degree of mind-wandering than might otherwise have been the case (Steinborn et al.,
Accepted Article
2016). Such time-on-task factors are not emphasized by the mental noise hypothesis
(Robinson & Tamir, 2005) and we therefore switched to shorter, but more plentiful, trials in
Study 2.
aversive events would most likely disrupt rather than facilitate performance. We did this by
issuing white noise blasts for shorter periods of time, in a somewhat unpredictable manner.
This form of stimulation is both stressful and startling (Glass & Singer, 1972) and it should
question was whether this would be particularly true at higher levels of neuroticism, which
Method
The Study 2 sample consisted of 185 (75.14% female 89.75% Caucasian; M age =
18.92) undergraduates seeking course credit for their psychology classes. They completed the
motor tracking task prior to reporting on their neuroticism levels and did so within private
computer rooms in groups of 6 or fewer. Neuroticism was assessed with the same Goldberg
We designed a tracking task that would assess motor control performance before,
during, and after aversive noise blasts. The computers and programming platform were
identical to Study 1. However, participants were asked to track moving targets using a
familiar desktop mouse rather than the less familiar joystick used in the first study.
appeared, either 25% leftward or rightward from the cursor. Participants started the trial by
moving the mouse cursor to the target (average duration = 966 ms), which then turned yellow
and started moving toward the center of the screen at the rate of 160 pixels per second.
Participants tracked each target for 8000 ms and then returned to center screen. There was
The task included 60 trials. On 30 of them, the computer blasted the participant with
irritating 90 dB white noise for 200 ms. The aversive-ness of the event was further increased
by varying its timing such that it could occur at 1 of 10 different time intervals subsequent to
the start of the trial (from 1.5 seconds to 6.5 seconds), determined at random. For comparison
purposes, another 30 trials involved a silent “event” with the same timing parameters. The
with respect to three time intervals. We sampled mouse and target position every 100 ms for
5 samples prior to the trial event (either noise or silence), for 7 samples after event offset, and
for 6 samples thereafter. We will refer to these intervals in terms of baseline, reactivity, and
Results
mouse and target position. As in Study 1, these distances were positively skewed. We
therefore began by replacing 1.50% of the distances with this outlier value (230). The
distances remained positively skewed and we therefore log-transformed them for analysis
purposes, though pixel-based means were also computed for presentation purposes
a main effect for Sound, F(1, 184) = 44.64, p < .001, ηp2 = .19, such that performance was
worse on trials with aversive noise (M = 21.72), relative to silence (M = 20.74). There was
also a main effect for Trial Phase, F(2, 368) = 86.10, p < .001, ηp2 = .32, which displayed the
expected curvilinear pattern (baseline = 20.37; reactivity = 23.74; recovery = 21.29). These
main effects, though, were qualified by a Sound by Trial Phase interaction, F(2, 368) = 94.20,
p < .001, ηp2 = .33. As shown in the top panel of Figure 1, participants exhibited pronounced
reactivity to the aversive noise blasts, but later recovered from them to some extent.
predictor, which should be done with designs of the present type (Robinson, 2007). In this
presentation, we will omit normative effects, which were parallel to those discussed above. In
a replication of Study 1, there was a main effect for Neuroticism, F(1, 183) = 6.31, p = .013,
ηp2 = .03. Estimated means revealed that performance was worse when neuroticism was high
(+1 SD: 24.06) relative to low (-1 SD: 19.56). Neuroticism, though, did not interact with
Sound, F(1, 183) = 2.25, p = .135, ηp2 = .01, or Trial Phase, F(2, 366) = 1.96, p = .142, ηp2 =
.01, and there was no 3-way interaction, F < 1. Thus, the destabilizing effects of aversive
noise were comparable across neuroticism levels and the main effect for neuroticism
the first 15 trials of each type (noise or silence) versus the second 15 trials. The addition of
this trial variable can reveal whether neuroticism effects were more pronounced earlier on,
perhaps due to performance anxiety, or later, perhaps due to fatigue. In this four-predictor
GLM, there was a main effect for Trial, F(1, 183) = 16.81, p < .001, ηp2 = .08, such that
However, there were no interactions that involved both neuroticism and the trial factor, all ps
Accepted Article
> .20. These results again point to the generality of the neuroticism-related performance
costs. Of further note, individual differences were very reliable across the 60 trials of the task
(α = .97).
