Is poor control over thoughts and emotions related to a higher tendency to delay tasks? The link between procrastination, emotional dysregulation and attentional control

Abstract

A tendency to procrastinate has previously been linked to low attentional control and poor emotion regulation skills. Building upon these findings, in the present study we investigated whether the relationship between procrastination and attention can be corroborated and explained by emotion dysregulation and dispositional spontaneous mind‐wandering. University students completed questionnaires along with the Attentional Networks Test for Interactions and Vigilance – executive and arousal components. The results showed that trait procrastination is inversely related to some indices of executive and arousal vigilance. Interestingly, the direct effects between trait procrastination and vigilance scores disappeared when emotion dysregulation or dispositional spontaneous mind‐wandering were included in the model. Obtained findings suggest that difficulties in managing emotional reactions and poor control over the focus of one's thoughts might explain the relationship between low attentional control and increased chronic procrastination.

BACKGROUND

Procrastination is a voluntary and irrational delay of task completion or initiation. Voluntary, as there is no important obstacle that prevents an individual from performing the task; and irrational, as a procrastinating person is aware that not completing the task on time might have adverse outcomes, including feelings of guilt and anxiety. These are the characteristics that differentiate procrastination from other forms of delay, such as strategic delay, in which the perceived positive consequences of postponing the task outweigh anticipated negative outcomes (see Mahy et al., ²⁰²⁴; Klingsieck, ²⁰¹³ for reviews).

The problem of procrastination appears to be particularly prevalent in educational settings – studies often show that rates of procrastination are higher among students compared to other populations (Steel, 2007; Svartdal et al., ²⁰¹⁶; although see also Wypych et al., ²⁰¹⁸ for contradictory results). This can be partially explained by the characteristics of academic tasks, as the probability of procrastinatory behaviours might depend on different environmental factors, such as work supervision or the proximity of deadlines (e.g. Corkin et al., ²⁰¹⁴; Kawczyński et al., ²⁰²³). However, some individuals procrastinate more than others and this tendency is relatively stable across time (Rice et al., 2012), which provokes questions about the underlying mechanisms of this maladaptive disposition. This is of particular importance, as chronic procrastination among students has been associated with multiple negative consequences, such as lower academic achievements and learning outcomes, or increased levels of negative emotions (Goda et al., 2015; Kim & Seo, ²⁰¹⁵; Rahimi et al., ²⁰²³). In the present study, we aimed to gain a deeper understanding of this problem by investigating the relationships between a tendency to procrastinate, attentional control and emotion regulation among higher education students.

Procrastination and attentional control

Temporal motivation theory coined by Steel and König (2006) states that a tendency to procrastinate can be exacerbated by such factors as impulsivity and distractibility, linked to higher sensitivity to delay and lower control of attention (Steel et al., 2018). Being unable to maintain focus on task objectives can facilitate the attentional shift towards alternative goals and activities, such as watching series or meeting with friends, despite the initial plans to study for the exam. Indeed, some studies have shown that trait procrastination is related to reported attentional dysfunctions, such as inability to sustain attention on the performed task, being easily distracted by external stimuli, or frequently experiencing spontaneous mind‐wandering (e.g. Fernie et al., ²⁰¹⁶; Moon et al., ²⁰²⁰; Rebetez et al., ²⁰¹⁸; Steel et al., ²⁰¹⁸). Interestingly, a recent study showed that problems with focusing on a task were perceived by students as one of the causes of their procrastination (Kawczyński et al., 2023). Taking into account these findings, it comes as no surprise that procrastination has been related to increased symptoms of attention deficit hyperactivity disorder (ADHD) and, conversely, that individuals diagnosed with ADHD often declare an increased tendency to procrastinate (Altgassen et al., 2019; Bolden & Fillauer, ²⁰¹⁹). Moreover, stimulants used to treat ADHD symptoms have been shown to modulate procrastination‐related factors by increasing motivation for cognitive effort and reducing avoidance of tasks that require waiting for rewards over those providing immediate gratification (i.e. avoiding delay aversion; e.g. Addicott et al., ²⁰¹⁹; Hofmans et al., ²⁰²⁰; Low et al., ²⁰¹⁸).

The link between attentional deficits and procrastination has also been confirmed with the use of more objective measures. For example, when compared to low procrastinating students, chronic procrastinators performing a Parametric Go/No‐Go task presented increased reaction time variability (RTV; Michałowski et al., ²⁰²⁰; Wiwatowska et al., ²⁰²²), which seems to reflect reduced vigilance (Lim & Dinges, 2008; Luna et al., ²⁰¹⁸) or difficulties in sustaining attention during task completion (MacDonald et al., 2009; Weissman et al., ²⁰⁰⁶).

The above‐mentioned attentional dysfunctions in procrastination might be rooted in structural and functional changes within certain brain regions. Namely, a tendency to delay tasks has been linked to reduced activity and decreased volume of the dorsolateral prefrontal cortex (dlPFC; Chen et al., ²⁰²⁰; Hu et al., ²⁰¹⁸; Wypych et al., ²⁰¹⁹), which is a brain area responsible for maintaining cognitive control and sustained attention throughout task completion (Qiao et al., 2022; Sarter et al., ²⁰⁰¹). A recent study (Xu et al., 2023) found that stimulating this brain area increases motivation to complete tasks. Furthermore, increased dispositional procrastination has been related to higher activity of regions that comprise the default mode network (DMN; Zhang et al., ²⁰¹⁶). Higher activation of this network has been linked to states of internally oriented attention (e.g. mind‐wandering episodes) as opposed to states of attention focused on external stimuli (e.g. Poerio et al., ²⁰¹⁷). Therefore, increased reactivity of the DMN along with reduced dlPFC activation might compromise the ability to control procrastinators' focus of attention and to suppress unwanted task‐unrelated thoughts.

In the presented study, we aimed to investigate the relationship between a tendency to procrastinate and a capacity to deploy attentional control measured during the completion of a cognitive task. Although, as described above, this link has already been identified in previous studies, they either relied on self‐reported measures of attentional control (e.g. Fernie et al., ²⁰¹⁶) or compared indices of attention between extreme groups of low vs. high procrastinating participants (e.g. Michałowski et al., ²⁰²⁰). Moreover, it remains unclear which components of attentional control are impaired in procrastination. The present study aims to fill this gap by measuring different components of attention, with a particular focus on vigilance.

