Experimental induction of state rumination: A study evaluating the efficacy of goal-cueing task in a sample of athletes

Abstract

Athletes’ behavior, thoughts, and feelings are influenced by whether athletes achieve their personal goals or not. The present experiment aimed to evaluate the application of a paradigm that uses unresolved goals to elicit state rumination in athletes. For this purpose, 101 athletes (f = 53, m= 48, M_age = 22.40, SD_age = 2.80) participated in an experiment and were divided into three different conditions – two experimental and one control condition. Using a 3 × 3 mixed ANOVA, we examined (1) whether we could significantly elicit rumination in athletes with a goal-related procedure (2) whether the context in which goals were formulated (general, sport-specific) mattered, and (3) whether the application of the procedure had an impact on performance as well as on cognitive processes in a subsequent go/no-go task. In addition, we assessed also athletes’ mood as well as their affect to account for emotional processes in addition to cognitive ones. Results showed that regardless of context, cueing unresolved goals resulted in a significant increase in state rumination after the paradigm compared to the control condition and subsequent recovery during the go/no-go task. In addition, only temporal changes in athletes’ mood and affect were evident; group differences were largely absent. Approaches for future research and further application in the sports context are discussed.

Keywords:

• Rumination
• goals
• athletes
• sports context
• mood
• affect

1. Introduction

Goals are ubiquitous in sports (Jeong et al., 2021), and can be defined in different ways. For example, a goal may be to improve one’s techniques or strategies (i.e., process goals), or to achieve a personal best (i.e., performance goals), a specific placement, or a standard (i.e., outcome goals) to qualify for a higher-level competition, a selection team, or for the squad in the particular discipline (Healy et al., 2018; Locke & Latham, 2019; Weinberg & Butt, 2014). These goals serve as the main motivation for athletes and structure their daily routine in the form of various training sessions. Depending on the individual goal, the athletes’ goal achievement process can extend over a varying period (e.g., several weeks, months or even years [Olympic cycle]), often placing enormous importance on the goal and its achievement compared to other life goals (Ellis, 2002). Moreover, goal setting and achievement is a complex process in which factors such as contextual and environmental factors, characteristics of goals being pursed, or personal differences with those individuals setting the goals always play an important role (Healy et al., 2018). Especially in competitive sports, goal attainment (or lack thereof) can be directly related to an athlete’s career success or failure (Williams, 2013), which in turn may have consequences on, for example, financial support, qualifications for higher-level competitions, or nomination to a squad. However, when goal pursuit does not proceed as desired, consequently a discrepancy between the desired goal and the current state is perceived. If we are unable to resolve this discrepancy through goal adjustment, alternative paths, or by discarding the goal, repetitive and negative thoughts may occur, a type of thinking that is referred to goal-directed rumination (Krys et al., 2020; Schultheiss et al., 2008). Goal-directed rumination is an everyday phenomenon, which can be described as an involuntary and unconscious coping process (Connor-Smith et al., 2000) as well as intrusive and perseverative thoughts about a goal-failure, a negative event, or a problem that is personally relevant and whose outcome is perceived as being uncontrollable (Brosschot et al., 2006; Martin & Tesser, 1996; Schultheiss et al., 2008). Moreover, rumination can affect an individual’s well-being, mood, perceived strain, or individual performance (Kircanski et al., 2018; Krys et al., 2020; Krys, 2020; Lyubomirsky et al., 2003; Moberly & Watkins, 2010). From a practical sports perspective, it is therefore important to understand (1) which thoughts are associated with the athletes’ individual goals in order to provide them with the best possible support in pursuing them. (2) Furthermore, it is essential for successful goal achievement that athletes are able to continue performing at their best after negative events (e.g., a missed qualification) or failures (e.g., a missed penalty kick) instead of getting lost in repetitive, negative, and intrusive thoughts. According to this, goal achievement is a powerful mechanism for guiding thoughts, feelings, and behaviors, which allows us to understand rumination in a more general context (Martin & Tesser, 1996). For this reason, we modified a goal-directed procedure – the so-called goal-cueing task (Roberts et al., 2013 see also Michel-Kröhler, Wessa & Berti, 2023) for application in the sports context that has the potential to identify the goals that may act as ruminative triggers for individuals, which is of value in examining both the causes and consequences of personally relevant goal-related rumination.