Discussion
The Study 2 procedures were different than the Study 1 procedures in several notable
respects. Participants were able to use a mouse rather than a joystick to track moving targets
and performance was better with this tracking device. In addition, the trials were fairly short,
which should minimize losses of attention that occur when sustained attention is required
(Sanders, 1998, chap. 9). Despite these changes, we were able to replicate the main effect for
neuroticism found in Study 1: Neurotic individuals, relative to those who tend to be more
stable and unperturbed (Watson & Clark, 1984), displayed noisier tracking performance even
when trials were short and the task was comparatively easy.
Study 2 also allowed us to contrast the mental noise hypothesis with threat-reactivity
views of neuroticism in a more effective manner. Aversive stimulation that was brief, intense,
and unpredictable (Westman & Walters, 1981) disrupted tracking performance in a relatively
pronounced way. Even so, this threat-reactivity effect was equal in magnitude at higher
versus lower levels of neuroticism, as reflected in the lack of a neuroticism by sound by trial
phase interaction. Or, expressed in other terms, the neuroticism effect was evident on all trials
rather than only some of them and was evident before as well as after the occurrence of
aversive stimulation. These findings are consistent with the mental noise hypothesis.
Study 3
The Study 2 procedures were relatively effective in contrasting mental noise versus
threat-reactivity perspectives of neuroticism. Because this was true, and because direct
Unless otherwise noted, the procedures for Study 3 were identical to Study 2. In Study
seeking (and receiving) course credit. Participants completed a continuous tracking task and
The tracking task was identical to Study 2 except that there were 40 rather than 60
trials and each lasted 5 seconds. On 20 of the trials, there was an aversive noise blast that
occurred 1, 2, 3, or 4 seconds after the start of the trial. There was no such blast on the
remaining 20 trials, though the computer program did log a silent “event” for comparison
purposes. Each participant received a different randomized order of the two conditions.
To study the influence of the noise blasts, we quantified performance with respect to
three time intervals. The baseline period consisted of 5 samples immediately preceding event
onset (either noise or silence). The reactivity period consisted of 7 samples after event offset.
And the recovery period consisted of the 6 samples following the reactivity period.
Results
Distances were positively skewed. We therefore replaced 0.97% of the high outliers
with this value (100). Because the distance distributions were still skewed, we log-
values, but means and estimated means will be reported in their original pixel-based units.
Study 2, there were main effects for Sound (silence M = 14.96; noise M = 17.64), F(1, 120) =
51.86, p < .001, ηp2 = .30, and Trial Phase (baseline M = 15.85; reactivity M = 17.29;
Trial Phase interaction, F(2, 240) = 32.17, p < .001, ηp2 = .21. As shown in the bottom panel
Accepted Article
of Figure 1, aversive noise blasts disrupted performance, which then recovered to a certain
extent.
2007). In a direct replication of Study 2, there was again a main effect for Neuroticism, F(1,
119) = 5.53, p = .023, ηp2 = .04: Continuous tracking performance suffered at higher (+1 SD:
17.73), relative to lower (-1 SD: 14.91), levels of this trait. Such neuroticism-linked
decrements in performance did not interact with Sound, F < 1, or Trial Phase, F(2, 238) =
1.81, p = .165, ηp2 = .02, nor was there a 3-way interaction, F < 1. Accordingly, individuals
high in neuroticism exhibited difficulties with the task that were orthogonal to the effects of
emotional reactivity.
first 10 trials of each type (either noise or silence) versus the second 10 trials of each type.
