Trait self-control does not predict attentional control: Evidence from a novel attention capture paradigm.

To what extent are low-level visual and attentional phenomena related to higher-level personality traits? Trait self-control is thought to modulate behavior via two separate mechanisms: 1) by preventing initial temptation and, 2) by inhibiting temptation when it occurs (disengagement). Similarly, the control of visual attention often entails preventing initial distraction by irrelevant but tempting (goal-similar) objects, and disengaging attention when it has been inappropriately captured. Given these similarities, we examined whether individuals higher versus lower in trait self-control would differ in their susceptibility to attention capture using mouse-tracking as a sensitive, online measure of how attentional dynamics resolve over time and space in response to a distracting visual cue. Using a variety of metrics of attention capture, we found that differences among people in trait self-control did not predict initial selection of visual information nor subsequent disengagement. Overall, these results suggest that trait self-control and attention capture operate via separate mechanisms.

Introduction

Does the ability to successfully ignore colorful but task-irrelevant billboards while driving have anything in common with the ability to successfully refuse a delicious but unhealthy slice of cake? Both tasks involve inhibitory control—one to suppress distraction, the other to suppress temptation. Are these mechanisms related to one another or are they orthogonal? One possibility is that the cognitive mechanisms involved in demonstrating the self-control to say “no” to an unnecessary slice of cake when one holds the goal of eating healthfully are similar to the mechanisms involved in resisting visual distraction. Another possibility is that the macro-level processes involved in self-control are orthogonal to micro-level processes involved in attention capture. Here, we explore whether differences in trait self-control influence attention capture.

What is attention capture?

Attention capture is the involuntary and transient prioritization of task-irrelevant stimuli. In other words, it is when we temporarily lose control of our attention and attend to something irrelevant to the task at hand. One of the most popular methods for examining this phenomenon is the contingent capture paradigm [1-2]. In a sense, this paradigm can be thought to measure distraction by irrelevant but tempting (goal-similar) objects. Participants are instructed to search for and identify a specific target (e.g., a red symbol among symbols of other colors), and response times are recorded. Prior to searching for the target, the display screen is cued with an irrelevant distractor that either shares the defining feature of the target (e.g., a red flash of color) or does not match the target (e.g., a green flash of color or the abrupt appearance of a new object). The typical finding is that only distractors that match what people are looking for capture attention (slowing responses when they appear at non-target locations), whereas ones that do not match have little ability to do so. For example, an observer searching for a red X may involuntarily direct attention toward an irrelevant flash of red, but not an irrelevant flash of green, even though both represent salient environmental change. Some researchers suggest that this pattern of attention capture is primarily related to the mechanism of attentional disengagement [3-4]. They argue that all salient stimuli are capable of capturing attention, but disengagement from that distractor is delayed to the extent that the distractor shares features with the search target, producing longer response times.

Individual differences in capture

Notably, susceptibility to attention capture differs across people: some are more or less prone to having their attention captured. People high in working memory capacity exhibit less capture by irrelevant information [5-6]. Similarly, higher working memory capacity is associated with faster recovery from capture [7]. In addition to the working memory literature, some work suggests that action video game experience reduces the involuntary capture of attention [8]. Finally, there is other work suggesting that mood may influence attention capture; specifically, it has been shown that depression can increase or decrease capture effects depending on the nature of the distractor [9]. Collectively, these studies suggest that basic mechanisms of attention may be susceptible to individual difference factors—with some factors increasing susceptibility to capture and others decreasing it. Along these same lines, we ask if trait self-control also influences susceptibility to attention capture.

Self-control

Self-control can be conceptualized as the ability to (a) override or inhibit prepotent responses, including thoughts, emotions, and behaviors, and (b) replace them with responses more consistent with social norms or one’s long-term goals [10-11]. Self-control varies according to situational demands or constraints [12] and also among individuals [13]. This stable trait is the focus of the current work. High trait self-control facilitates waiting for larger, later rewards over smaller, sooner rewards [14] and is associated with various positive life outcomes, including academic and career success, quality of interpersonal relationships, psychological wellbeing, avoidance of substance abuse and crime, and relatively better health [13, 15, 16].

How does self-control work and what mechanisms does it recruit? Self-control is thought to work by employing executive functions toward maintaining goal-pursuit [17-18]. Researchers have argued that self-control depends on behavioral inhibition, task-switching, and working memory [17-18]. It is also associated with deliberative cognitive processing, such as planning, decision making, and impulse control [15]; consideration of future consequences [13]; and need for cognition [19].

In the past, researchers assumed that people high in trait self-control differed from those low in trait self-control merely in the extent to which they could overcome temptation to violate a norm or undermine goals [20]. Augmenting that view, a burgeoning body of literature suggests that people high in trait self-control may not experience temptation or distracting desires to the same extent that people low in this trait do [17] and try to avoid or preclude distractions [21] in addition to resisting temptations more effectively [17]. Researchers have used experience sampling methods to map out everyday self-regulatory processes and found that people higher (vs. lower) in trait self-control did not experience as many urges or desires that conflicted with more important goals [22]. Of the temptations that broke through, people higher in trait self-control felt less tension and more often succeeded in overcoming them.

Moreover, people high (vs. low) in trait self-control actively avoid distractions in their environment that might undermine performance on an important task and when speed or accuracy garner larger rewards. For example, in laboratory settings, people high but not low in self-control chose in advance to forgo breaks in an economics experiment to read entertaining stories [23], to wait for a distraction-free room to become available to optimize performance on an anagram task rather than work in a noisy room that was available immediately [21], and to view a boring standard black and white version of an anagram task instead of a more aesthetically pleasing but distracting one (i.e., included pictures of classic and modern artwork on either side; [21]). Other work suggests that people higher (vs. lower) in trait self-control avoid goal conflicts by automatizing particular behaviors (i.e., forming habits), including in health and academic domains [12, 24, 25].