1.1. Rumination in the context of sports

Rumination and its consequences have attracted increasing theoretical and empirical interest in the past 30 years (Smith & Alloy, 2009; Whitmer & Gotlib, 2013). Although little research has been done on rumination in the context of sports, there is a growing interest. For example, Scott et al. (2002) reported negative correlations between different partial performances in tennis and rumination. Another study showed significant correlations between anger rumination—tendency to obsessively ruminate about experiences, accompanied by negative affect in the form of anger, and self-reported aggressive behavior in competitive athletes (Maxwell, 2004). Roy et al. (2016) investigated relations between rumination and performance in soccer and hockey players. The results suggested an expertise effect between athletes and non-athletes in terms of reflective rumination. In addition, the authors stated that lower rumination scores were associated with longer careers at a higher level in soccer players. Findings from another study demonstrated that ruminative thinking was related to state orientation, which is the individual’s tendency to stick on their own emotions and thoughts while performing under pressure compared to action orientation, which focuses on taking action (Kröhler & Berti, 2019). Furthermore, Tahtinen and colleagues (R. E. Tahtinen et al., 2021; R. Tahtinen et al., 2020) address the question of the role of rumination in relation to depressive symptoms in athletes. The results showed that brooding rumination predicted higher depressive symptoms scores over the study period. Moreover, Jansen et al. (2021) showed that female athletes had higher rumination scores than male athletes and individual athletes had higher values than athletes from team sports (see also Michel-Kröhler, Krys & Berti, 2023) and linked rumination to self-compassion and competitive anxiety (see also Casali et al., 2021). Finally, one study demonstrated that athletes who achieved their athletic goal at the end of the season reported less rumination compared to athletes who did not achieve their goals (Kröhler & Berti, 2017). To summarize, findings from previous studies underline that rumination can play a relevant role in the context of sport. However, the mentioned studies above were all based on a purely correlative design, which did not allow drawing causal conclusions. Accordingly, there is a need for experimental evidence on the various correlations with rumination to examine both the causes and consequences of rumination in the context of sport. One way to gain detailed insight into the effects of rumination as well as the mechanisms that cause these effects, is to induce ruminative thoughts within an experimental design so that a direct measurement of the potential effects of rumination on selected outcome variables is possible. However, to the best of our knowledge application of rumination induction in a sport specific context is lacking.

1.2. Application of the modified goal-cueing task in the context of sport

In the current study, we applied a paradigm called the goal-cueing task, which has been used previously in various experimental settings to elicit state of rumination (Edwards, 2017; Lanning, 2015; Michel-Kröhler, Wessa & Berti, 2023; Roberts et al., 2013, 2020). The paradigm is essentially based on the Goal Progress Theory of Martin and Tesser (1996; see also control theory by Carver & Scheier, 1982, 1990), which provides a detailed account of goal discrepancies as a central mechanism of rumination. Roberts et al. (2013) first applied the goal-cueing task and examined whether cueing an unresolved goal (i.e., a perceived discrepancy in the individual goal achievement process) elicited significantly more ruminative thinking than cueing a resolved goal. The results showed that participants in the unresolved condition experienced significantly more state rumination compared to participants in the resolved condition. Building on this, Michel-Kröhler, Wessa and Berti (2023) further modified the unresolved condition for future application in different contexts, which now serves as the basis for the present experiment (for more details see Methods). The chosen paradigm has the advantage compared to other forms of rumination induction such as the response manipulation task (cf. Lyubomirsky & Nolen-Hoeksema, 1993) that participants are free to choose the content so that a broader range of thoughts can be addressed to experience state rumination. In addition, participants can engage with their problem to the extent that they usually do. To summarize, pursuing and achieving goals is an essential part of (competitive) sports that largely determines athletes’ actions and behavior, which is why the application of the goal-cueing task in the context of sport seems very promising.

1.3. Aims of the experiment

In summary, there are few studies examining rumination in a sports context to date. These few studies are mainly based on correlational designs, so no causal conclusions can be derived. To deepen our understanding of the effects of rumination, experimental research approaches represent an important extension. Therefore, the aim of the present experiment was to investigate whether we could elicit significant rumination in athletes using the modified goal-cueing task. To this end, we proceeded as follows: First, to ensure that rumination can be reliably induced in athletes, we examined whether the context of a perceived discrepancy plays a role in the individual goal achievement process. For this, athletes were assigned to one of two experimental conditions, in which they identified either an ongoing and unresolved concern in their goal achievement process from their daily life (e.g., conflict with their partner; general problem condition [GPC]) or specific to their sport (e.g., conflict with their coach; sport-specific problem condition [SSPC]). Since people are more likely to ruminate about goals that are personally relevant and central to their well-being (Martin & Tesser, 1996), and since this can apply to both goals from everyday life and goals from the sport context, we assume that there are no significant differences between these two experimental conditions (ECs) in terms of state rumination assessed by two different measures, namely ruminative self-focus and a general rumination rating. However, to confirm the efficacy of using the goal-cueing task with athletes, state rumination should be significantly higher in both ECs after the goal-cueing task compared to before the goal-cueing task. In addition to the two ECs, we applied a control condition (CC) with an emotion-neutral task to identify the most effective form of inducing rumination and maximize its effect for future application in sports context. We assessed also further variables, using the same approach, to account for emotional processes in addition to cognitive processes. For this, we applied measures of mood and affect (for more details see Material & Measures).