There was no main effect for Trial, F(1, 119) = 2.24, p = .137, ηp2 = .02, and 2- and 3-way
interactions involving neuroticism and trial were also non-significant, Fs < 1. There was,
however, a 4-way interaction, F(2, 238) = 3.08, p = .048, ηp2 = .03. An inspection of the
pattern suggested that high-neuroticism individuals exhibited delayed recovery during the
earlier trials of the task, but not the later trials. Regardless, performance was worse at higher
levels of neuroticism for all 12 of the cells defined by the trial by sound by trial phase within-
subject design (all estimated mean difference scores > 1.5). We therefore conclude that the
neuroticism-related main effect is a robust one. As a final note, the reliability of performance
across the 40 trials of the task was .96. This figure indicates that individual differences in
The main purpose of Study 3 was to replicate Study 2. Though linked to a modest
Accepted Article
effect size, we again found that higher levels of neuroticism were problematic to performance
within a continuous tracking task. The introduction of sudden and intense sound also
disrupted performance, but this emotional reactivity effect was equivalent at higher versus
lower levels of neuroticism, as defined by the lack of a neuroticism by sound by trial phase
interaction. Accordingly, the link between neuroticism and continuous performance appears
to be a general one rather than one that is dependent on aversive stimulation or particular
phases of a trial.
Study 4
Although neuroticism by sound interactions were not found in Studies 1-3, one could
still argue that the presence of aversive sounds on some of the trials placed neurotic
individuals into a defensive mode of operation that undermined their performance in task-
general terms. To speak to this issue, we simply removed any aversive stimulation from the
the circumstances, we should be able to replicate a main effect for neuroticism even in such a
context.
In Study 4, we also sought to extend the scope of the analysis. If we are correct,
mental noise can essentially serve as an implicit measure of personality (Robinson &
Wilkowski, 2015). What we mean here is that regardless of whether tracking performance
assessed by the tracking task – should have some capacity to predict the sorts of states that
one is susceptible to in daily life (Robinson & Wilkowski, 2015). To speak to predictions of
this type, we joined the assessment procedures of the laboratory with a subsequent daily diary
protocol (Bolger, Davis, & Rafaeli, 2003) and hypothesized that poorer performance in the
experiences of negative affect. We also examined whether this relationship would depend on
Accepted Article
the occurrence of daily stressors. Based on the findings of Studies 2 and 3, no such stress-
Method
sought to meet, and slightly exceed, recommendations that have been made in the multilevel
modeling (MLM) literature (e.g., Scherbaum & Ferreter, 2009). Accordingly, we collected
data from slightly over 100 people in the laboratory and the number of daily observations was
In more specific terms, potential participants were told that the study consisted of two
parts – an initial laboratory session followed by 2 weeks of daily diary reports. We ran the
laboratory portion of the protocol for long enough to collect tracking data from 103
undergraduates. We also collected personality data from 70 of these participants, and thus
potential links with personality were considered a secondary focus for the study. Of the 103
participants who completed the tracking task, 85 (47.96% female; 85.42% Caucasian; M age
= 19.47) completed at least 9 daily reports, which was an a priori criterion for inclusion (the
excluded participants fell considerably short of this standard). Within the laboratory session,
Because neuroticism main effects do not seem to depend on the presence of aversive
stimulation, we omitted such stimulation from the Study 4 paradigm. That is, participants
simply tracked moving targets, continuously, as well as they could. Each of the 24 trials
began with a mouse cursor (a white + sign) at center screen. The assessment portion of the
(average latency = 918 ms), which then turned yellow and began moving at the rate of 160
Accepted Article
pixels/second. During the 8-second trial, the target moved across the screen and participants
tracked it. Subsequently, participants were asked to return to center screen; after 500 ms, the
next trial started. Mouse and target position were sampled every 100 ms, which resulted in 80
Personality Assessment
Daily Protocol
Subsequent to the laboratory portion of the study, participants were asked to complete
14 days of a daily diary protocol. On each of these days, we sent participants an email
questionnaire. Participants were told to complete each report after 5 p.m. (so that the report
captured their daily experiences) and prior to 9 a.m. the next morning (so that retrospection
would be minimized). Reports not received within this time frame were considered missing.