In addition to these macro-level behavioral differences, self-control is also associated with low-level cognitive processes, such as working memory [17-18]. Given that differences in working memory are associated with differences in attention capture [5, 7], it is reasonable to suspect that differences in trait self-control may be associated with differences in attention capture. Specifically, we might expect that people higher in trait self-control are less susceptible to attention capture and are better at filtering out goal-irrelevant stimuli. This would fit with the idea that effective self-control entails preclusion of distraction. On the other hand, it may be the case that people high (vs. low) in trait self-control cannot help but initially attend to distracting stimuli in the perceptual field, but they can better disengage from those stimuli. This would be consistent with the traditional view that self-control entails overcoming that which does not facilitate one’s goals [20]. Given that people with high self-control experience fewer distractions, preclude more distractions, and more easily overcome distractions relative to those low in trait self-control [17], it may even be the case that trait self-control is positively associated with both the initial selection of visual information and the subsequent disengagement from irrelevant visual information.

Current study

Research question

Self-control involves mechanisms of suppression and inhibition. Attention capture also involves analogous mechanisms of suppression and disengagement. To what extent are these mechanisms that operate on very different levels—self-control on a macro-level and attention capture on a micro-level—related to one another? On the one hand, we might expect that differences in self-control could influence basic visual attention processes. If so, the question then becomes: which aspects of attention capture does it affect—selection, disengagement, or both? On the other hand, it has also been argued that attention capture has strong bottom-up components [3]. The strong version of this approach is that bottom-up processes are impenetrable and unaffected by top-down processes. In this case, we would expect differences in trait self-control to be unrelated to differences in attention capture, but may be more related to disengagement.

Mouse-tracking

In order to test whether self-control is associated with attention capture—and if so, which mechanisms it affects—we employed a computer mouse-tracking paradigm [26-29]. Participants selected a colored X that appeared on the screen while the x and y the coordinates of their responses were continuously recorded. The underlying assumption of mouse-tracking is that the partial and tentative conclusions of perceptuo-cognitive systems continuously feed into and affect the motor system [28, 30, 31]. Thus, the parallel influence of various cognitive processes can be seen in the unfolding trajectory across space and time. The advantage of this methodology is that it provides a high-resolution measure of online behavior, one capable of disentangling mechanisms of selection and disengagement from one another. Thus, the assumption is that if a participant’s attention is captured, it will be reflected in the trajectory of their mouse movement [32-35].

Our decision to use mouse-tracking instead of a similar methodology like eye-tracking was based on a number of advantages related to measuring how cognitive processes play out over time. While saccades are largely ballistic in nature and their flight paths are often completed in tens of milliseconds, mouse-tracking unfolds over a much larger timescale, providing a rich look at both spatial and temporal dimensions of behavior. On another level, mouse-tracking was chosen over eye-tracking for a number of practical reasons. Mouse-tracking software is freely available [27, 29] to anyone who has access to a computer. This high-accessibility makes it an appealing method of conducting research, one that is particularly conducive to encouraging replication and reproducibility. Finally, mouse-tracking is a relatively new methodology compared to eye-tracking, especially within the field of attention capture and self-control; thus, given the novelty of the methodology, we wanted to deepen the field’s knowledge.

Predictions

Broadly speaking, there are four different ways trait self-control and attention capture can interrelate. One, self-control may only correlate with selectivity—the degree to which features similar to our goals (i.e., matching cues) capture attention over and above features that are dissimilar with our goals (i.e., mismatching cues). People high in self-control may be more resistant to the involuntary capture of their attention by visually salient but goal-similar features. Two, self-control may only correlate with disengagement—the ability to release attention after it has been captured. People high in self-control may be just as susceptible to being captured by salient stimuli but may be better at disengaging from it. Three, self-control may correlate with both selectivity and disengagement. People high in self-control may be less susceptible to the involuntary capture of attention by a goal-similar distractor and better at disengaging from it when they are captured. Finally, self-control may be uncorrelated to either selectivity or disengagement. There may be no difference between people with high and low self-control in how their attention is captured.

To preview our results, our data are most consistent with the fourth possibility. Namely, self-control scores did not correlate with either selectivity or disengagement. This suggests that attention capture and trait self-control operate via different mechanisms.

Methods

All procedures complied with and were approved by Florida State University’s IRB; participants were compensated with course credit. Data were collected from 103 participants; however, three participants were excluded from analysis because they did not complete the self-control survey (final N = 100, 71 females, Mage = 20.03). This sample size was determined by an a prior power analysis. Using an alpha level of .05 and a power level of .8, our sample of 100 participants is sufficiently powered to detect a medium correlation (r = .28; see [36]).

Note, the procedures, apparatus, and preprocessing in this paper were identical to previous research [35]. Our hypotheses, predictions, and analyses were preregistered (https://osf.io/twhnj/) and our raw data, scripts, and Supplemental Materials are publicly available (https://osf.io/xnc7k/).

Apparatus and stimuli

The experiment was programmed, preprocessed, and analyzed using the following software: Programming was done in OpenSesame version 3.1.9 [37] using the mousetrap package [29] and the legacy backend; Data importing and preprocessing were done in R [38] using the readbulk [39] and mousetrap [40] packages; Data were analyzed using a combination of R packages [41] and JASP 0.8.2.0 [42]. Data were collected on a computer running Windows 7 with default mouse settings. In OpenSesame, the script was set to record coordinates at a temporal resolution of 10 ms. Participants sat approximately 30 cm from an 18-inch Dell Trinitron CRT monitor. The resolution of the screen was set to 1280 X 800 by OpenSesame.

Participants saw a screen with a start button and four response boxes. Stimuli consisted of colored Xs (red, blue, green, and yellow) that appeared inside one of the four response boxes. The start button was gray with the word “Start” in black text and was located in the bottom center of the screen. We refer to the other four boxes as B1, B2, B3, and B4 from left to right (see Fig 1). The boxes on the outside (B1 and B4) appeared halfway up the center of the screen. The boxes in the center (B2, B3) appeared at the top. All four of these boxes started off as white outlines with black centers. Each box had the following dimensions: width = 160 pixels (1/8 of the screen’s width), height = 100 pixels (1/8 of the screen’s height). The boxes were 6° along the diagonal in visual angle.