Vigilance is a term frequently used interchangeably with sustained attention, defining the ability to maintain sufficient levels of attention over time. It is usually evaluated with tasks in which participants have to respond to infrequent, auditory or visual targets (Liu & Falahpour, 2020). With time on task, people usually react slower, with more variability and lower accuracy. This performance decline has been conceptualized as vigilance decrement. According to the different theories (Esterman & Rothlein, 2019; Thomson et al., ²⁰¹⁵), this decrement is caused by changes within arousal, executive control and motivation. Arousal levels determine the available resources and the selectivity of attention, while executive control is responsible for distributing these cognitive resources between the performed task and task‐unrelated processes, such as mind‐wandering. It seems that the mind has a natural tendency to wander towards task‐unrelated thoughts, especially when the performed task is monotonous and boring. Opposing this tendency in such contexts is often perceived as effortful and requires motivation to further engage in the task while maintaining sufficient vigilance levels (Bonnefond et al., 2011; Dillard et al., ²⁰¹⁹). What is important, deficits in any of the aforementioned components (arousal levels, executive functions, or motivation) might underlie diminished capacity to sustain attention over time (Bozhilova et al., 2018; Bubnik et al., ²⁰¹⁵).

As increased trait procrastination has been previously inversely related to self‐reported attentional control abilities (e.g. Steel et al., ²⁰¹⁸), perseverance (e.g. Rebetez et al., ²⁰¹⁸), or willingness to engage in mentally effortful activities (Rakes & Dunn, 2010; Ziegler & Opdenakker, ²⁰¹⁸), we predicted that it would also be correlated with higher vigilance decrement during task completion. Furthermore, we speculated that the tendency to experience spontaneous mind‐wandering would mediate this relationship, considering earlier findings on the positive link between a tendency to delay tasks and the frequency of daydreaming (Rebetez et al., 2018). Our reasoning behind this hypothesis was that a diminished capacity to maintain focus on the tasks might underlie a higher tendency to experience spontaneous mind‐wandering episodes, during which attention is redirected away from task completion towards alternative activities, thereby increasing the probability of procrastinatory behaviours.

Procrastination and emotion regulation

Aside from considering impulsivity and low attentional control as potential causes of procrastination, other scholars point out that this behaviour is a maladaptive way of coping with negative emotions evoked by aversive (e.g. difficult or boring) tasks. Given the aversive nature of cognitive control (Inzlicht et al., 2015), engaging in easier, more pleasurable (and therefore not attentionally demanding) activities can temporarily increase one's mood, even if it is at the expense of negative consequences for the future self's wellbeing (Sirois & Pychyl, 2013). The reward in the form of short‐term mood improvement might reinforce the formation of habitual delay in certain contexts, leading to chronic procrastination (Mahy et al., 2024).

Frequent use of this dysfunctional strategy of dealing with task‐related aversion might be caused by deficits in emotion regulation, which is broadly defined as an ability to properly identify and manage one's affective reactions in order to achieve goals (Gratz & Roemer, 2004; McRae & Gross, ²⁰²⁰). Indeed, several studies have found that an increased tendency to procrastinate is associated with emotion dysregulation (Mohammadi Bytamar et al., 2020; Wartberg et al., ²⁰²¹; Wypych et al., ²⁰¹⁸). Xu et al. (2023) showed that this link might be explained by weaker functional connections between the insula and medial lateral superior frontal gyrus – regions that play an important role in processing and regulating emotions (Frank et al., 2014; Li et al., ²⁰²²). Moreover, improving emotion regulation skills has been shown to decrease the frequency of procrastinatory behaviours (Eckert et al., 2016; Schuenemann et al., ²⁰²²), which reveals the causal relationship between these phenomena.

Although attentional control and emotion regulation might seem as independent processes, research shows that they are highly interconnected (Birk et al., 2018; Fox & Calkins, ²⁰⁰³; O'Bryan et al., ²⁰¹⁷; Tortella‐Feliu et al., ²⁰¹⁴) and to some extent share neural substrates (e.g. Taylor & Liberzon, ²⁰⁰⁷). For example, it has been shown that lower capacity for attentional control predicts higher levels of fear in response to threatening stimuli (Richey et al., 2012), worse recovery from negative emotions elicited by recalling traumatic experiences (Bardeen & Read, 2010), and increased risk for developing post‐traumatic stress (Bardeen et al., 2015). What is more, attentional training can have a positive impact on emotion regulation (see Wadlinger & Isaacowitz, ²⁰¹¹ for a review).

A recent study showed that the abilities to regulate emotions mediated the relationship between ADHD symptoms and procrastination (Bodalski et al., 2022). Moreover, in a study conducted by Wypych et al. (2018) higher impulsivity predicted increased procrastination tendencies via the habitual use of maladaptive emotion regulation strategies. As both ADHD and impulsivity encompass attentional deficits (Malloy‐Diniz et al., 2007; Shaw et al., ²⁰¹¹) it might be expected that emotion dysregulation would also mediate the link between low attentional control and increased tendency to procrastinate. However, to our knowledge, none of the previously conducted studies tested this hypothesis, which is another goal of the present research.

Moreover, procrastinators' difficulties in regulating emotions and suppressing spontaneous mind‐wandering might contribute to higher disengagement from the performed task along with an increase in task‐related negative emotional reactions, which might further reduce motivation and facilitate procrastinatory behaviours. However, the link between a tendency to procrastinate and these subjective experiences evoked by attentionally demanding tasks has not been sufficiently examined in previous studies. Because of that, we secondarily aimed to investigate the relationship between trait procrastination and task‐related engagement, worry and distress.

Vigilance decrement measurement and the summary of formulated hypotheses

One of the tasks used to generate and measure decrement in executive control and vigilance is the Attentional Networks Test for Interactions and Vigilance – executive and arousal components (ANTI‐Vea; Luna et al., ²⁰¹⁸). The main advantage of this task is that it differentiates two aspects of sustained attention: executive vigilance, which relates to the ability to discriminate between infrequent targets and similar but more frequent non‐targets; and arousal vigilance, linked to the capacity to quickly react to stimuli without exerting much control. It has been observed that both vigilance types decline with time on this task, namely the response accuracy gets lower and the reaction time (RT) gets slower and more variable (e.g. Luna, Barttfeld, et al., ²⁰²¹; Luna, Lupiáñez, et al., ²⁰²¹; Luna, Roca, et al., ²⁰²¹).