Furthermore, we measured the extent and impact of residual rumination about goal discrepancies on performance on a simple go/no-go paradigm. To be consistent with previous studies, we used the Sustained Attention Response Task (SART; Robertson et al., 1997), even though the results were mixed in the past (Edwards, 2017; Michel-Kröhler, Wessa & Berti , 2023; Roberts et al., 2013, 2020). Therefore, we did not expect a high level of rumination during the implementation of the SART and rather small to no differences between conditions in terms of SART performance.

2. Methods

2.1. Procedure

Students of sports science or comparable courses of study from different universities were contacted via email through their respective academic offices. In addition, athletes from various team and individual sports throughout Germany were invited to participate in the study by e-mail via their respective clubs or sports associations. Participation requirements were a minimum age of 16 and a competitive experience of at least two years. The experiment consisted of two parts, which both were completed online at home due to the ongoing pandemic at the time of the experiment. Participants completed an initial online survey and, afterwards, performed the online-experiment.

The initial survey was conducted using SoSci Survey (Leiner, 2019) and comprised biographical, sociodemographic questions as well as training-related information, and different personality questionnaires (see below for detailed descriptions of the utilized questionnaires). The questionnaire data was also the basis for a stratified randomization of the participants. In detail, we first grouped athletes according to their level of trait rumination (measured with the Perseverative Thinking Questionnaire; PTQ, Ehring et al., 2011) and their gender, and second randomly assigned them either to the GPC, the SSPC or the CC. We thus ensured that athletes in the three conditions did not differ significantly from each other in these characteristics at the group level.

The study protocol was approved by the local Review Board of the Institute for Psychology of the Johannes Gutenberg-University Mainz and was conducted according to the guidelines of the Declaration of Helsinki (World Medical Association, 2013). Participants were informed about the nature and the procedure of the experiment and gave informed consent before completing the questionnaires. Participation in this experiment was voluntary and athletes received 12 Euro as compensation.

2.2. Quality of performance

To evaluate the quality of performance in the online experiment, we asked athletes at the end of the experiment (1) how concentrated they were, (2) how seriously they tried to implement the instructions on how to perform the experiment, and (3) how well they succeeded. Participants answered these three questions on a 5-point scale ranging from “1” (not at all) to “5” (very). In addition, the athletes indicated the location where they conducted the experiment and how quiet it was at that location. The latter was measured on an 11-point scale ranging from “0” (not at all quiet) to “10” (very quiet). Overall, 102 athletes completed the experiment. However, we excluded one participant who reported not making a serious effort to follow the implementation instructions from further data analysis. Moreover, the mean concentration of the athletes was 3.55 (SD = 1.00), they made an average effort of 4.52 (SD = 0.69) to follow the instructions seriously, and that they succeeded with 4.14 (SD = 0.88). Ninety-six athletes stated that they completed the experiment at home, whereas five athletes completed the experiment at the university or the library. The average rating for the quietness of the environment was 8.74 (SD = 1.58).

2.3. Participants

Table 1 summarizes the individual characteristics of athletes separately by condition and for the entire sample. Due to the specific population of athletes, we based our planned sample on the average size in quasi-experimental studies in sports context (N = 91 with an interquartile range from 45 to 206, Schweizer & Furley, 2016). Consequently, taking into account potential dropouts, we recruited 110 athletes, of whom 102 completed both parts of the experiment. However, due to the reasons mentioned above, the final sample consisted only of 101 athletes (female: n = 53, male: n = 48, M_age = 22.40, SD_age = 2.80) from different team and individual sports. Of these, 34 athletes assigned to an experimental condition, which means that 34 athletes performed the task under the general problem condition (GPC) while the other 34 athletes performed the task under the sports-specific problem condition (SSPC); the remaining 33 athletes were assigned to the CC. Group assignment was randomized with the restriction that trait rumination and gender was equally distributed across the groups (i.e., stratified randomization by gender and trait rumination). In general, athletic identity was very high in our sample, and athletes did not differ across conditions in this regard, F(2,98) = 0.30, p = .74, ƞp² < .01 (see Table 1). Furthermore, they also did not significantly differ in their general rumination assessed with the PTQ (Ehring et al., 2011): F(2,98) = 0.01, p = .98, ƞp² < .001 as well as in their sport-specific rumination captured with the Sport Competition Rumination Scale (SCRS; Michel-Kröhler et al., 2023), F(2,98) = 0.25, p = .77, ƞp² < .01.