Fig 1

Experiment 1 procedure.

Half of the participants searched for a green X and half searched for a red X. Participants clicked the start button to initiate the trial. Thereafter, there was a 500 ms delay. Then one of the four boxes was cued with a flash of color for 75 ms that either matched or mismatched the color of the target. The cue could appear in any of the four boxes (labeled B1-B4 from left to right). Finally, there was a 50 ms delay where the boxes were shown, followed by a screen with colored Xs inside the boxes. Participants clicked on the box containing the colored X they were searching for. Capture was evidenced by mouse trajectories that curved toward the cue when it appeared on the opposite side of the screen compared to the target.

Procedure

The trial structure is shown in Fig 1. Participants were told they would see colored Xs and that they would have to click on their target. Half the participants searched for a green X and half for a red X. Participants started the trial by clicking on the start button. After clicking, there was a 500 ms delay where the start box was removed (but the other four boxes remained). This was followed by a flash of color. The outline of one of the boxes changed color for 75 ms. This flash of color was the cue and it matched the target on 50% of trials (i.e., green flash when searching for a green X) and mismatched on 50% (i.e., red flash when searching for a green X). This screen was followed by a screen showing the four boxes with the white outlines again for 50 ms. At this point, the mouse was reset to the bottom center of the start screen to ensure that all trajectories started from the same origin point. Finally, four colored Xs appeared inside the boxes on the last screen. Participants moved their mouse to the box containing their target and clicked on it. Note, participants did not have to click the X itself; they could click anywhere inside the boxes containing the correct X. Participants had 2000 ms to make their response. The locations of the colored Xs were randomized and counterbalanced such that the target appeared at each location an equal number of times. Trials were marked correct if participants clicked on the box containing the appropriately colored X (either red or green) within the response deadline (2000 ms). Participants were explicitly warned to ignore the flash as this was intended to distract them and compromise their performance. Participants completed 32 practice trials followed by five blocks of 96 trials each. The location of the target and the cue were randomized and counterbalanced within each block, meaning that the cue occurred on the target location about 25% of the time. A blank screen with a fixation dot in the center appeared for 1000 ms between trials. As a reminder, the boxes were labeled B1 to B4 from left to right. Trials where the target appeared at B1 or B4 were dropped from analyses as noted in the Data Exclusion section of our preregistration (consistent with previous research [35]). If the target was B2 and the cue appeared at B1 or B2, this was counted as a “same side” trial. However, if the target was B2 and the cue appeared at B3 or B4, this was counted as a “different side” trial. The converse is true for when the target appeared at B3.

After this first task, participants then completed the Self-Control Scale [13]. This scale has been found to be negatively associated with delayed discounting using preference based tasks [43], point based tasks [44], and, most importantly, using monetary based tasks (in the form of gift cards [45]). The scale consists of 36 statements (24 reverse scored), such as I am good at resisting temptation and I am self-indulgent at times (reverse scored). Participants rated the extent to which they agreed with each item on a scale of 1 (not at all like me) to 5 (very much like me). A single trait self-control score was created for each participant by calculating the reverse coded items and averaging across the 36 items (M = 3.35, SD = .52, α = .90).

Data preprocessing

Before conducting our analyses, we first excluded all practice trials, incorrect trials, and trials where the participant did not respond within the response deadline (2000 ms). In addition, we filtered out trials where the response time was faster than 300 ms or slower than 3.5 standard deviations above the participant’s mean as determined a priori (see the Data Exclusion section of the preregistration). Finally, only trials appearing at the top central boxes (B2 and B3) were analyzed. All data filtering and preprocessing were specified in the preregistration.

For ease of visualization and comparison, trajectories were remapped to the top left response so that all responses began in the same spot and ended at the same response box. Afterwards, trajectories were normalized to 101 steps [27, 29, 30] to ensure that responses had an equivalent number of coordinates regardless of differences in response time. This was done for visualization purposes. The dependent measures for each trajectory were calculated based on the raw trials (un-normalized). These dependent measures were aggregated up to the participant level for analyses.

Results

Manipulation check

In addition to our primary analyses reported below, we ran secondary analyses to investigate attention capture effects in general. These analyses can be seen as manipulation checks. Given that they are secondary to the primary purpose of the paper, we report these in the Supplementary Materials (https://osf.io/xnc7k/). Here, we merely note that these analyses are consistent with previous research of ours [35] and verify that our manipulation worked. Namely, capture occurred for both cues matching and mismatching the color of the target, as indicated by mouse trajectories that initially curved toward distractors when they occurred on the opposite side of the screen compared to the same side of the screen as the target, but capture was substantially smaller for cues not sharing the color of the target since trajectories corrected toward the target soon after this initial curvature. This demonstrates that both matching and mismatching cues captured attention but mismatching cues were easier to disengage from. Having established that captured did occur, we now turn our attention to the trait self-control analyses.

Primary analyses

Our primary analyses used two mouse parameters to derive dependent variables: area under the curve and time of maximum deviation. Area under the curve (AUC) is calculated as the difference in area (measured in pixels) between a direct trajectory from the start position of the mouse to the target and the actual trajectory of the mouse movement; this measures the degree of spatial attraction to the competing, but ultimately unchosen, responses [46]. In contrast, time of maximum deviation (TMAD) is a temporal measure calculated as the latency in time between the beginning of the trial and the point of greatest divergence from a direct line to the target; this measure provides a temporal index of the resolution in attraction [47].

In order to test whether trait self-control affects attention capture, we created different sets of capture scores: one to primarily index attentional selectivity and one to primarily index disengagement (see Fig 2). We created a “selectivity” index by subtracting mouse parameters for trials featuring mismatching (non-target color) cues on the target-opposite side of the screen from trials featuring matching (target color) cues on the opposite side (i.e., Matching Cue Opposite Side–Mismatching Cue Opposite Side). This measure compares a person’s distraction to goal-similar cues (e.g., red cue, red target) against their distraction to goal-dissimilar cues that mismatch the target (e.g., green cue, red target). For area under the curve, higher selectivity scores reflect greater distraction by goal-similar features (relative to goal-dissimilar features). For time of maximum deviation, higher scores reflect a longer period of distraction specifically for goal-similar relative to goal-dissimilar information.