ANTI‐Vea also allows for measuring three attentional networks derived from Posner and Petersen (1990) theory: orienting, which is the capacity to prioritize sensory input by selecting a modality or location; alertness, which reflects a temporary increase in arousal that facilitates response readiness; and executive control, reflecting the ability to select goal‐oriented responses in conflict situations. Although the primary focus of the study was to verify whether a tendency to delay tasks is related particularly to vigilance, we also decided to exploratively investigate the links between procrastination and these three attentional functions. We hypothesized that procrastination might be related to lower executive control, taking into account a previously observed relationship between procrastination and reported executive dysfunctions (Rabin et al., 2011). We did not formulate any hypotheses regarding the link with alertness and orienting as, in our opinion, the available evidence is insufficient to support them. Nevertheless, as the literature on the nature of attentional dysfunctions in procrastination is very limited, we believe that the obtained results might contribute to a better understanding of this problem. In addition, we decided to analyse the relationship between procrastination and attention span, a recently conceptualized aspect of vigilance that quantifies the amount of time for which an individual is able to maintain an efficient attentional state, as indexed by optimal sustained performance (Simon et al., 2023).

To sum up, based on the above‐mentioned findings, we have formulated and preregistered (https://osf.io/7m2au/?view_only=d9e169b4d7a5416d9fdc4a9f9434172e) the following hypotheses (it is important to emphasize that the term “procrastination” used below relates to a disposition – a tendency to procrastinate):

Higher procrastination would be related to reduced executive and arousal vigilance.

Procrastination would be related to larger decrements in executive and arousal vigilance with time on task.

The relationship between procrastination and vigilance scores would be mediated by higher emotional dysregulation and a higher tendency to experience spontaneous mind‐wandering.

Procrastination would be positively related to an increase in task‐related worrying and distress as well as to a decrease in task engagement.

In addition, we added the following hypotheses to the original preregistration plan:

Procrastination would be related to lower executive control.

Procrastination would be related to lower attention span.

The link between procrastination and attention span would be mediated by emotion dysregulation and a higher tendency to experience spontaneous mind‐wandering.

Although our hypotheses included mediation analyses to examine indirect effects, it is important to emphasize that the present study was not designed to identify causal relationships between the indicated variables, as causality can be conclusively established only through controlled experimental designs (which in some cases are difficult, or even impossible to deploy). However, in our models, we aimed to integrate described theory and research in order to gain deeper understanding of the relationship between attentional control and procrastination and to verify the statistical plausibility of the presented idea, which suggests a potential explanation for the tendency to procrastinate.

METHODS

Sample size justification

We used the recommendations formulated by Fritz and MacKinnon (2007) to identify the minimum sample size for simple mediation analyses. The recommended sample size is equal to 162 participants, taking into account that we planned to conduct the percentile bootstrap test of mediation and that, based on previous results (Michałowski et al., 2020; Wiwatowska et al., ²⁰²²) we expected to find small to medium effect sizes.

We also used the G*Power software (Faul et al., 2009) to conduct a power analysis for a two‐tailed correlation r = .20, with an alpha level of .05 and power of .80. The result of this analysis indicated that the required sample size is equal to 193.

Based on the above‐mentioned suggestions and analyses, we decided to recruit a minimum of 200 participants for our study.

Participants

Participants were students from different universities and colleges in Poland. They were recruited via social media platforms and universities' mailing lists. The exclusion criteria were as follows: (1) declared diagnosis of psychiatric or neurological disorders; (2) declared chronic use of psychoactive drugs or substances; and (3) uncorrected vision and hearing impairment.

226 participants completed all questionnaires and the whole ANTI‐Vea. Out of this sample, 20 participants were excluded: 10 subjects answered incorrectly to the attention check question; 6 subjects showed a hit rate lower than 60% in ANTI trials in at least one of the ANTI‐Vea blocks; and 4 subjects showed error rates or mean RTs higher than 3 SD from the group mean. The final sample consisted of 206 participants (M _age = 24.44, SD _age = 7.05; 110 identified as female, 93 identified as men, 3 identified as non‐binary).

Procedure

The study was conducted online. It was approved by the local Ethics Committee at the SWPS University and performed in accordance with the Declaration of Helsinki. Each participant has given informed consent to participate in the research. For participation in the study, the subjects had a chance to win a financial award in a lottery. One out of every twenty participants was randomly selected to win the 100 PLN (~22 EUR) award. Additionally, students from the SWPS University (69% of participants) were also awarded the course credit for taking part in the study.

Participants were asked to complete the study on a computer with working speakers in a quiet place without distractions as well as to turn off their mobile phones and other entertainment devices, such as television or radio. First, they signed informed consent of participation. Then, they answered questions regarding the inclusion criteria as well as age and gender. Next, they completed a set of questionnaires (see below), after which they were asked to perform a version of the ANTI‐Vea embedded in the online platform (Coll‐Martín et al., 2023) without any breaks during task completion. After task completion, participants completed the post‐task version of the Dundee Stress State Questionnaire (Matthews & Zeidner, 2012; see the Questionnaires section). The whole study was about 1 h long.

Task

The design of the task is presented in Figure 1. The detailed description can be found at https://anti‐vea.ugr.es/method.html. The task consisted of 6 blocks of 80 trials. Each block included the following trial types (presented in a randomized order):

1. ANTI (60% of trials). In these trials, participants have to react to five arrows presented on the screen by pushing either the left or right button according to the direction pointed by a central arrow (target), while ignoring the arrows horizontally placed at the sides. In half of these trials, the central arrow points in the same direction as the side arrows (congruent) and in the other half, the central arrow points in the opposite direction to the side arrows (incongruent). In two‐thirds of the trials, arrows are preceded by a visual cue presented 100 ms beforehand, either above or below the fixation point, in the same (valid cue) or in the opposite (invalid cue) location as the central arrow. In the remaining third of the trials, arrows are preceded by a fixation point, without any additional cue (no cue). In half of all the trials, an auditory warning signal (tone) appears 500 ms before the arrows presentation. In the other half, no signal is presented (no tone).