Eighty-five participants completed at least 9 of the 14 reports (M = 12. 14) and were therefore
The key question was whether we could predict daily experiences of negative emotion
from tracking performance in the laboratory. To speak to this question, we asked participants
to report on their daily experiences of two negative affect markers (“distressed” & “nervous”)
from the PANAS (Watson, Clark, & Tellegen, 1988). Participants indicated how strongly
they experienced each feeling on each day (1 = not at all; 5 = extremely) and we averaged
3.09; SD = 1.08; α = .70). Finally, to examine whether possible relationships with negative
affect were stress-contingent, we asked about the occurrence (1 = not a single time; 4 = more
than two times) of four common stressors (e.g., “had a lot of responsibilities”, “too many
things to do at once”). The latter scale has demonstrated validity in daily diary studies (e.g.,
Klein, Liu, Diehl, & Robinson, 2017) and we averaged across items (M = 1.77; SD = 0.71; α
= .82).
Results
Distances from target, in the continuous tracking task, were computed using the
Pythagorean theorem. Because these distances were positively skewed, we replaced 1.19% of
the longer distances with this cutoff value (150). The distances still displayed some positive
GLM analyses were used to examine possible relationships between the Big 5
personality traits and continuous tracking means. In the most pertinent GLM, there was a
main effect for Neuroticism, F(1, 69) = 6.48, p = .013, ηp2 = .09, such that performance was
worse at high (+1 SD: 19.43) relative to low (-1 SD: 15.47) levels of this personality
continuum. By contrast, there were no personality-related main effects for Extraversion, F <
continuous tracking issues, t = 2.36, p = .022, β = .31, but the remainder of the Big 5 traits did
not share a relationship with the tracking measure, ps > .15. These data highlight the special
and performance.
Accepted Article
We next turned to the daily diary data, which were analyzed using the multilevel
(MLM) modeling procedures recommended by Singer (1998). In these models, the level 2
predictor (tracking performance) was z-scored, the level 1 predictor (daily stress) was person-
centered, and the daily outcomes (affect) retained their original units. In the model of central
interest, daily NA levels were higher among participants exhibiting poor tracking
performance, b = .127, t = 2.45, p = .017. By contrast, tracking performance did not predict
daily PA levels, b = .003, t = 0.05, p = .963. Thus, there was specificity to these relationships.
the relationship between tracking performance and daily NA levels. In this cross-level model,
NA levels increased when daily stressors occurred, b = .149, t = 4.85, p < .001. There was no
cross-level interaction, however, b = -.024, t = -0.83, p = .410, meaning that poor trackers
were prone to daily NA regardless of whether particular days were stressful or not. These
findings complement those of Studies 1-3, which similarly showed that neuroticism predicted
circumstances.
Discussion
tracking performance even in the absence of any stressors within the task. Thus, the
relationship between neuroticism and mental noise (as we have assessed it) appears to be a
general one. Moreover, this relationship was unique in that other members of the Big 5 were
highlight, though, is that people prone to poor tracking performance were also prone to
contingent and tracking performance did not predict positive affect levels. Lapses of
Accepted Article
momentary control therefore seem to share a special relationship with states of distress, a
General Discussion
deficient control. Perhaps not surprisingly, neurotic people report that they have greater
experiencing strong emotion (John & Gross, 2007). Such apparent difficulties are not limited
thoughts (Munoz, Sliwinski, Smyth, Almeida, & King, 2013) and in controlling their
consequence of such difficulties, neurotic people complain that they are prone to cognitive
failures (Flehmig, Steinborn, Langner, & Westhoff, 2007) and they report less mindfulness in
their daily activities (Fetterman et al., 2010). Although such correlates are self-reported in
nature, and one has to be dubious concerning self-reports of cognition, all of these findings
When one examines the personality-performance literature, though, one does not find
consistent effects for neuroticism. For example, neuroticism is not a good predictor of
performance (Perkins & Corr, 2005) in main effect terms. A possible reason for such null
findings is that people with high levels of neuroticism can compensate for their mental noise
(Robinson & Tamir, 2005) or inefficiency (Eysenck & Calvo, 1992) with higher levels of
effort, in effect eliminating possible performance deficits when the task tolerates brief
the emotional features of anxiety can motivate the person to try harder even while the
Accepted Article
cognitive features of anxiety interfere with on-task monitoring (Bresin et al., 2011; Sarason,
1984).