Fig 2

Selectivity and disengagement.

Graphical representation of our selectivity and disengagement indices based on Area Under the Curve (AUC). (A) Selectivity is the difference in area under the curve between trials on which the distracting cue matched vs. mismatched the color of the target when the cue appeared on the opposite side of the screen as the target. (B) Disengagement is the area under the curve when the cue matches the color of the cue and the cue appears on the opposite side of the screen as the target. It corresponds to the difficulty in disengaging (i.e., inhibiting) the goal-similar cue.

In addition, our disengagement index was merely mouse parameters for trials on which the cue matched the color of the target and appeared on the side of the screen opposite the target. The greater the area under the curve for matching cues (e.g., red cue, red target) the more difficult it was for a person to disengage from the goal-similar distractor. Similarly, the greater the time of maximum deviation, the slower a person was overall from disengaging with the goal-similar distractor.

Reliability analyses

Before any meaningful individual difference relationships can be explored, attention capture indices must first be examined for reliability. This is in response to previous research [48] which found that most attention capture paradigms produced capture scores with little to no reliability. We calculated reliability by splitting the data into odd and even trials, aggregating them to the participant level, calculating selectivity and disengagement indices, and then correlating the odd data set against the even data set. The reliability scores for our selectivity and disengagement indices are summarized in Table 1. Reliability for the disengagement measure was r = .85 for area under the curve derived index and r = .91 for time of maximum deviation derived index. Reliability for the selectivity index was r = .67 for area under the curve derived scores, but only r = .38 for time of maximum deviation derived scores. Overall, these reliability scores are either within the same range or much higher than similar measures derived from response times [48]. For their RT-based contingent capture paradigm, Roque, Wright, & Boot [48] report split-half reliability scores that ranged from as low as r = .34 (experiment 1) to as high as r = .48 (experiment 2). Thus, our reliability scores were comparable at worst, but considerably higher at best.

Table 1

Primary analyses.

	Reliability Correlations						TSC Correlations
	AUC			TMAD			AUC			TMAD
	r	t-value	p	r	t-value	p	r	t-value	p	r	t-value	p
Selectivity	0.67	8.94	< .001	0.38	4.05	< .001	0.02	0.233	0.816	0.06	0.546	0.587
Disengagement	0.85	15.65	< .001	0.91	21.73	< .001	-0.01	-0.094	0.925	-0.04	0.364	0.717

Reliability and trait self-control correlations for selectivity and disengagement for area under the curve (AUC) and time of maximum deviation (TMAD).

As shown in Table 1, trait self-control did not correlate with selectivity or disengagement for either area under the curve or time of maximum deviation. Bayes Factors were also calculated (B10) to examine relative evidence for the null vs. alternative hypotheses. B10 ranged from .13 to .15, indicating moderate support for the null [49]. For ease of visualization, we have provided a plot of the area under the curve broken down by participants with low, medium, and high self-control (see Fig 3). Trajectories look virtually identical across different levels of trait self-control.

Fig 3

Trajectories.

The trajectories across low, medium, and high trait self-control (TSC) groups. Note, data was analyzed continuously; these categorical groups are for visualization purposes only.

Note, it is also possible that disengagement from goal dissimilar cues would be a better measure of an individual’s trait self-control. To explore this, we reran the correlations using mouse tracking parameters for mismatching cues. This did not change the pattern of results (all rs < .1, all ps > .5). In addition, we also ran analyses controlling for other individual differences such as basic demographic information (age, gender, and parents’ education and income). Controlling for these demographic variables did not change the pattern of results. See Table B in S2 Materials.

Exploratory analyses

Given these null results, we also conducted exploratory correlations to test the relationships between trait self-control and initiation time and response time. Initiation time represents the time between a trial starting and the first movement of the mouse, whereas response time represents the time it takes to make one’s selection. As shown in Table 2, neither the selectivity nor disengagement measures for response time and initiation time correlated with trait self-control scores. Note that reliability was quite low for response time selectivity (r = .23) and initiation time selectivity (r = .16).

Table 2

Exploratory analyses.

	Reliability Correlations						TSC Correlations
	RT			IT			RT			IT
	r	t-value	p	r	t-value	p	r	t-value	p	r	t-value	p
Selectivity	0.23	2.168	0.032	0.16	1.607	0.111	0.13	1.338	0.184	-0.07	-0.66	0.511
Disengagement	0.9	21.039	< .001	0.95	31.208	0.001	-0.06	-0.567	0.572	-0.04	-0.445	0.657

Reliability and trait self-control correlations for selectivity and disengagement for response time (RT) and initiation time (IT).

Discussion

Overall, attention capture was unrelated to differences in trait self-control. Trait self-control did not correspond to differences in measures of capture nor in subsequent disengagement. This suggests that the relatively micro-level phenomenon of attention capture and the relatively macro-level personality trait of self-control are orthogonal to one another.

A closer look: Trait self-control

Is it surprising that trait self-control did not affect attention capture? On the one hand it is surprising because a number of studies have found differences in response dynamics due to differences in trait self-control. For instance, prior research [50] presented participants with healthy and unhealthy food items (e.g., apple and cake) and had them pick the healthy option while tracking their mouse movements. Compared to participants with lower trait self-control, participants with higher self-control showed less conflict in their decision-making processes (i.e., less area under the curve) and did so with smoother, less abrupt trajectories. In contrast, other research [51] also looked at trait self-control differences using mouse-tracking but found a different pattern of results. While they did not find spatial differences (e.g., area under the curve or maximum absolute deviation) between participants with high and low self-control, they did find temporal differences (i.e., time of maximum deviation) in trajectories. Participants with high self-control inhibited their responses quicker than low self-control participants, as evidenced by faster time of max deviation. These temporal differences are consistent with research showing that “tastiness” information influences trajectories sooner than “healthfulness” information for people with low self-control [52-53]. Nonetheless, our study found neither spatial nor temporal differences within the domain of attention capture, which is inconsistent with prior research linking trait self-control to spatial differences reflective of magnitude differences [50] and to temporal differences reflective of processing differences [51, 52]. Perhaps this is due to the relatively different cognitive processes involved in the experimental paradigms. The contingent capture paradigm is an abstract, relatively low-level task whereas the selection of dietary preferences is a relatively high-level task that depends on motivation and values and closer approximates real world behavior.