Table 1. Mean (M), standard deviations (SD), 95% confidence intervals (95% CI) and Cronbach’s alpha (α) of sample characteristics separated by conditions

	General problem condition		Sport-specific problem condition		Control condition		Total
	M (SD)	95%CI	M (SD)	95%CI	M (SD)	95%CI	M (SD)	95%CI	α
	n  = 34 (f = 18, m = 16) individual = 17, team = 17		n  = 34 (f = 18, m = 16) individual = 16, team = 18		n  = 33 (f = 17, m = 16) individual = 18, team = 15		n  = 101 (f = 53, m = 48) individual = 51, team = 50
Sample characteristics
Age	22.18 (2.84)	-	22.53 (2.98)	-	22.48 (2.63)	-	22.40 (2.80)	-
Discipline-specific training/week	3.88 (2.36)		3.26 (1.56)		2.75 (1.00)		3.31 (1.78)
Additional training/week	2.97 (1.83)		3.00 (2.07)		2.85 (1.73)		2.94 (1.87)
Training hours/week	11.38 (4.18)		10.53 (4.06)		9.64 (4.25)		10.52 (4.18)
Competitions/year*	18.53 (12.01)		16.51 (12.75)		18.47 (15.17)		17.81 (13.22)
Experience (in years)	10.82 (4.67)		13.70 (5.14)		12.54 (4.66)		12.36 (4.93)
Performance level 1 2 3 4 5 6	1 1 4 4 16 8		1 2 4 4 16 7		0 2 0 7 15 9		2 5 8 15 47 24
Baseline Stress	28.82 (6.15)	[26.67, 30.97]	28.35 (5.09)	[26.58, 30.12]	26.70 (6.87)	[24.26, 29.13]	27.97 (6.08)	[26.77, 29.17]	.86
Personality-related information
Athletic identity	37.73 (5.11)	[35.95, 39.52]	36.68 (6.34)	[34.46, 38.89]	36.88 (6.32)	[34.64, 39.12]	37.09 (5.91)	[35.93, 38.26]	.73
Sport-specific rumination	18.73 (5.57)	[16.79, 20.68]	19.26 (5.71)	[17.27, 21.26]	18.27 (5.85)	[16.20, 20.35]	18.76 (5.67)	[17.64, 19.88]	.87
Perseverative thinking	25.23 (11.15)	[21.34, 29.12]	25.65 (12.25)	[21.37, 29.92]	25.67 (10.91)	[21.80, 29.53]	25.51 (11.34)	[23.27, 27.75]	.94
Brooding	10.44 (2.20)	[9.67, 11.21]	11.03 (2.92)	[10.01, 12.05]	10.48 (3.17)	[9.36, 11.61]	10.65 (2.78)	[10.10, 11.20]	.60
Reflection	10.62 (2.81)	[9.64, 11.60]	10.73 (2.89)	[9.72, 11.74]	9.97 (3.62)	[8.69, 11.25]	10.44 (3.11)	[9.83, 11.06]	.69
General self-efficacy	29.56 (3.11)	[28.47, 30.64]	29.32 (3.76)	[28.01, 30.63]	30.70 (3.65)	[29.40, 31.99]	29.85 (3.53)	[29.15, 30.55]	.78

Note. f = female, m = male, individual = individual athlete, team = team athletes, * answered only by 95 athletes. Performance level (PL) corresponds to the athletes’ assignment either to a squad (e.g., A-, B-, C-squad) or to a league (e.g., 1^st Bundesliga, 2^nd Bundesliga, 3^rd League): first PL (A-squad/Olympia-squad or 1^st Bundesliga), second PL (B-/C-squad equivalent to perspective or supplementary squad or 2^nd Bundesliga), third PL (Junior squad [NK1 & NK2], C-/DC-squad not belonging to perspective-squad or third highest league), forth PL (D-squad or forth highest league), fifth PL (another squad or another level below) and sixth PL (no squad status or no competition level).