2. Executive Vigilance (20% of trials). These trials are similar to the ANTI trials, but the central arrow appears vertically displaced (±8 px from the side arrows) for participants to detect and respond to this change by pressing the spacebar.

3. Arousal Vigilance (20% of trials). During these trials, a red millisecond down counter is displayed on the screen. Participants have to stop it as quickly as possible by pressing any key on the keyboard.

FIGURE 1.

The Attentional Networks Test for Interactions and Vigilance – executive and arousal components (ANTI‐Vea). Source: https://anti‐vea.ugr.es/method.htmL; published with the permission of the authors.

In both ANTI and executive vigilance trials, each of the five arrows is horizontally and/or vertically displaced at random up to ±2 px from its original position to generate some noise. Consequently, difficult ANTI trials are those ANTI trials with a vertical displacement >2 px from the central arrow to at least one of its two adjacent arrows.

Measures in the ANTI‐Vea

The attentional network functions – alertness, orienting and executive control – were measured by computing the differences in mean RTs and error rates (wrong button presses) between two trial types: tone vs. no tone (only for no cue trials) for alertness index; invalid cue vs. valid cue for orienting index; incongruent vs. congruent for executive control index. Mean RTs were measured only for correct button presses that were faster than 1500 ms, but slower than 200 ms, to exclude potential lapses and anticipated reactions accordingly.

Executive vigilance was measured with the following indices: the percentage of correct reactions (hits) in executive vigilance trials; the percentage of false alarms (FAs; i.e. spacebar presses in difficult ANTI trials); response sensitivity (A′) and bias (B″) calculated according to the formulas specified by Stanislaw and Todorov (1999) based on signal detection theory.

Arousal vigilance was measured with the following indices calculated only in arousal vigilance trials: mean RT, SD of RT and lapses (the percentage of trials with RTs longer than 600 ms or missed responses). Apart from SD of RT, We also included another measure of RTV: coefficient of variation (CV), calculated by dividing each participant's SD of RTs by their mean RT.

Executive and arousal vigilance decrement was measured as the change in each of the above‐mentioned indices across blocks, expressed as a linear slope. While indices of executive vigilance tend to diminish throughout the task (negative slope; with the exception of response criterion, which tends to rise), arousal vigilance indices tend to increase (positive slope).

Attention span was computed based on Simon et al. (2023) as the mean number of consecutive correct responses between 200 ms and 2 standard deviations (for ANTI and executive vigilance trials) or 1 standard deviation (for arousal vigilance trials) above participants' mean RT.

Questionnaires

The Cronbach′ α′ values presented below are computed on data collected in the presented study.

Pure Procrastination Scale (PPS; Steel, ²⁰¹⁰; Polish adaptation by Stępień & Topolewska, ²⁰¹⁴), which consists of 12 items measuring a tendency to procrastinate (e.g. I often find myself performing tasks that I had intended to do days before; I am continually saying “I'll do it tomorrow”). The participants respond on a scale from 1 (describes me completely inaccurately) to 5 (describes me completely accurately). The scale showed high internal consistency (Cronbach ɑ = .92).

Difficulties in Emotion Regulation Scale (DERS; Gratz & Roemer, ²⁰⁰⁴; Polish version by Dragan, ²⁰²⁰), which includes 36 items divided into six subscales: Nonacceptance of Emotional Responses (e.g. When I'm upset, I become embarrassed for feeling that way); Difficulties Engaging in Goal Directed Behaviour (e.g. When I'm upset, I have difficulty getting work done); Impulse Control Difficulties (e.g. When I'm upset, I lose control over my behaviour); Lack of Emotional Awareness (e.g. I pay attention to how I feel – reversed item); Limited Access to Emotion Regulation Strategies (e.g. When I'm upset, I believe there is nothing I can do to make myself feel better); and Lack of Emotional Clarity (e.g. I am confused about how I feel). Participants indicate the frequency of described behaviours and feelings on a scale from 1 (0–10% – almost never) to 5 (91–100% – always). In the present study, only the sum score was included in the analyses. The scale showed high internal consistency (Cronbach's ɑ = .92).

Spontaneous Mind‐Wandering Scale (MW‐S; Carriere et al., ²⁰¹³), which was translated into Polish using the back‐translation procedure. The scale contains 4 items (e.g. I mind‐wander even when I'm supposed to be doing something else) measuring the frequency of experiencing spontaneous mind‐wandering on a daily basis. Participants answer on a 7‐point Likert scale ranging from 1 (very rarely) to 7 (very frequently). The scale showed high internal consistency (Cronbach's ɑ = .83).

Short version of the Dundee Stress State Questionnaire (DSSQ; Matthews & Zeidner, ²⁰¹²; Polish version by Zajenkowski & Zajenkowska, ²⁰¹⁵) in two forms: the first one was completed right before the beginning of the task, measuring participant's current state; the second one was filled out right after the ANTI‐Vea completion, retrospectively evaluating the subjective experiences related to participant's task performance. Both versions include 24 analogous items clustered into three subscales: Engagement (e.g. I was determined to succeed on the task), Worry (e.g. I felt concerned about the impression I was making.) and Distress (e.g. I felt that I could not deal with the situation effectively). The wording of the two versions is adjusted to pre‐ and post‐task measurements (e.g. I feel confident about my performance vs. I felt confident about my performance). Participants answer on a 5‐point Likert scale ranging from 0 (definitely false) to 4 (definitely true). In the previous studies (Zajenkowski et al., 2016; Zajenkowski & Zajenkowska, ²⁰¹⁵), the Polish version of DSSQ achieved high internal consistency (Cronbach's ɑ for Engagement = .80; Distress ɑ = .76; Worry ɑ = .84). However, in this study we obtained lower values of Cronbach's alpha (Engagement ɑ = .38; Distress ɑ = .67; Worry ɑ = .83 for pre‐task assessment; Engagement ɑ = .73; Distress ɑ = .55; Worry ɑ = .83 for post‐task assessment).