On the other hand, it may not be surprising that self-control did not affect attention capture. There is evidence that successful self-regulation may largely depend on automatized, habitual processes rather than deliberative ones [15, 24– 25, 55]. One possibility is that self-control differences do affect attention capture but that self-control strategies must be deployed deliberately and repeatedly in order to later automatize selection and/or disengagement from attention capture. Prior work demonstrates that briefly training participants to automatically approach goal-supporting targets (i.e., pushing a joystick toward healthy stimuli) and avoid goal-conflicting targets (i.e., pulling a joystick away from unhealthy stimuli) led to increased efficiency in an attention task as well as to a higher likelihood of selecting a healthy, but not unhealthy, snack (similar results were demonstrated with school vs. party-related stimuli and intentions to study; [54]). Hence, future research may benefit by testing whether training participants in the attention capture task would lead to more efficient automatization of selectivity and disengagement for people higher, but not lower, in trait self-control. As a proxy for a training task, we evenly divided trials into early, middle, and late and then correlated selectivity and disengagement scores with trait self-control scores. Learning within the task did not affect correlations (all ps > .5).

The goal in the current study was to test whether the mechanisms related in trait self-control are related to the mechanisms involved in the relatively low-level phenomenon of attention capture. Although the mechanisms appear to be unrelated, it is quite possible that trait self-control would be correlated with “higher order” tasks, such as a mouse-tracking task where participants select between healthy and unhealthy items. Indeed, we have previously cited some research showing that this the case [50–52). As such, future research may wish to investigate exactly how “low” the mechanisms of trait self-control operate. At what point in the continuum between perceptual tasks and cognitive tasks does trait self-control function?

Reliability

Our results contribute to the attention capture literature in general by providing a paradigm that has a greater degree of reliability, particularly with spatial metrics such as area under the curve. Previous research [48] found that reliability across classic attention capture paradigms was generally low. In bottom-up paradigms, split-half reliability ranged from as low as .12 in the irrelevant singleton task to no higher than .28 in the onset cueing task; similarly, in top-down paradigms, split-half reliability ranged from as low as .34 in contingent cuing (experiment 1) to no higher than .48. In contrast, our modified contingent capture paradigm had higher reliability. The reliability of area under the curve derived measures—which provide crucial spatial information of capture—was .67 for our selectivity measure (a difference measure) and .85 for our disengagement measure. In addition, the time of maximum deviation had a reliability of .91 for our disengagement measure. These reliability scores suggest that continuously tracking a participant’s mouse is more reliable a measure of capture than outcome based temporal measures like response time.

Limitations

The current study has a number of limitations to consider. For one, although we created selectivity and disengagement scores as proxies for different mechanisms of attention capture, exploratory analyses found a number of correlations amongst the scores (e.g., AUC: selectivity and disengagement .864; a full correlation table is shown in Table C in S2 materials). These correlations could be indicative of several different things. One possibility is that selectivity and disengagement scores are measuring the same mechanism, despite our intention to have them measure different mechanisms of attention capture. Another possibility is that selectivity and disengagement scores measure different mechanisms, but the two mechanisms are highly related to one another. For example, it may be that selectivity and disengagement are independent mechanisms, but people with high selectivity may tend to also have high disengagement. Future research is needed to tease these possibilities apart.

Conclusion

Our study suggests that attention capture is impenetrable by trait self-control. Despite the mechanistic similarities, it appears self-control and visual attention recruit and rely on different mechanisms of information selection and behavioral inhibition. The ability to say no to a slice of cake and the ability to resist attention capture appear to be independent and orthogonal to one another. One may be distracted by a cake and not want to eat it too.

Area under the curve.

Errors bars represent confidence intervals.

(EPS)

Click here for additional data file.

Response time.

Errors bars represent confidence intervals.

(EPS)

Click here for additional data file.

Initiation time.

Errors bars represent confidence intervals.

(EPS)

Click here for additional data file.

Time of maximum absolute deviation.

Errors bars represent confidence intervals.

(EPS)

Click here for additional data file.

(DOCX)

Click here for additional data file.

(DOCX)

Click here for additional data file.

6 Sep 2019

PONE-D-19-20912

Trait Self-Control Does Not Predict Attentional Control: Evidence from a Novel Attention Capture Paradigm

PLOS ONE

Dear Mr. Dieciuc,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Thank you for submitting your very nice paper to PLOS One.  The paper  is almost ready for publication, but before I accept it I would like you to perform the following additional analyses of the data:

1) As suggested by referee 1, please compare behavior  in the early rounds of a session with behavior in the later ones.

2) As suggested by referee 2, please run additional analyses that include the available demographic information. If you find that this has little impact and/or that the controls are statistically insignifcant then you could mention that in a footnote; otherwise it would seem natural to include it in the text.

3) As also suggested by referee 2, please do some analysis of the te the influence of self-control on attentional capture- regression is one way to do that but you may prefer others.

Both referees also make interesting suggestions for further experiments. I encourage you to consider following up on these suggestions in future work.

sincerely

Drew Fudenberg

==============================

We would appreciate receiving your revised manuscript by Oct 21 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Drew Fudenberg

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Additional Editor Comments (if provided):

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This study investigates whether trait self-control correlates with individual susceptibility to attention capture. Self-control is measured using self-reported perception of self-control with various questions on a scale, while susceptibility to attention capture is measured using a novel experimental paradigm which relies on mouse tracking.

The article makes 2 main contributions.

The first contribution is methodological: the authors propose a novel experimental task to measure low-level attention capture and show that their resulting (pre-registered) quantitative indexes are much more reliable than those previously used in the literature (as measured by the degree of internal correlation within-subject).