2.4. Measures & material

We used the German versions of the following questionnaires. In Table 1, mean values, standard deviations, 95% confidence intervals (95% CI) as well as Cronbach’s alpha (α) of the respective scales are presented. Unless otherwise stated, we report the sum score of the scale used.

Perseverative Thinking. The Perseverative Thinking Questionnaire (PTQ; Ehring et al., 2011) is a content-independent self-report questionnaire of repetitive negative thoughts. The PTQ consists of 15 items (e.g., “Thoughts come to my mind without me wanting them to”) and is rated on a 5-point scale, ranging from “0” (never) to “4” (almost always). Here, we report the general PTQ score.

Brooding and Reflection. Huffziger and Kühner (2012) validated the 10-item short version of the Response Styles Questionnaire (RSQ; original English version: Treynor et al., 2003; long version: Nolen-Hoeksema, 1991) with the facets Reflection and Brooding. In the RSQ, it is assumed that brooding describes dysfunctional ruminating about an unattained goal (e.g., “What am I doing to deserve this?”), while reflection describes a more goal- and solution-oriented self-reflection (e.g., “I write down what I am thinking and analyze it.”). Each scale comprises five items. Participants rated all 10 items on a 4-point Likert Scale ranging from “1” (almost never) to “4” (almost always).

Sports-specific rumination. We used the Sports Competition Rumination Scale (SCRS; Michel-Kröhler, Krys & Berti, 2023) to assess rumination about competition-related problems. The SCRS consists of eight items (e.g., “I can’t stop thinking about competition-related problems.”) and is answered on a 5-point scale ranging from “1” (does not apply at all) to “5” (fully applies).

Athletic identity. We measured athletic identity using the Athletic Identity Measurement Scale (AIMS-D; Schmid & Seiler, 2003 English original by Brewer et al., 1993), which consists of seven items and is answered on a seven-point scale ranging from “1” (does not apply at all) to “7” (fully applies).

Self-efficacy. We used the General Self-Efficacy Scale (GSE; Hinz et al., 2006 English version: Schwarzer & Jerusalem, 1995) to assess participants’ general sense of perceived self-efficacy, for instance in relation to coping with everyday life or after experiencing all kinds of stressful life events. The GSE comprises 10 items (e.g., “I can always manage to solve difficult problems if I try hard enough.”) and is answered on a 4-point scale ranging from “1” (not at all true) to “4” (exactly true).

Perceived stress. We used the Perceived Stress Scale (PSS-10, Schneider et al., 2020 English original by Taylor, 2015) to assess the degree of how participants perceive situations in their lives as uncontrollable, unpredictable and overloaded relative to their subjective coping abilities. The PSS-10 comprises 10 Items (e.g., “In the last month, how often have you felt nervous and stressed?”) and is answered on a 5-point Likert scale ranging from “1” (never) to “5” (very often).

State Rumination. We tested state rumination with two measures: First, we assessed the momentary ruminative self-focus (Moberly & Watkins, 2010). For this purpose, two items (“How much did you focus on your feelings?”; “How much did you focus on your problems?”) are combined into one index and rated on an 11-point scale ranging from “0” (not at all) to “10” (very). We refer to this measure as the ruminative self-focus and reported the mean of the two items. Second, we used a single item (“To which extent did you ruminate over something?”; see also Koval et al., 2015) to assess general state rumination, which was also rated on the same 11-point scale. We refer to this measure as the general state rumination rating.

Mood. We assessed the Multidimensional Mood Questionnaire (MDMQ; Wilhelm & Schoebi, 2007) to measure three basic dimensions of mood—valence, energetic arousal, and calmness. The MDMQ consists of six items and is a bipolar measure that comprises three pairs of adjectives rated on a 7-point scale describing opposite end points of different mood dimensions (e.g., energetic arousal: tired vs. awake, full of energy vs. without energy).

Affect. We used the Positive Affect Negative Affect Schedule (PANAS; Krohne et al., 1996 English original version by Watson et al., 1988) to measure affective state. The PANAS includes 20 items to assess both positive and negative state affectivity, each with 10 items. Athletes were asked to rate the degree to which they feel the emotional state described in each item, on a five-point Likert scale, ranging from “1” (very slightly or not at all) to “5” (extremely).