Statistical analyses

The analyses were performed in IBM SPSS Statistics 27, R Statistical Software (v3.6.1.; R Core Team, ²⁰¹⁹) and MATLAB (The Mathworks, Inc., Natick, MA, USA). Pearson correlation coefficients were computed between the scores in the procrastination scale (PPS) and the results of other questionnaires as well as the above‐mentioned indices in the ANTI‐Vea. To control for multiple comparisons, false discovery rate (FDR; Benjamini & Hochberg, 1995) correction was applied for correlation analyses. However, as our analyses were preregistered and based on specific hypotheses, we also present and cautiously interpret the uncorrected results in order to minimize the risk of Type II error. Both results are presented in the text and in the correlation tables, with results that remained significant after correction marked in bold.

To identify the mediating roles of emotion dysregulation and spontaneous mind‐wandering in a relationship between procrastination and vigilance, a series of percentile bootstrap mediation tests were performed with the use of PROCESS macro, model 4 (implemented via R script; Hayes, ²⁰²²), using 10,000 bootstrap samples and a 95% confidence interval for indirect effects. Procrastination score was an outcome variable, and executive and arousal vigilance indices were independent variables. Scores in DERS and MW‐S were included as mediators.

For the analysis of the relationship between procrastination and a change in task‐related worry, distress and engagement, repeated measures analyses of covariance (ANCOVAs) were performed with the time point (a score in the pre‐ vs. post‐task version of the DSSQ) as the within‐subject variable and procrastination score (PPS) as the covariate. The analyses were conducted separately for each of the subscales in DSSQ.

Transparency and openness

The collected data as well as the scripts used for analyses in the presented study were published in the public repository (https://osf.io/67kms/), in accordance with FAIR Data principles (Wilkinson et al., 2016). The preregistration of the study can be found at https://osf.io/7m2au.

We have made the following changes with regard to the preregistration:

1. Inclusion of three additional hypotheses (see the Introduction section), which were not initially included in the preregistration, as the primary focus of this research was on the relationship between procrastination and vigilance. Also, the attention span metric was not published at the time of the preregistration.

2. Exclusion of participants with a hit rate lower than 60% in ANTI trials in at least one of the ANTI‐Vea blocks. We observed that some participants obtained relatively high accuracy in the whole task, except for one or two blocks. This might have resulted from a temporary disengagement from the task, or from a temporal network disconnection, and because of that, we decided to introduce this additional exclusion criterion.

3. Calculation of CV as another measure of RTV in arousal vigilance trials. We decided to add this index, as it was measured in previous studies on procrastination‐related attentional deficits (e.g. Michałowski et al., ²⁰²⁰; Wiwatowska, Pietruch, et al., ²⁰²³; Wiwatowska, Wypych, et al., ²⁰²³). Also, CV is less dependent on mean RT than SD of RT (Wagenmakers & Brown, 2007). However, since CV was included after the preregistration, these analyses should be treated as explanatory and interpreted with proper caution.

4. Applying FDR correction instead of Bonferroni to correlation analyses. After publishing the preregistration we realized that Bonferroni correction might be too conservative, given the relatively large number of examined indices and their statistical dependence (Blakesley et al., 2009). This approach would require a substantial reduction in alpha level, thereby increasing the risk of Type II errors and potentially obscuring meaningful effects. Instead, we applied the FDR correction (Benjamini & Hochberg, 1995), which provides a more balanced approach by controlling for false positives with lower statistical power loss. However, given that our analyses were guided by preregistered, theory‐driven hypotheses, we decided to additionally present and exploratorily interpret significant results which did not survive the correction.

RESULTS

Correlations between procrastination and vigilance ( H1 )

The correlation coefficients can be found in Table 1. Procrastination correlated negatively with two out of four indices of executive vigilance: hit rate (r = −.158; p = .024) and A' (r = −.174; p = .012), although the correlation with hit rate did not survive the FDR correction (p _FDR > .05). There were no significant correlations with the arousal vigilance indices (ps > .05).

TABLE 1.

Pearson correlation coefficients between vigilance indices and questionnaire results.

	M	SD	1	2	3	4	5	6	7	8	9	10	11
1. EV: Hits	70.89	19.19	–
2. EV: FAs	6.33	6.28	.403 ***	–
3. EV: A'	0.90	0.05	.900 ***	−.011	–
4. EV: B″	0.51	0.40	−.587 ***	−.886 ***	−.240 ***	–
5. AV: RT	514.23	61.92	−.216 **	−.142*	−.165*	.169*	–
6. AV: SD of RT	92.44	38.20	−.232 ***	−.064	−.212 **	.141*	.635 ***	–
7. AV: CV	0.18	0.06	−.196 **	−.022	−.192 **	0.105	.396 ***	.954 ***	–
8. AV: Lapses	14.92	15.45	−.189 **	−.129	−.137*	.151*	.903 ***	.639 ***	.438 ***	–
9. PPS	33.72	10.64	−.158*	−.047	−.174 *	.069	.061	.072	.059	.041	–
10. DERS	83.60	20.68	−.175 *	−.048	−.176 *	.110	.021	.173 *	.192 **	.023	.527 ***	–
11. MW‐S	16.03	6.15	−.230 ***	−.029	−.240 ***	.150*	.048	.089	.081	.067	.435 ***	.437 ***	–

Note: Correlations highlighted in bold are those that remained significant at a p‐value of < .05 after adjusting for multiple comparisons.

Abbreviations: AV, arousal vigilance indices; DERS, Difficulties in Emotion Regulation Scale; EV, executive vigilance indices; FA, false alarms; MW‐S, Spontaneous Mind Wandering Scale; PPS, Pure Procrastination Scale; RT, reaction time.

^* p < .05.

^** p < .01.

^*** p < .001.

Correlations between procrastination and vigilance decrement ( H2 )

The correlation coefficients can be found in Table 2. Procrastination was significantly related to the slope of CV (r = .138; p = .048), although this result did not survive the FDR correction (p _FDR > .05). Other indices of vigilance decrement were not significantly related to procrastination (ps > .05).

TABLE 2.

Pearson correlation coefficients between vigilance decrement indices (slopes) and questionnaire results.