The second key contribution is a null result: trait self-control is not correlated with susceptibility to attention capture. This contribution is significant because:

- Both self-control and resistance to attention capture are thought to employ similar mechanisms, relying on inhibitory control, albeit at a high cognitive level for self-control and at a low cognitive level for attention capture;

- Self-control and resistance to attention capture have been shown to correlate in an experimental paradigm which involves utility-relevant distractions (e.g. food items). Yet it was unclear if this correlation would survive if the attention capture environment was made lower-level with purely perceptual stimuli. This study provides a negative answer to this question, suggesting the correlation does rely on the nature of the distraction and is not due to low-level cognitive mechanisms involved in attention control (note: this is an important consideration and I find the article would benefit if this was mentioned in the introduction, not only in the conclusion) ;

- Both self-control and susceptibility to attention capture have been shown to correlate with working memory.

Minor comments:

- An eye-tracking based version of the experiment would seem like a more natural paradigm, because it would not rely on the assumption that “the partial and tentative conclusions of perceptuo-cognitive systems continuously feed into and affect the motor system” (which the authors correctly mention). The article would benefit from a discussion justifying the choice of mouse-tracking, unless it was a purely practical choice.

- In the “Predictions” section, the verb “affect” is repeatedly used (“self-control may affect only selectivity” etc.), suggesting a causal link. However nothing in the literature cited nor in the data collected supports a causal relationship. Hence it would be more appropriate to use a verb such as “correlates with”.

- The self-control scale used suffers from being purely based on a self-report. The study would benefit from comments about how well the specific scale used here predicts revealed preferences measures such as actual intertemporal choices.

- The authors mention that it would be interesting to investigate whether learning arises and self-control does help resist attention capture but only after some training. I agree that this is an interesting follow-up question. But while the authors have no data on training, they do have a lot of trials per participant. They thus could provide a partial answer by comparing behavior in the later trials to behavior in the earlier trials in the experiment. It would be beneficial to run such tests and add some words on these.

Reviewer #2: Summary:

This work measures attention capture—the tendency to briefly respond to extraneous stimuli before returning to a primary task—via a mouse-tracking experiment, with the goal of assessing the correlation between that behavior and a questionnaire measuring self-control. Specifically, the work seeks to examine the effect of self-control on selectivity (whether the strength of attention capture increases as the distractor becomes more similar to the goal) and disengagement (how easy it is for the subject to return to the goal after being distracted). In the experiment, subjects are tasked with clicking a colored X which appears in one of four boxes, and are distracted by the flashing outline of a randomly-selected box before doing so. By changing the color of the distracting stimulus to match or differ from that of the goal, the work replicates findings in the literature wherein similar stimuli are more effective in capturing attention.

To provide a description of the magnitude of attention capture, the work computes both the distance (in pixels) added to a straight-line trajectory by the distracting stimulus and the time taken by the subjects to begin moving their mouse towards the goal. Selectivity is measured by comparing each subject’s behavior when the distractor’s color matched the goal to their behavior when it did not; disengagement is measured by restricting analysis to the case where the distractor’s color matched the goal, but the distractor was on the opposite side of the screen. The work splits the data set in two and examines between-subsample correlations in both the attention-capture and self-control measures, finding a high degree of reliability. However, the work finds near-zero correlations between self-control and their attention capture measures, and the mouse trajectories of the participants show no clear differences regardless of self-control scores. This result is taken as grounds to claim that attention capture and self-control operate through different cognitive mechanisms.

Comments:

Following the seven PLOS ONE Criteria for Publication,

(1) Primary results of original research: criterion met; no further comments.

(2) Results not published elsewhere: criterion met, this work has only been submitted to PLOS ONE; no further comments.

(3) Experiments, statistics, analyses are high-standard and described in detail: descriptions are clear and reproducible, but some additional analysis is needed to confirm the reliability and consistency of the measures of attention capture and address other potential variables of interest

a. It would be good to see the correlation between the two measures of attention capture, both within-sample (i.e., compare each participant to themselves) and “out-of-sample” (i.e., compare on aggregate or using subsampling). These results would provide stronger evidence of whether these measures are really capturing the same behavior.

b. Along the lines of (2), there are no demographic/socioeconomic/other controls included in the correlation analysis. The work notes that attention-capture and self-control may vary due to individual-level traits (e.g. video-game-playing); is the goal of the study to capture the correlation after taking into account these personal idiosyncrasies or to extract some more “baseline” measure, or one that can be adjusted for interactions with various personal factors? If it is the latter, then including controls would help to provide a clearer picture of how self-control and attention capture interact without the influence of particular personal traits.

(4) Conclusions supported by the data: some scope for clearer causal analysis and further experiments to more directly involve the processes of self-control

a. In addition to examining the correlation between attention capture and self-control, it seems valuable to make some effort at examining the influence of one on the other, e.g., regressing attention-capture measures on self-control scores. While there is potential for reverse causality (sensitivity to attention capture influences the ability to develop self-control), given that self-control is the macro-level process it seems more likely to be the overall driver of behavior. Especially if both attention-capture measures are used, care must be taken in interpreting the coefficients, but this approach could still provide another angle on the relationship between the two variables.

b. The measure of disengagement includes only goal-similar distractors; it is unclear why this is the optimal measure of disengagement. It may be that self-control does not influence goal-similar attention capture but does improve disengagement from goal-dissimilar attention capture—in the dessert vs. healthy eating example, self-control may improve the ability to reject a goal-dissimilar slice of cake, but not a goal-similar fruit tart. Investigating the effect of self-control on goal-dissimilar distractors, or the overall measures of attention capture not split by goal-similarity, would provide a richer picture of this effect.

c. In addition to using questionnaire measures of self-control, and in order to more directly invoke self-control behavior, additional tasks that take advantage of the mouse-tracking framework could be implemented. The work notes that self-control can vary according to “situational demands,” so using tasks that more directly invoke self-control (e.g., clicking on the “healthy” food from among various “unhealthy” options) would provide a valuable point of comparison to the more abstract task used here.