To test these ideas would seem to require a task in which momentary lapses of
attention or control cannot be rectified through spurts of on-task activity. The continuous
tracking task seemed ideally suited to solve such problems. The target for the task moved
and, therefore, participants had to continuously monitor the target as well as their own
position in order to do well. We sampled position 10 times a second and therefore any brief
distances. And, because the task is such a simple one, performance in it should not necessary
performance and these relationships were quite robust. They were equally present on early
trials and late trials, which could shift one’s task state from anxiety to fatigue (Hockey,
1997). They were also equally present on long trials and short trials and within both aversive
and non-aversive sound conditions (Glass & Singer, 1972). Indeed, a close analysis of the
latter time course revealed that neuroticism-linked deficits were apparent before, during, and
after aversive noise bursts, as reflected in a main effect for neuroticism that did not depend on
the sound by trial phase interaction that defined a reactivity effect (see Figure 1). It was also
the case that the neuroticism-performance relationship was observed whether participants
used a joystick controller or a mouse and the relationship was a unique one in that there were
no parallel effects for other Big 5 traits (Studies 1 & 4). Thus, a task with excellent temporal
resolution appears to possess considerable value in documenting the sorts of lapses and slips
Robinson, in press).
Accepted Article
The results were not limited to such relationships, however. In Study 4, we were also
able to show that variations in mental and behavioral control possess value in predicting the
sorts of emotions or feelings that an individual is likely to experience in the upcoming weeks.
Specifically, poor controllers were vulnerable to states of negative affect in their daily lives,
irrespective of their levels of trait neuroticism. Thus, and even though the continuous tracking
task would seem to be a relatively neutral one, with few direct implications for emotional
experience, individuals whose attention exhibited greater variability during the task were
more vulnerable to distress as they went about their lives. This was true regardless of whether
particular days were stressful or not and the findings therefore offer positive support for the
idea that even brief lapses of attention or control can reflect the sorts of processes that give
disconnections between the mind and its control of the external environment. These
disconnections will be due to attentional lapses and to variations in attention across time as
attention plays a key, if not the key, role in the regulation of motor behavior (Lohse, Jones,
Healy, & Sherwood, 2014). At the same time, the task should not be viewed as a vigilance
task because these sorts of tasks typically require individuals to maintain alertness across
fairly long periods of time in order to respond to infrequent target events (Sanders, 1998,
chap. 9). By contrast, the continuous tracking task requires constant behavioral adjustments
that operate over much shorter periods of time (e.g., less than 10 seconds in Studies 2 and 3).
an easy task – one merely needs to follow a clearly visible target as it moves across the
Accepted Article
screen. While doing so, the person is not asked to inhibit dominant responses, to switch task
sets, or to move information in and out of working memory. Accordingly, the task is
structurally distinct from executive function tasks (Miyake & Friedman, 2012), though the
individual will still need to control internal forms of distraction (e.g., mind-wandering or
rumination) in the service of good performance. From this perspective, the continuous
tracking task could be considered a sort of mindfulness task in that doing well on it mostly
requires intention, attention, and awareness concerning a simple activity (Jazaieri & Shapiro,
2017).