	M	SD	1	2	3	4	5	6	7	8	9	10	11
1. EV: Hit rate	−2.73	4.07	–
2. EV: FAs	−0.37	1.69	.014	–
3. EV: A'	−0.01	0.02	.858 ***	−.356 ***	–
4. EV: B″	−0.04	0.12	−.210 **	−.546 ***	.050	–
5. AV: RT	5.24	12.50	.049	.065	.045	−.152*	–
6. AV: SD of RT	1.60	3.90	−.024	−.041	.023	.006	.623 ***	–
7. AV: CV	0.01	0.02	−.038	−.065	.021	.053	.451 ***	.966 ***	–
8. AV: Lapses	6.42	12.87	.035	.030	.057	−.140*	.851 ***	.517 ***	.393 ***	–
9. PPS	33.72	10.64	−.051	−.034	−.055	−.009	−.005	.120	.138*	.008	–
10. DERS	83.60	20.68	.022	.051	−.021	−.012	.184 **	.252 ***	.232 ***	.166*	.527 ***	–
11. MW‐S	16.03	6.15	−.129	.019	−.169 *	.058	.056	.153*	.165*	.062	.435 ***	.437 ***	–

Note: Correlations highlighted in bold are those that remained significant at a p‐value of < .05 after adjusting for multiple comparisons.

Abbreviations: AV, arousal vigilance decrement indices; DERS, Difficulties in Emotion Regulation Scale; EV, executive vigilance decrement indices; FA, false alarms; MW‐S, Spontaneous Mind Wandering Scale; PPS, Pure Procrastination Scale; RT, reaction time.

^* p < .05.

^** p < .01.

^*** p < .001.

Correlations between procrastination and attentional network functions and attention span (npH1 & npH2)

Procrastination did not correlate significantly with any of the attentional network functions: alerting, orienting and executive control (ps > .05; see Table 3). However, higher attention span was significantly related to lower procrastination (r = −.143; p = .040), although this result did not survive the FDR correction (p _FDR > .05).

TABLE 3.

Pearson correlation coefficients between attentional networks, attention span and procrastination.

	M	SD	1	2	3	4	5	6	7	8	9	10
1. Alerting – RT	42.85	33.48	–
2. Alerting – errors	0.84	2.47	−.070	–
3. Orienting – RT	41.25	23.40	.066	−.049	–
4. Orienting – errors	0.79	3.46	.100	−.060	.192 **	–
5. Congruency – RT	52.88	27.28	.131	−.008	.023	−.053	–
6. Congruency – errors	2.04	6.37	−.084	−.102	−.108	.119	.228 ***	–
7. Attention span	6.81	1.92	−.019	.086	.036	−.270 ***	−.069	−.271 ***	–
8. PPS	33.72	10.64	−.118	.002	−.057	−.047	−.020	.098	−.143*	–
9. DERS	83.60	20.68	−.030	.013	−.102	−.087	.048	.130	−.152*	.527 ***	–
10. MW‐S	16.03	6.15	−.063	.032	−.128	−.033	−.054	.040	−.191 **	.435 ***	.437 ***	–

Note: Correlations highlighted in bold are those that remained significant at a p‐value of < .05 after adjusting for multiple comparisons.

Abbreviations: DERS, Difficulties in Emotion Regulation Scale; MW‐S, Spontaneous Mind Wandering Scale; PPS, Pure Procrastination Scale; RT, reaction time.

^* p < .05.

^** p < .01.

^*** p < .00.

Mediation analyses ( H3 & npH3)

Mediation analyses were conducted only for task‐related variables that significantly correlated with procrastination, namely: hit rate and A' (executive vigilance), slope of CV (arousal vigilance decrement) and attention span. The standardized coefficients are presented in Figures 2 and 3.

FIGURE 2.

The results of mediation analyses with emotion dysregulation as a mediator in the relationships between procrastination and executive vigilance indices (a&b), arousal vigilance decrement (c) and attention span (d). All indirect effects are significant (p < .05). *p < .05; **p < .01; ***p < .001.

FIGURE 3.

The results of mediation analyses with spontaneous mind‐wandering as a mediator in the relationships between procrastination and executive vigilance indices (a, b), arousal vigilance decrement (c) and attention span (d). *p < .05; **p < .01; ***p < .001.

Emotion dysregulation fully mediated the relationship between procrastination and executive vigilance indices (i.e. hit rate and A′; see Figure 2a,b), arousal vigilance decrement index (i.e. slope of CV; see Figure 2c) and attention span (see Figure 2d).

Mediation analyses with spontaneous mind‐wandering as a mediator yielded similar results: the score in MW‐S fully mediated the relationship between procrastination and all of the above‐mentioned indices (see Figure 3).

Relationship between procrastination and change in task‐related engagement, distress and worry (H4)

Correlation analyses revealed that procrastination was related to lower task engagement as well as increased task‐related distress and worry. These associations were significant with all pre‐and post‐task measurements, although the correlation with post‐task engagement did not survive the FDR correction (Table 4). The results of repeated measures ANCOVA showed that after (vs. before) task completion participants reported higher distress (F(1, 204) = 4.90; p = .028; η ² _p  = .023 for the main effect of time), but decreased worry (F(1, 204) = 9.00; p = .003; η ² _p  = .042) and engagement (F(1, 204) = 23.11; p < .001; η ² _p  = .102). However, there were no significant interactions between procrastination and the time of measurement (Fs <2; ps > .1; see Figure 4).

TABLE 4.

Pearson correlation coefficients between procrastination and pre‐ & post‐task‐related engagement, distress and worry.

	M	SD	1	2	3	4	5	6	7
1. PPS	33.72	10.64	–
2. DSSQ pre‐task engagement	3.37	4.95	−.185 **	–
3. DSSQ pre‐task distress	−3.19	5.52	.417 ***	−.361 ***	–
4. DSSQ pre‐task worry	22.00	6.78	.177 *	−.208 **	.435 ***	–
5. DSSQ post‐task engagement	−2.48	5.26	−.163*	.443 ***	−.076	−.136	–
6. DSSQ post‐task distress	−1.08	5.77	.357 ***	−.189 **	.509 ***	.163*	−.274 ***	–
7. DSSQ post‐task worry	18.73	6.52	.271 ***	−.057	.258 ***	.402 ***	−.088	.198 **	–

Note: Correlations highlighted in bold are those that remained significant at a p‐value of < .05 after adjusting for multiple comparisons.