(5) Presented in intelligible fashion: criterion met; no further comments.

(6) Meets ethics standards: criterion met; IRB documentation provided, no further comments.

(7) Data availability: criterion met; analysis is preregistered and data is available, no further comments.

Conclusions

This work lays out an experimental method that, while it could use refinement, provides a great level of detail about the temporal and spatial behaviors associated with attention capture without requiring particularly resource-intensive hardware and software. There is room to take further advantage of this method to obtain a more complete picture of attention capture, which may include a greater variety of tasks, the introduction of more true-to-life goals and distractions, and clearer measures of disengagement in different contexts. In addition to these extensions of the existing experiment—extensions that could provide more direct experimental evidence of the role of self-control—the analysis of existing data could be strengthened by considering alternative statistical approaches and a richer set of potential independent and dependent variables. These modifications would help take advantage of what appears to be a powerful approach to answering important psychological questions.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

22 Oct 2019

We would like to thank the reviewers for their overall positive assessment of the manuscript. We found their feedback to be helpful in clarifying some ambiguities and in strengthening the message of the manuscript. Below we have distilled each reviewers major suggestions and provided a detailed explanation of how they were addressed.

Reviewer # 1

-I find the article would benefit if this [the lack of correlation between attention capture and TSC] was mentioned in the introduction, not only in the conclusion)

Thank you for this suggestion, we now preview this result in the introduction on page 10.

“To preview our results, our data are most consistent with the fourth possibility. Namely, self-control scores did not correlate with either selectivity or disengagement. This suggests that attention capture and trait self-control operate via different mechanisms.”

-The article would benefit from a discussion justifying the choice of mouse-tracking [compared to eye-tracking], unless it was a purely practical choice.

We addressed this by adding the following text to the manuscript, pages 8-9:

“Our decision to use mouse-tracking instead of a similar methodology like eye-tracking was based on a number of advantages related to measuring how cognitive processes play out over time. While saccades are largely ballistic in nature and their flight paths are often completed in tens of milliseconds, mouse-tracking unfolds over a much larger timescale, providing a rich look at both spatial and temporal dimensions of behavior. On another level, mouse-tracking was chosen over eye-tracking for a number of practical reasons. Mouse-tracking software is freely available (Kieslich & Henninger, 2017; Freeman & Ambady, 2010) to anyone who has access to a computer. This high-accessibility makes it an appealing method of conducting research, one that is particularly conducive to encouraging replication and reproducibility. Finally, mouse-tracking is a relatively new methodology compared to eye-tracking, especially within the field of attention capture and self-control; thus, given the novelty of the methodology, we wanted to deepen the field’s knowledge.”

In the “Predictions” section, the verb “affect” is repeatedly used…suggesting a causal link. However nothing in the literature cited nor in the data collected supports a causal relationship. Hence it would be more appropriate to use a verb such as “correlates with”.

Thank you for pointing out this mistake. The text has been changed so that affects has been replaced with “correlates with.”

The self-control scale used suffers from being purely based on a self-report. The study would benefit from comments about how well the specific scale used here predicts revealed preferences measures such as actual intertemporal choices.

The trait self-control scale used here (Tangney, Baumeister, & Boone, 2004) reflects a macrolevel disposition that effects a wide variety of behavioral outcomes, including those in the domains of school and work, eating and weight, interpersonal functioning, and wellbeing and adjustment (for meta-analysis, see de Ridder et al., 2012). With respect to future discounting in particular, self-control as assessed with this scale has been negatively associated with delayed discounting (e.g., Franco-Watkins, Mattson, & Jackson, 2016; Kool, McGuire, Wang, & Botvinick, 2013; Pang, Otto, & Worthy, 2015)

Franco‐Watkins, A. M., Mattson, R. E., & Jackson, M. D. (2016). Now or later? Attentional processing and intertemporal choice. Journal of Behavioral Decision Making, 29(2-3), 206-217.

Kool, W., McGuire, J. T., Wang, G. J., & Botvinick, M. M. (2013). Neural and behavioral evidence for an intrinsic cost of self-control. PloS one, 8(8), e72626.

Pang, B., Otto, A. R., & Worthy, D. A. (2015). Self‐Control Moderates Decision‐Making Behavior When Minimizing Losses versus Maximizing Gains. Journal of Behavioral Decision Making, 28(2), 176-187.

-The authors mention that it would be interesting to investigate whether learning arises and self-control does help resist attention capture but only after some training…They thus could provide a partial answer by comparing behavior in the later trials to behavior in the earlier trials in the experiment. It would be beneficial to run such tests and add some words on these.

We followed this suggestion and evenly split up our data into early, middle, and late trials. Thereafter, we created selectivity and disengagement measures and correlated them with trait self-control scores. We did this once for early trials and once for late trials. It seems that learning within the context of the experiment did not affect correlations (all ps > .05). The exact statistics for each of these correlations is shown in the table below.

We explained this by adding the following text to a footnote in the manuscript on page 22:

“As a proxy for a training task, we evenly divided trials into early, middle, and late and then correlated selectivity and disengagement scores with trait self-control scores. Learning within the task did not affect correlations (all ps > .5).”

REVIEWER # 2

It would be good to see the correlation between the two measures of attention capture, both within-sample (i.e., compare each participant to themselves) and “out-of-sample” (i.e., compare on aggregate or using subsampling). These results would provide stronger evidence of whether these measures are really capturing the same behavior.

As requested we have provided a within-sample correlation matrix of the attention capture measures. We were not entirely certain what the reviewer meant by “out-of-sample” correlations. If the reviewer wishes to specify, we would be happy to follow up with it.