Goal-cueing task. We used a modified version of the goal-cueing task by Michel-Kröhler, (Wessa and Bert,i 2023; original version by Roberts et al., 2013), which consists of three successive steps. First, athletes are instructed to identify an ongoing unresolved goal that repeatedly troubled them, causing them to feel sad, negative or stressed during the previous week. Athletes in the GPC should refer to a general problem (i.e., from their everyday life or their private life, such as a conflict with their partner), whereas athletes in the SSPC should refer to a sport-specific problem, such as problems with the coach, teammates, or their own performance development. Athletes were then asked to outline their problem in in 5 to 10 sentences (step 1 –goal identification). Second is the evaluation of certain characteristics of the identified goals by answering 12 different questions (step 2 –goal evaluation). They were asked to what extent the unresolved goal had been bothering them at its worst, and at the time of the experiment. Furthermore, participants indicate how important the goal is, how much the problem in the individual goal achievement process exemplifies more general problems, how long the problem exists and how much time they spend thinking about the problem last week. In addition to these questions, which were used as standard in the studies by Michel-Kröhler, Wessa and Berti (2023) and Roberts et al. (2013), we added six more questions related to the Goal Progress Theory (Martin & Tesser, 1996). Thus, athletes were also asked to indicate how much they are still currently pursuing this goal, how much the goal is linked to other goals, how high a priority the goal is compared to other goals in their life, how much their personal happiness depends on achieving the goal, how much they have focused on new goals, and how satisfied they are with their life in general. The questions were each answered on an 11-point scale from “0” to “10”. Furthermore, these can be used alongside the state ruminations measures as a manipulation check for the goal-cueing task. The third step consists of a 5-minute goal focus phase (Step 3 –goal focus period) that encourages athletes to reflect on their identified problem. Example instructions are “Think about what is important about this difficulty in terms of your personal goals”, or “Focus on the aspects of the difficulty that repeatedly come to mind”. Athletes in the CC were given a neutral writing task that did not relate to a personal goal but had the same extent. Athletes were asked to describe what they did from the morning until the time of the experiment (see Konig et al., 2014; Yu-Hsin Liao et al., 2012) instead of identifying a problem of their individual goal achievement process.

Sustained Attention to Response Task (SART). The SART (Robertson et al., 1997) is a simple go/no-go paradigm in which neutral words (e.g., “father”, “shirt”, or “green”) are presented in white text on a black background in the center of the screen. Each word appeared individually on screen for 300 ms followed by a 900 ms mask. The participants’ task was to respond to words in lowercase letters as quickly as possible by pressing the space bar (go-trials; e.g., “flower”). When a word appeared in upper case, participants were required to withhold their response (no-go-trial, e.g., “CHURCH”). The SART comprises in its shortened version two blocks of 450 trials each, consisting of 45 words repeated ten times in a different order. Within each set of 45 words, five uppercase words appeared randomly among 40 lowercase words. After 450 trials (i.e., one block), we allowed the participants to take a break, where they could determine the length of the break themselves up to a maximum of 3 minutes. The SART took approximately 30 minutes to complete.

Post evaluation of the identified problem. After completing the experimental part, participants indicated (1) how difficult it was to stop thinking about the concern, (2) to what extent their focus was mainly on negative aspects, or (3) on bad feelings, (4) to what extent thinking about the problem made it seem worse and (5) made them feel worse (questions were adapted from Mosewich et al., 2013). Participants rated all questions on a 5-point scale ranging from “1” (not at all) to “5” (very).

2.5. Statistical analyses

Collection of experimental data was carried out with the Inqusit Webplayer (Free Inquisit 5 Player app, Millisecond Software, Seattle, WA) and data preparation and all statistical analyses were performed with the software R Studio (R Core Team, 2018).

Statistical Tests. To analyze the effects of the goal-cueing task on state-rumination, mood and affect, we applied single mixed analyses of variance (ANOVAs) using the ezANOVA-function (“ez”-package; Lawrence, 2016) with time (before goal-cueing task, after goal-cueing task, after SART) and condition (general goal, sport-specific goal, control) as a factor. Beforehand, we checked the requirements for the application (normal distribution and homogeneity of variances). We conducted a Shapiro-Wilk Test for testing the assumption of normality (p > .05) and a Levene’s Test for testing the homogeneity of variance (p > .05). In case of violation of sphericity, we used Mauchly’s Test (p < .05) and applied a Greenhouse-Geisser correction for ANOVAs. Effect sizes were specified in the form partial eta squared (ηp²) with the following criteria for small, medium, and large effect: 0.01, 0.06, and>0.14 (Fritz et al., 2012).