Abbreviations: DSSQ, Dundee Stress State Questionnaire; PPS, Pure Procrastination Scale.

^* p < .05.

^** p < .01.

^*** p < .001.

FIGURE 4.

Task‐related engagement (a), distress (b) and worry (c) measured before and after the completion of the attentional task in relation to the trait procrastination score (three groups were differentiated only for visualization purposes).

DISCUSSION

In this study, we aimed to investigate the link between dispositional procrastination among higher education students and their capacity for maintaining sustained attention as well as the mediating role of emotion dysregulation and propensity to experience spontaneous mind‐wandering in this relationship. We predicted that a higher tendency to procrastinate would be related to decreased executive and arousal vigilance as well as to higher vigilance decrement over time. We also expected that this link would be mediated by difficulties in regulating emotions and by more frequent experiences of spontaneous mind‐wandering episodes. The results partially confirmed the preregistered hypotheses, indicating that some of the executive vigilance indices were significantly linked with a tendency to procrastinate, while there were no significant correlations with arousal vigilance. However, concerning vigilance decrement, only one index of arousal time‐related deterioration was related to procrastination, but there were no associations with executive vigilance decrements. Crucially, the relationships between procrastination and task‐related indices were fully mediated by declared difficulties in emotion regulation and a tendency to experience spontaneous mind‐wandering, both of which were related to dispositional procrastination.

One of the indices that negatively correlated with procrastination was the percentage of hits in executive vigilance trials. This was probably due to lower ability to detect and properly react to rare stimuli, taking into account that procrastination was also associated with sensitivity index (A′), but not with false alarm rate or criterion bias (B″). This finding goes in line with previous studies showing a negative relationship between trait procrastination and sustained attention (Michałowski et al., 2020). Moreover, a recent study (Wiwatowska et al., 2022) showed that high procrastination is related to lower proactive control, which is responsible for maintaining and actively updating goal‐related information in one's mind. Therefore, it might be more difficult for highly procrastinating individuals to maintain the goal of detecting rare stimuli, which can lead to a lower rate of correct responses. However, a tendency to delay tasks was not associated with a time‐related decrease in executive vigilance. This finding might be explained by the lack of significant relationship between procrastination and executive control, which seems to modulate executive vigilance decrement (Luna et al., 2022). Although nonsignificant correlation between tendency to procrastinate and executive control is in contrast to previous questionnaire studies (e.g. Rabin et al., ²⁰¹¹), it goes in line with other research using more objective measures. For example, a recent study found no differences between high and low procrastinating participants in behavioural or neural reactions to distractors (Wiwatowska, Pietruch, et al., 2023) or to distinct stimuli that required withdrawal of motor reactions (Michałowski et al., 2020). Therefore, it seems that dispositional procrastination is not related to problems with inhibitory/executive control, as it has been previously suggested.

Another finding, which was contrary to our hypothesis, was that procrastination was not related to measures of overall arousal vigilance (i.e. RT, RTV and the percentage of lapses – slow and missed reactions). This observation is inconsistent with the results of previous research showing that a high tendency to procrastinate is linked with increased RTV (e.g. Michałowski et al., ²⁰²⁰; Wiwatowska et al., ²⁰²²). The potential reason for this discrepancy is that in these past studies RTV was measured in response to repetitive stimuli, which required top‐down control. In ANTI‐Vea, arousal vigilance is measured based on reactions to rare stimuli, which might attract attention in a bottom‐up manner, as they are presented in the form of a red down counter. Therefore, the procrastination‐related increase in RTV found in previous research might have reflected lower top‐down attentional control, rather than decreased arousal vigilance.

On the other hand, in the presented study we observed that a higher tendency to procrastinate was linked to a bigger increase in RTV with task progression, but there were no correlations with other measures of arousal vigilance decrement. This suggests that trait procrastination is probably not related to progressive loss in tonic alertness, which would be manifested in the elongation of RTs and a higher rate of lapses. Instead, with more time on task, highly procrastinating individuals might find it more difficult to maintain motivation and steady focus on the task, which can lead to more variable performance. Moreover, additional analyses showed a negative correlation between a tendency to procrastinate and the attention span index – the mean number of consecutive trials with optimal performance. This result further supports the notion of difficulties in maintaining attention on tasks for longer periods of time among highly procrastinating individuals.

The observed relationships between dispositional procrastination, vigilance scores and attention span were fully mediated by emotion dysregulation as well as a tendency to experience spontaneous mind‐wandering. Lower attentional control might hinder maintaining focus on task‐related goals and ignoring thoughts related to alternative goals, current concerns, or negative emotional reactions evoked by the tasks. This attentional reorientation could facilitate task avoidance and engagement in other, less aversive activities (although it is important to point out the correlational nature of our findings, which do not allow for inferring causation). Therefore, our study integrates two important views on procrastination: emphasizing the role of low attentional control (Steel, 2007; Steel et al., ²⁰¹⁸) as well as poor emotion regulation (Sirois & Pychyl, 2013; Wypych & Potenza, ²⁰²¹). Obtained findings are particularly relevant for educational settings, as academic tasks usually require focused attention and proper self‐regulation, given that they are often perceived as effortful and performed in distraction‐rich environments. Consistent with this notion, our results align with evidence suggesting that chronic procrastination among students could be mitigated by training in emotion regulation (Schuenemann et al., 2022) and mindfulness (Rad et al., 2023) – the latter being associated with reduced spontaneous mind‐wandering (Cásedas et al., 2023; Mrazek et al., ²⁰¹³).