The correlation matrix below shows a number of correlations amongst our measures (e.g., AUC: selectivity and disengagement, r = .864). While these correlations do, in fact, suggest that the selectivity and disengagement scores are measuring the same overall behavior, it does not tease apart whether selectivity and disengagement scores are measuring the same or different mechanisms. For instance, the high correlation may suggest that selectivity and disengagement scores are actually measuring the same single mechanism of attention capture. However, another possibility is that selectivity and disengagement are measuring two different mechanisms that are themselves highly rated. For instance, people with high selectivity may tend to also have high disengagement. The current study cannot tease these different possibilities apart. To address this, we added the correlation table to our supplementary materials (Table S3) and the following text to the manuscript on p23:

“The current study has a number of limitations to consider. For one, although we created selectivity and disengagement scores as proxies for different mechanisms of attention capture, exploratory analyses found a number of correlations amongst the scores (e.g., AUC: selectivity and disengagement .864; a full correlation table tis shown in Table S3). These correlations could be indicative of several different things. One possibility is that selectivity and disengagement scores are measuring the same mechanism, despite our intention to have them measure different mechanisms of attention capture. Another possibility is that selectivity and disengagement scores measure different mechanisms, but the two mechanisms are highly related to one another. For example, it may be that selectivity and disengagement are independent mechanisms, but people with high selectivity may tend to also have high disengagement. Future research is needed to tease these possibilities apart.”

Along the lines of (2), there are no demographic/socioeconomic/other controls included in the correlation analysis. The work notes that attention-capture and self-control may vary due to individual-level traits (e.g. video-game-playing); is the goal of the study to capture the correlation after taking into account these personal idiosyncrasies or to extract some more “baseline” measure, or one that can be adjusted for interactions with various personal factors? If it is the latter, then including controls would help to provide a clearer picture of how self-control and attention capture interact without the influence of particular personal traits.

This is a great point. Although we did not collect data on individual differences such as video game playing (which we mention as relevant to highlight that individual differences can modulate attention capture and disengagement processes), we collected data on basic demographic variables, such as age, gender, and parents’ education and income. (Note that because our sample includes student participants, their educational attainment and income do not really vary.) Our primary concern is the association between trait self-control and individual differences in distractibility, namely, attention capture and disengagement. Hence, although we are not focused on the interaction between self-control and demographic variables here, it is important to control for them.

Accordingly, we conducted correlation analyses controlling for these variables. Notably, this did not change the pattern of results that emerged without controls. This table is included in the Supplementary Materials and the results are mentioned in footnote 2 of the manuscript.

Pearson Correlations, controlling for age, gender, ethnicity, race, parental education, and parental income.

TSC AUC_SELECTIVITY AUC_DISENGAGE TMAD_SELECTIVITY TMAD_DISENGAGE

TSC —

AUC_SELECTIVITY .057 —

AUC_DISENGAGE .026 .866*** —

TMAD_SELECTIVITY .121 -.110 .043 —

TMAD_DISENGAGE -.036 -.643*** -.684*** .004 —

*p < .05, **p < .01, ***p < .001

It seems valuable to make some effort at examining the influence of one on the other, e.g., regressing attention-capture measures on self-control scores.

We ran linear regressions and used our attention capture scores to predict trait self-control. As shown in the table below, neither attention capture score significantly predicted trait self-control scores.

B SE B β t p

Intercept 3.362 1.03E-01 1.03E-16 32.745 < .001

AUC Selectivity 2.58E-06 4.13E-06 0.125 0.623 0.535

Disengagement -2.00E-06 3.42E-06 -0.118 -0.586 0.559

Intercept 3.49 0.344 3.487 10.147 < .001

TMAD Selectivity 0.0012956 0.002 0.056 0.551 0.583

Disengagement 0.0003971 0.001 -0.038 0.373 0.71

The measure of disengagement includes only goal-similar distractors… Investigating the effect of self-control on goal-dissimilar distractors, or the overall measures of attention capture not split by goal-similarity, would provide a richer picture of this effect.

To address the possibility that goal-dissimilar distractors might be a better measure of disengagement, we ran correlations between trait self-control and area under the curve and time of maximum deviation on mismatching cues. As shown in the table below, none of the measures correlated.

r t-value p

AUC -0.06 -0.57 .570

TMAD -0.06 -0.561 .576

We supplemented the text with the following footnote found on page 17:

“Note, it is also possible that disengagement from goal dissimilar cues would be a better measure of an individual’s trait self-control. To explore this, we reran the correlations using mouse tracking parameters for mismatching cues. This did not change the pattern of results (all rs < .1, all ps > .5).”

In addition to using questionnaire measures of self-control, and in order to more directly invoke self-control behavior, additional tasks that take advantage of the mouse-tracking framework could be implemented. The work notes that self-control can vary according to “situational demands,” so using tasks that more directly invoke self-control (e.g., clicking on the “healthy” food from among various “unhealthy” options) would provide a valuable point of comparison to the more abstract task used here.

We agree with the reviewer that these are interesting questions to ask and worthy of further research. Throughout the paper we cited a few studies that have examined trait self-control’s relationship to nutritional decisions in a mouse-tracking framework. We more explicitly address this by adding in the following text to page 22:

“The goal in the current study was to test whether the mechanisms related in trait self-control are related to the mechanisms involved in the relatively low-level phenomenon of attention capture. Although the mechanisms appear to be unrelated, it is quite possible that trait self-control would be correlated with “higher order” tasks, such as a mouse-tracking task where participants select between healthy and unhealthy items. Indeed, we have previously cited some research showing that this the case (Stillman, Medvedev, & Ferguson, 2017; Gillebaart et al., 2016; Sullivan et al., 2015). As such, future research may wish to investigate exactly how “low” the mechanisms of trait self-control operate. At what point in the continuum between perceptual tasks and cognitive tasks does trait self-control function?”

24 Oct 2019

Trait Self-Control Does Not Predict Attentional Control: Evidence from a Novel Attention Capture Paradigm

PONE-D-19-20912R1

Dear Dr. Dieciuc,

Thank you for your responsive revision,  I am now happy to accept the paper for publication. Bolierplate follows below

----.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Drew Fudenberg

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

21 Nov 2019

PONE-D-19-20912R1

Trait Self-Control Does Not Predict Attentional Control: Evidence from a Novel Attention Capture Paradigm

Dear Dr. Dieciuc:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Drew Fudenberg

Academic Editor

PLOS ONE