To compare the SART task performance across athletes in different conditions, we examined the errors of commission (i.e., incorrect responses: key presses in no-go trials) as well as the mean reaction times (mean RTs) for correct go-trials. According to Cheyne et al. (2009), mean RTs were calculated for all response latencies over 200 ms. Reaction times less than 100 ms were coded as anticipatory and reaction times between 100 ms and 200 ms were coded as ambiguous. We then applied mixed ANOVAs with time (SART: Block 1 vs. Block 2) and condition (general goal, sport-specific goal, control) as factors.

With regard to participants’ characterization of problems in the individual goal achievement process and their evaluation of the identified problem after the experimental setting, we analyzed mean differences between conditions with independent t-tests. In case of non-parametric distribution, we reported the significance of Wilcoxon signed-rank test as robust alternative for independent t-tests (p_wilcox; Field et al., 2012). We report the effect size of mean differences between conditions with Cohen’s d (Cohen, 1988) with the following criteria: d = 0.20, d = 0.50 and d = 0.80 for small, medium, and large effects (see also Fritz et al., 2012 for interpretation). In case of non-parametric distribution, we reported corresponding robust effect size (r) with 0.10 to<0.30 for a small effect, 0.30 to<0.50 for a moderate effect and≥0.50 for a large effect (Cohen, 1992).

3. Results

3.1. Impact of goal manipulation on state rumination, affect and mood

Before the experiment started, we measured athletes’ perceived stress with the PSS-10 (Schneider et al., 2020) to ensure that athletes in the three conditions participated in the experiment with the same baseline levels (see Table 1). Results of a one-way ANOVA indicated that there were no significant differences between the conditions in terms of athletes’ perceived stress, F(2,98) = 1.13, p = .33, ηp² = .02.

Wilcoxon signed-rank test confirmed that athletes in the ECs did not significantly differ in the extent to which the goal had bothered them at its worst (p_wilcox = .58, r = .07), the extent to which the goal had bothered them at the time of the experiment (p_wilcox = .05, r = .23), or the amount of time they spent on the problem (p_wilcox = .14, r = .11), indicating that the goal manipulation worked and athletes in the ECs did not differ in subjective evaluations of their severity. In addition, there was a significant condition difference in the evaluation of the goal importance (p_wilcox < .01, r = .32) and the priority of the goal compared to other goals (p_wilcox <.001, r = .41). Athletes in the GPC reported a significant higher goal importance and higher priority of their goal than athletes in the SSPC. Table 2 shows means, standard deviations and t-test results of the respective ratings of the goal characteristics.

Table 2. Mean (M), standard deviations (SD), t-test results, and effect sizes of the goal characteristics for both experimental conditions

	GPC		SSPC
Goal characteristics	M	SD	M	SD	Group comparison	Effect size
(1) Trouble at the time of the experiment	8.44	1.33	7.20	2.45	pwilcox = .05	r = .23
(2) Trouble in its worst …	9.15	1.13	9.00	1.23	pwilcox = .58	r = .07
(3) Goal Importance	8.62	1.81	7.65	1.74	pwilcox < .01	r = .32
(4) Example for general problems	5.44	3.30	4.97	2.84	pwilcox = .46	r = .09
(5) Duration of the problem	5.59	2.62	6.35	2.90	pwilcox = .03	r = .26
(6) Time spent with the problem	7.23	1.98	5.79	2.81	pwilcox = .14	r = .11
(7) Current goal pursuit	7.76	1.94	6.17	2.85	pwilcox = .02	r = .27
(8) Dependence personal happiness	6.68	3.03	5.20	3.12	pwilcox = .05	r = .24
(9) Connection to further goals	6.35	2.95	6.09	2.81	pwilcox = .59	r = .06
(10) Focus on new goals	5.05	2.92	5.00	2.65	t(65.38) = 0.09, p = .93	d = .02
(11) Satisfaction with life	6.76	2.34	6.85	2.00	pwilcox >.99	r < .001
(12) Priority of the goal	6.59	2.55	4.29	2.70	pwilcox <.001	r = .41
Goal evaluation
(1) Difficulties stop thinking about the problem	2.70	1.27	2.65	1.23	pwilcox = .87	r = .02
(2) Focus on negative aspects	2.88	1.17	3.32	1.27	pwilcox = .14	r = .18
(3) Focus on bad feelings	2.73	1.26	3.26	1.31	pwilcox = .10	r = .20
(4) Problem seemed worse	2.73	1.19	2.82	1.53	pwilcox = .81	r = .03
(5) Athlete felt worse	2.85	1.39	2.91	1.42	pwilcox = .88	r = .02

Note. GPC = General problem condition, SSPC = sport-specific problem condition, with exception of question 10, we reported the significance of Wilcoxon signed-rank test (p_wilcox) as robust alternative for independent t-tests with the corresponding effect size (r), d = Cohens d.

