Sense of control buffers against stress

  1. Jennifer C Fielder
  2. Jinyu Shi
  3. Daniel McGlade
  4. Quentin JM Huys
  5. Nikolaus Steinbeis

  • Curated by eLife

    eLife logo

    eLife Assessment

    This important research addresses the effects of subjective control and task difficulty on experienced stress using a novel behavioral task in two, large online samples. Convincing evidence is provided, establishing internal and external task validity and a relationship with individual differences in relevant mental health constructs. Evidence for the core claims could be strengthened by disentangling the effects of controllability from those of reward rate and adjusting data parcellation for computing internal consistency. This work will be of interest to psychologists and clinicians studying controllability, stress, and psychopathology.

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Stress is one of the most pervasive causes of mental ill-health across the lifespan. Subjective dimensions of stress perception, such as perceived control, are especially potent in shaping stress responses. While the impact of reduced or no control over stress is well understood, much less is known about whether heightened feelings of control buffer against the negative impact of later stress. We designed a novel paradigm with excellent psychometric properties to sensitively capture and induce different states of subjective control. Across two studies with a total of 768 neurotypical adults, we show a robust association between sense of control and stress as well as symptoms of mental ill-health. More importantly, in a subsample of 295 participants we show that compared to a neutral control group, inducing a heightened state of subjective control buffers against the impact of later stress. These findings demonstrate a causal role for a heightened sense of control in mitigating the negative impact of stressful experiences and spell out important directions for future preventative interventions.

Article activity feed

  1. Version published to 10.7554/elife.105025.1 on eLife
  2. Version published to 10.7554/elife.105025 on eLife
  3. eLife

    eLife Assessment

    This important research addresses the effects of subjective control and task difficulty on experienced stress using a novel behavioral task in two, large online samples. Convincing evidence is provided, establishing internal and external task validity and a relationship with individual differences in relevant mental health constructs. Evidence for the core claims could be strengthened by disentangling the effects of controllability from those of reward rate and adjusting data parcellation for computing internal consistency. This work will be of interest to psychologists and clinicians studying controllability, stress, and psychopathology.

  4. eLife

    Reviewer #1 (Public review):

    Summary:

    This work investigated how the sense of control influences perceptions of stress. In a novel "Wheel Stopping" task, the authors used task variations in difficulty and controllability to measure and manipulate perceived control in two large cohorts of online participants. The authors first show that their behavioral task has good internal consistency and external validity, showing that perceived control during the task was linked to relevant measures of anxiety, depression, and locus of control. Most importantly, manipulating controllability in the task led to reduced subjective stress, showing a direct impact of control on stress perception. However, this work has minor limitations due to the design of the stressor manipulations/measurements and the necessary logistics associated with online versus in-person stress studies.

    Nevertheless, this research adds to our understanding of when and how control can influence the effects of stress and is particularly relevant to mental health interventions.

    Strengths:

    The primary strength of this research is the development of a unique and clever task design that can reliably and validly elicit variations in beliefs about control. Impressively, higher subjective control in the task was associated with decreased psychopathology measures such an anxiety and depression in a non-clinical sample of participants. In addition, the authors found that lower control and higher difficulty in the task led to higher perceived stress, suggesting that the task can reliably manipulate perceptions of stress. Prior tasks have not included both controllability and difficulty in this manner and have not directly tested the direct influence of these factors on incidental stress, making this work both novel and important for the field.

    Weaknesses:

    One minor weakness of this research is the validity of the online stress measurements and manipulations. In this study, the authors measure subjective stress via self-report both during the task and also after either a Trier Social Stress Test (high-stress condition) or a memory test (low-stress condition). One concern is that these stress manipulations were really "threats" of stress, where participants never had to complete the stress tasks (i.e., recording a speech for judgment). While this is not unusual for an in-lab study and can reliably elicit substantial stress/anxiety, in an online study, there is a possibility for communication between participants (via online forums dedicated to such communication), which could weaken the stress effects. That said, the authors did find sensible increases and decreases of perceived stress between relevant time points, but future work could improve upon this design by including more complete stress manipulations and measuring implicit physiological signs of stress.

  5. eLife

    Reviewer #2 (Public review):

    Summary:

    The authors have developed a behavioral paradigm to experimentally manipulate the sense of control experienced by the participants by changing the level of difficulty of a wheel-stopping task. In the first study, this manipulation is tested by administering the task in a factorial design with two levels of controllability and two levels of stressor intensity to a large number of participants online while simultaneously recording subjective ratings on perceived control, anxiety, and stress. In the second study, the authors used the wheel-stopping task to induce a high sense of controllability and test whether this manipulation buffers the response to a subsequent stress induction when compared to a neutral task, like looking at pleasant videos.