Having discussed the nature of the task, we should also briefly discuss the sound-
related results of Studies 2 and 3. Prominent theories of emotion suggest that emotional
reactions can co-opt the systems responsible for behavior (Frijda, 2010). Consistent with this
idea, we found that sudden, intense, and aversive noise blasts disrupted performance on the
continuous tracking task and this disruption effect lasted for about a second, which is
compatible with the time course of fear or startle (Lang, 1995). This reactivity effect was
equivalent across levels of neuroticism, but could vary by other factors that implicate the fear
system in more particular terms. For example, there is some evidence that startle reactions are
less pronounced in psychopathy (Lang, 1995) and probes of the Study 2 sort could be
The primary purpose of the studies, though, was to support noise- or efficiency-based
views of neuroticism. Such views contend that the neurotic mind is a noisier, more chaotic
one that is prone to slips of attention and control (Flehmig et al., 2007). Neurotic people may
be able to compensate for such issues through extra effort (Eysenck et al., 2007) when they
scales. The current findings, we think, offer some of the best available evidence for such
Accepted Article
ideas.
Although neuroticism did not modulate reactivity to aversive events in Studies 2 and 3, those
disruptive to nearly everyone (Westman & Walters, 1981). Had the aversive events
implicated the ego to a greater extent, interactions with neuroticism might have been found.
Consistent with this idea, the mind-wandering linked to neuroticism tends to be self-relevant
in nature (Perkins et al., 2015) and it often takes the form of rumination, worry, or self-
criticism (Suls & Martin, 2005). It is possible to manipulate processes of this type (e.g.,
Watkins & Teasdale, 2001) and such conditions could amplify the neuroticism-related
after continuous tracking performance. Building on earlier research (e.g., Sarason, 1984),
Matthews et al. (2002) compiled a large set of items to measure the motivational, cognitive,
and emotional states that can occur when individuals are trying to accomplish a task. A factor
analysis of these measures revealed that they can be parsimoniously described in terms of
task engagement, distress, and worry. There are reasons for thinking that worry should be
be problematic as well (Matthews et al., 2002). It is further possible that states of this type
might mediate the association between neuroticism and continuous tracking performance,
though our analyses did suggest some degree of state-independence (e.g., similar
performance and operations of the default mode network (DMN). The DMN, which includes
Accepted Article
brain regions such as the lateral parietal cortex, the medial prefrontal cortex, and the anterior
cingulate cortex, has been implicated in mind-wandering (Schooler et al., 2011) as well as
symptoms of anxiety and depression (Coutinho et al., 2016). Our behavioral task might have
value in tracking DMN activity, though one would have to establish systematic links of this
Conclusions
Robinson and Tamir (2005) proposed that neuroticism varies positively with mental
noise, which should affect performance under some circumstances. Although Robinson and
Tamir (2005) were able to show that neurotic people exhibited greater reaction time
variability, the temporal resolution of the tasks was insufficient for demonstrating
performance deficits. The present studies sought to rectify such problems through the use of a
continuous tracking task – one in which even short-duration lapses in control would be
evident. Across four studies, positive relationships between neuroticism and performance
deficits were observed, and these findings were not dependent on punishment-sensitivity
processes. The findings, more so than previous studies, link neuroticism to basic deficits in
momentary self-regulation.
The author(s) declared no potential conflicts of interest with respect to the research,
Accepted Article
authorship, and/or publication of this article.
Funding
No funding to report
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting
Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life as it is lived.
Borkovec, T. D. (2002). Life in the future versus life in the present. Clinical Psychology:
Bresin, K., Robinson, M. D., Ode, S., & Leth-Steensen, C. (2011). Driven, distracted, or
Brown, K. W., & Ryan, R. M. (2003). The benefits of being present: Mindfulness and its role
848.
Chamorro-Premuzic, T., Moutafi, J., & Furnham, A. (2005). The relationship between
15, 578-592.
Coutinho, J. F., Fernandesl, S. V., Soares, J. M., Maia, L., Goncalves, O. F., & Sampaio, A.
(2016). Default mode network dissociation in depressive and anxiety states. Brain
Crocq, M.-A. (2015). A history of anxiety: From Hippocrates to DSM. Dialogues in Clinical
Derryberry, D., & Rothbart, M. K. (1988). Arousal, affect, and attention as components of
Eysenck, M. W., & Calvo, M. G. (1992). Anxiety and performance: The processing
Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive
Fetterman, A. K., Robinson, M. D., Ode, S., & Gordon, K. H. (2010). Neuroticism as a risk
Frijda, N. H. (2010). Impulsive action and motivation. Biological Psychology, 84, 570-579.
Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe (Vol. 7, pp. 7-28).
mindfulness and the Five Factor Model of personality. Personality and Individual
Jazaieri, H., & Shapiro, S. (2017). Mindfulness and well-being. In: M. D. Robinson & M. Eid
(Eds.), The happy mind: Cognitive contributions to well-being (pp. 41-58). Cham,
John, O. P., & Gross, J. J. (2007). Individual differences in emotion regulation. In: J. J. Gross
(Ed.), Handbook of emotion regulation (pp. 351-372). New York, NY: Guilford Press.
Judge, T. A., Klinger, R., Simon, L. S., & Yang, I. W. F. (2008). The contributions of
Klein, R. J., Liu, T., Diehl, D., & Robinson, M. D. (2017). The personality-related
241-256.
Lang, P. J. (1995). The emotion probe: Studies of motivation and attention. American
Lohse, K. R., Jones, M., Healy, A. F., & Sherwood, D. E. (2014). The role of attention in
Mathews, A. (1990). Why worry? The cognitive functions of anxiety. Behaviour Research
Matthews, G., Campbell, S. E., Falconer, S., Joyner, L. A., Huggins, J., Gilliland, K., et al.
Miyake, A., & Friedman, N. P. (2012). The nature and organization of individual differences
Ormel, J., Rosmalen, J., & Farmer, A. (2004). Neuroticism: A non-informative marker of
39, 906-912.
Perkins, A. M., Arnone, D., Smallwood, J., & Mobbs, D. (2015). Thinking too much: Self-
492-498.
Perkins, A. M., & Corr, P. J. (2005). Can worriers be winners? The association between
worrying and job performance. Personality and Individual Differences, 38, 25-31.
research methods in personality psychology (pp. 345-359). New York, NY: Guilford
Press.
Robins (Eds.), Handbook of personality: Theory and research (3rd ed.). New York:
Guilford Press.
Robinson, M. D., & Tamir, M. (2005). Neuroticism as mental noise: A relation between
neuroticism and reaction time standard deviations. Journal of Personality and Social
Robinson, M. D., & Tamir, M. (2011). A task-focused mind is a happy and productive mind:
Designing positive psychology: Taking stock and moving forward (pp. 160-174). New
Sarason, I. G. (1984). Stress, anxiety, and cognitive interference: Reactions to tests. Journal
Scherbaum, C. A., & Ferreter, J. M. (2009). Estimating statistical power and required sample
Schooler, J. W., Smallwood, J., Christoff, K., Handy, T. C., Reichle, E. D., & Sayette, M. A.
Singer, J. D. (1998). Using SAS PROC MIXED to fit multilevel models, hierarchical models,
and individual growth models. Journal of Educational and Behavioral Statistics, 23,
323-355.
Steinborn, M. B., Langner, R., Flehmig, H. C., & Huestegge, L. (2016). Everyday life
time distributions and delta plots. Applied Cognitive Psychology, 30, 92-102.
Suls, J., & Martin, R. (2005). The daily life of the garden-variety neurotic: Reactivity,
73, 1485-1510.
353-357.
Watson, D. (2000). Mood and temperament. New York, NY: Guilford Press.
Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief
measures of positive and negative affect: The PANAS scales. Journal of Personality
Watson, D., Wiese, D., Vaidya, J., & Tellegen, A. (1999). The two general activation systems
Westman, J. C., & Walters, J. R. (1981). Noise and stress: A comprehensive approach.
30
Distance from Target (Pixels)
25
20
Silence
Noise
15
10
Baseline Reactivity Recovery
Trial Phase
22
Distance from Target (Pixels)
19
16
Silence
Noise
13
10
Baseline Reactivity Recovery
Trial Phase