Moreover, in the present study, we observed that an increased tendency to procrastinate is related to lower task engagement as well as higher task‐related distress and worry. It means that tasks requiring attentional control might be more aversive for highly procrastinating students. This aversion might be rooted in past negative experiences (e.g. failures or low grades) associated with performing this type of tasks, which can be more difficult to complete for individuals with attentional dysfunctions. Further, dealing with these aversive reactions might be compromised among procrastinators, given their difficulties in emotion regulation observed in this and previous studies (e.g. Wypych et al., ²⁰¹⁸). This might create a vicious circle of attentional control difficulties, negative aversive reactions and task delay. However, contrary to our expectations, a tendency to procrastinate was not associated with the change (pre‐ vs. post‐task) in declared engagement, worry and distress. This suggests that procrastination‐related aversive reactions towards the task do not escalate during task completion. Hence, it appears that mere expectations of facing difficulties or boredom might reduce motivation and encourage avoidance behaviour. Nonetheless, these results should be interpreted with caution, as some subscales of DSSQ (i.e. Engagement and Distress) showed low internal consistency in the presented study, which means that they may not measure homogeneous constructs. Moreover, this questionnaire assesses task‐specific reactions, meaning the results could differ if participants were asked to complete a different type of attentionally demanding activity or if procrastination was measured as a task‐specific behaviour rather than a trait.

LIMITATIONS AND FUTURE DIRECTIONS

There are several limitations of the presented research that should be addressed in future studies. Firstly, it should be pointed out that after applying FDR correction to control for multiple comparisons in our analysis, some of the correlations initially observed were no longer significant, although it could have resulted from an insufficient number of participants or the fact that we put all variables into one correlation matrix. Moreover, some of the relationships observed in this study (i.e. the correlations between procrastination and A' as well as the slope of CV) did not reach significance in an ongoing project in which similar measures were used (Aguirre et al., 2024). These discrepancies might result from differences in sample size (larger in the presented study), experimental procedure (in Aguirre et al., ²⁰²⁴, the ANTI‐Vea was slightly modified), or participants' nationality (Polish vs. Spanish), but there is also a possibility that some of the findings from this study were obtained by chance and it should be taken into account before drawing any strong conclusions. Nevertheless, Aguirre and collaborators' study confirmed that the tendency to procrastinate correlates with attention span index as well as with hit rate in the executive vigilance trials. These results strengthen the evidence for the existence of the link between a tendency to delay tasks and difficulties in some aspects of attentional control.

Secondly, the recruited participants were from different universities, and some (but not all) received additional credit for their participation, which may have introduced heterogeneity into the collected data. Future studies could address this issue by collecting a larger and more balanced sample, enabling additional analyses to test the generalizability of our findings both across and beyond participants' origin and type of incentive offered in the study.

Thirdly, it is important to emphasize the cross‐sectional nature of the observed relationships, which does not allow for concluding whether vigilance or attentional control directly influence a tendency to delay tasks. Nevertheless, we believe that the theory and research presented in the Introduction (supporting a potential causal relationship between the variables of interest) provide a solid foundation for the mediation models proposed in this paper. Although some authors caution against using cross‐sectional data in mediation analyses (Maxwell et al., 2011), others argue that in some circumstances, it is a valid approach (Hayes, 2022) and that observational data can indeed be utilized for causal inference (Bailey et al., 2024; Pearl & Mackenzie, ²⁰¹⁸). While longitudinal studies are recommended to better capture the temporal dynamics and potential causal pathways (Maxwell & Cole, 2007), they alone may not conclusively confirm causality. This could be established in future studies by investigating the impact of attentional training or stimulation on the frequency of procrastinatory behaviours. These interventions should incorporate measures of emotional dysregulation and spontaneous mind‐wandering in order to test our mediational models in a design that enables drawing more robust causal inferences. It would also be beneficial to experimentally verify if states of lower attentional control could explain specific instances of procrastination by inducing task‐unrelated thoughts or difficulties in regulating emotions evoked in a particular situation.

Lastly, it should be acknowledged that the observed correlations between questionnaire results and task‐related indices were low, which does not align with the findings of past studies that used self‐report measures of attentional control and found much larger correlations with procrastination (e.g. Steel, ²⁰⁰⁷; Steel et al., ²⁰¹⁸). Although these previous results might have reflected preconceptions about one's functioning instead of real cognitive capacities (Quigley et al., 2017), smaller effect sizes might also stem from lower reliability typically observed in indices of cognitive‐behavioural tasks (Dang et al., 2020). In fact, reports on the reliability scores of the ANTI‐Vea in previous studies (Cásedas et al., 2022; Coll‐Martín et al., ²⁰²¹; Luna, Roca, et al., ²⁰²¹) have been particularly low for indices of vigilance decrement, precisely the ones for which almost no relationships were observed. Also, while previous research has shown acceptable test–retest stability of some of the executive and arousal vigilance measures, studies using the Psychomotor Vigilance Task have shown quite low test–retest correlation coefficients for arousal vigilance decrement (Luna, Barttfeld, et al., 2021; Langner et al., ²⁰²³; Thompson et al., ²⁰²²). However, this may indicate poor measurement precision (which presumably would increase with more trials), but not necessarily a lack of stability of the underlying construct. In any case, future studies should address this problem by collecting larger sample sizes to compensate for the loss of statistical power in correlations that involve indices with poor internal consistency (e.g. Coll‐Martín et al., ²⁰²⁴). The ease of the ANTI‐Vea platform to measure different attentional components online and remotely makes this large‐scale data collection reasonably feasible.

CONCLUSIONS

The presented study provided novel findings on the association between dispositional procrastination and attentional control components as well as on the mediating role of emotion dysregulation and spontaneous mind‐wandering in these relationships. Obtained results allow for a better understanding of a tendency to procrastinate among higher education students. The conclusions from this study might also contribute to the development of therapeutic strategies aimed at mitigating this problem.

AUTHOR CONTRIBUTIONS

Ewa Wiwatowska: Funding acquisition; project administration; conceptualization; data curation; writing – original draft; writing – review and editing; investigation; methodology. Magdalena Prost: Investigation; writing – review and editing; methodology. Tao Coll‐Martin: Formal analysis; writing – review and editing; visualization; methodology. Juan Lupiáñez: Conceptualization; supervision; writing – review and editing; methodology.

CONFLICT OF INTEREST STATEMENT

The authors report there are no competing interests to declare.

Wiwatowska, E. , Prost, M. , Coll‐Martin, T. , & Lupiáñez, J. (2025). Is poor control over thoughts and emotions related to a higher tendency to delay tasks? The link between procrastination, emotional dysregulation and attentional control. British Journal of Psychology, 116, 807–830. 10.1111/bjop.12793

DATA AVAILABILITY STATEMENT

The data and code are available to download from a public repository at the following link: https://osf.io/67kms/