    Strengths:

    (1) The authors validate a method to manipulate stress.
    (2) The authors use an experimental manipulation to induce an enhanced sense of controllability to test its impact on the response to stress induction.
    (3) The studies involved big sample sizes.

    Weaknesses:

    (1) The study was not preregistered.

    (2) The control manipulation is conflated with task difficulty, and, therefore the reward rate. Although the authors acknowledge this limitation at the end of the discussion, it is a very important limitation, and its implications are not properly discussed. The discussion states that this is a common limitation with previous studies of control but omits that many studies have controlled for it using yoking.

    (3) The methods are not always clear enough, and it is difficult to know whether all the manipulations are done within-subjects or some key manipulations are done between subjects.

    (4) The analysis of internal consistency is based on splitting the data into odd/even sliders. This choice of data parcellation may cause missed drifts in task performance due to learning, practice effects, or tiredness, thus potentially inflating internal consistency.

    (5) Study 2 manipulates the effect of domain (win versus loss WS task), but the interaction of this factor with stressor intensity is not included in the analysis.

    This study will be of interest to psychologists and cognitive scientists interested in understanding how controllability and its subjective perception impact how people respond to stress exposure. Demonstrating that an increased sense of control buffers/protects against subsequent stress is important and may trigger further studies to characterize this phenomenon better. However, beyond the highlighted weaknesses, the current study only studied the effect of stress induction consecutive to the performance of the WS task on the same day and its generalizability is not warranted.

  6. eLife

    Reviewer #3 (Public review):

    Summary:

    This is an interesting investigation of the benefits of perceiving control and its impact on the subjective experience of stress. To assess a subjective sense of control, the authors introduce a novel wheel-stopping (WS) task where control is manipulated via size and speed to induce low and high control conditions. The authors demonstrate that the subjective sense of control is associated with experienced subjective stress and individual differences related to mental health measures. In a second experiment, they further show that an increased sense of control buffers subjective stress induced by a trier social stress manipulation, more so than a more typical stress buffering mechanism of watching neutral/calming videos.

    Strengths:

    There are several strengths to the manuscript that can be highlighted. For instance, the paper introduces a new paradigm and a clever manipulation to test an important and significant question. Additionally, it is a well-powered investigation that allows for confidence in replicability and the ability to show both high internal consistency and high external validity with an interesting set of individual difference analyses. Finally, the results are quite interesting and support prior literature while also providing a significant contribution to the field with respect to understanding the benefits of perceiving control.

    Weaknesses:

    There are also some questions that, if addressed, could help our readership.

    (1) A key manipulation was the high-intensity stressor (Anticipatory TSST signal), which was measured via subjective ratings recorded on a sliding scale at different intervals during testing. Typically, the TSST conducted in the lab is associated with increases in cortisol assessments and physiological responses (e.g., skin conductance and heart rate). The current study is limited to subjective measures of stress, given the online nature of the study. Since TSST online may also yield psychologically different results than in the lab (i.e., presumably in a comfortable environment, not facing a panel of judges), it would be helpful for the authors to briefly discuss how the subjective results compare with other examples from the literature (either online or in the lab). The question is whether the experienced stress was sufficiently stressful given that it was online and measured via subjective reports. The control condition (low intensity via reading recipes) is helpful, but the low-intensity stress does not seem to differ from baseline readings at the beginning of the experiment.

    (2) The neutral videos represent an important condition to contrast with WS, but it raises two questions. First, the conditions are quite different in terms of experience, and it is interesting to consider what another more active (but not controlled per se) condition would be in comparison to the WS performance. That is, there is no instrumental action during the neutral video viewing (even passive ratings about the video), and the active demands could be an important component of the ability to mitigate stress. Second, the subjective ratings of the stress of the neutral video appear equivalent to the win condition. Would it have been useful to have a high arousal video (akin to the loss condition) to test the idea that experience of control will buffer against stress? That way, the subjective stress experience of stress would start at equivalent points after WS3.

    (3) For the stress relief analysis, the authors included time points 2 and 3 (after the stressor and debrief) but not a baseline reading before stress. Given the potential baseline differences across conditions, can this decision be justified in the manuscript?

    (4) Is the increased control experience during the losses condition more valuable in mitigating experienced stress than the win condition?

    (5) The subjective measure of control ("how in control do you feel right now") tends to follow a successful or failed attempt at the WS task. How much is the experience of control mediated by the degree of experienced success/schedule of reinforcement? Is it an assessment of control or, an evaluation of how well they are doing and/or resolution of uncertainty? An interesting paper by Cockburn et al. 2014 highlights the potential for positive prediction errors to enhance the desire for control.

    (6) While the authors do a very good job in their inclusion and synthesis of the relevant literature, they could also amplify some discussion in specific areas. For example, operationalizing task controllability via task difficulty is an interesting approach. It would be useful to discuss their approach (along with any others in the literature that have used it) and compare it to other typically used paradigms measuring control via presence or absence of choice, as mentioned by the authors briefly in the introduction.

    (7) The paper is well-written. However, it would be useful to expand on Figure 1 to include a) separate figures for study 1 (currently not included) and 2, and b) a timeline that includes the measurements of subjective stress (incorporated in Figure 1). It would also be helpful to include Figure S4 in the manuscript.

  7. eLife

    Author response:

    Conflation of control, difficulty and reward rate

    In response to the comment of control being conflated with task difficulty (and thus reward rate) that the reviewer feels is not adequately discussed in the paper, we will add more to this point in our discussion, especially in relation to previous literature. It is important to note, however, that our measure of perceived difficulty was included in analyses assessing the fluctuations in stress and control. Subjective control still had a unique effect on the experience of stress over and above perceived difficulty, suggesting that subjective control explains variance in stress beyond what is accounted for by perceived difficulty. We will also include additional analyses in which we include the win rate (i.e. percentage of all trials won) as a covariate when assessing the relationship between subjective control, perceived difficulty and subjective stress, which shows that win rate does not predict stress, but subjective control and perceived difficulty still uniquely predict subjective stress. The results of this will be added and elaborated further in the discussion.

    Neutral video condition

    In response to the comment of the neutral video condition not being active enough, we believe that any task with action-outcome contingencies would have a degree of controllability. To better distinguish experiences of control (WS task) to an experience of no/neutral control (i.e., neither high nor low controllability), we decided to use a task in which no actions were required during the task itself, although concentration was still required (attention checks regarding the content of the videos and ratings of the videos).

    The suggestion of having a high arousal video condition would indeed be interesting to test how experiencing ‘neutral’ control and high(er) stress levels preceding the stressor task influences stress buffering and stress relief. This is a good suggestion for future work that we can include in the discussion section.

    The TSST version (online and anticipatory)

    We will add more information regarding prior literature that the Trier Social Anticipatory Stress test has found physiological and psychological correlates (e.g. Nasso et al., 2019, Schlatter et al., 2021, Steinbeis et al., 2015), suggesting that the anticipation is still a valid stress manipulation despite participants not performing the actual speech task. Further, the TSST had a significant impact on subjective stress in the expected direction demonstrating that it was effective at eliciting subjective stress.

    Internal consistency

    We will parcellate the timepoints differently (not just odd/even sliders) to test the internal consistency, for example a random split or first half/second half.

    Effect of win-loss domain in Study 2

    We will run additional analyses testing the interaction of Domain (win or loss) with stressor intensity when predicting the stress buffering and stress relief effects. To test whether the loss domain is more valuable at mitigating experiences of stress than the win condition, we will run additional analyses with just the high control conditions (WS task) to test for a Domain*Time interaction, as we cannot test a Control*Domain*Time interaction in the full model given that we do not have ‘Domain’ for the video (neutral control) condition.

    Stress relief analyses

    Regarding the stress relief analyses (timepoints 2 and 3) and ‘baseline’ stress (timepoint 1), we will add to the manuscript that there is no significant difference in stress ratings between the high control and neutral control (collapsed across stress and domain) after the WS/video task, hence why we do not think it’s necessary to include in the stress relief model. Nevertheless, we will include a sensitivity analysis in the supplementary material to test the Timepoint*Control interaction (of stress relief – timepoints 2 and 3) when including timepoint 1 stress as a covariate.

    Clarity

    We will add more clarity in the methods section regarding within- and between-subject manipulations. We will also add Figure S4 to the main manuscript and expand Figure 1 to include both Studies 1 and 2 and a timeline of when subjective stress was assessed throughout the experiment.

  8. Version published to 10.1101/2024.12.05.626945v1 on bioRxiv