Ruminative self-focus. Two 3 (condition: general goal, sport-specific goal, control) x 3 (time: before goal-cueing task, after goal-cueing task, after SART) mixed ANOVAs examined the effects of the goal-cueing task on state rumination as assessed by ruminative self-focus and the general state rumination rating. There was a significant main effect of condition on ruminative self-focus, F(2,98) = 4.20, p = .02, ηp² = .08, and a significant main effect of time, F(2,196) = 81.92, p[GG] < .001, ηp² = .45, which were qualified by a significant interaction between time and condition, F(4,196) = 13.39, p[GG] < .001, ηp² = .21. Pairwise comparisons revealed a significant increase in ruminative self-focus between time 1 (before goal-cueing task) and time 2 (after goal-cueing task) in both ECs (both ECs: p_bonf <.001), and no significant change in the CC. The conditions did not differ in ruminative self-focus prior to the goal-cueing task, but significantly differed after the goal-cueing task, reflecting higher values for athletes in both ECs compared to athletes in the CC (p_bonf <.001). Athletes in the GPC and the SSPC did not differ in their ruminative self-focus at time 2, indicating that cueing unresolved goals in general and sport-specific context increased similarly ruminative self-focus from before to after the goal-cueing task. Moreover, there was a significant reduction in ruminative self-focus between time 2 (after the goal-cueing task) and time 3 (after the SART) in both ECs (both ECs: p_bonf <.001), with no difference found in the CC. Finally, at time 3, there were no differences between the three conditions. Table 3 shows means, and standard deviations of the respective state measures separated by condition and time.

Table 3. Mean values (M) and standard deviations (SD) of relevant experimental state variables separated by condition and time

	General problem condition			Sport-specific problem condition			Control condition
	Time 1	Time 2	Time 3	Time 1	Time 2	Time 3	Time 1	Time 2	Time 3
	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)
State Rumination
Ruminative self-focus	3.76 (1.69)	5.97 (0.77)	2.37 (1.49)	4.17 (1.41)	5.97 (0.98)	3.10 (2.09)	4.47 (1.19)	3.82 (1.28)	2.95 (1.75)
General rumination rating	3.47 (1.81)	5.82 (1.14)	2.65 (1.81)	3.53 (1.65)	5.32 (1.39)	2.88 (2.21)	3.97 (1.81)	3.54 (1.70)	2.91 (1.96)
Affect
Positive affect	14.94 (5.09)	18.26 (5.92)	15.79 (5.05)	15.20 (4.90)	17.44 (5.57)	17.09 (5.43)	16.61 (5.98)	14.42 (4.68)	16.09 (5.82)
Negative affect	30.15 (6.78)	26.65 (6.52)	21.09 (7.46)	30.41 (5.12)	28.62 (6.19)	21.15 (6.37)	32.21 (6.42)	32.30 (7.55)	27.06 (9.75)
Mood
Energetic Arousal	3.59 (1.41)	3.23 (1.23)	2.62 (1.48)	3.54 (1.23)	3.44 (1.13)	2.25 (1.34)	3.61 (1.36)	3.70 (1.33)	2.68 (1.61)
Valence	3.68 (1.27)	2.87 (0.95)	2.82 (1.31)	3.87 (0.99)	3.12 (1.21)	2.50 (1.18)	3.82 (1.27)	3.97 (1.41)	3.36 (1.32)
Calmness	3.81 (1.33)	3.20 (1.27)	2.38 (1.26)	4.00 (1.34)	3.35 (1.47)	2.60 (1.18)	3.92 (1.40)	3.94 (1.38)	2.92 (1.40)

Note. time 1 = before the goal-cueing task, time 2 = after the goal-cueing task, time 3 = after the SART.

General rumination rating. Regarding the general state rumination rating, a significant main effect of time, F(2,196) = 46.54, p[GG] < .001, ηp² = .32 as well as of the interaction between time and condition, F(4,196) = 9.11, p[GG] < .001, ηp² = .16 was obtained. The effect of condition remained non-significant, F(2,98) = 1.72, p = .18, ηp² = .03. Subsequent post-hoc test revealed a significant increase in the general state rumination rating between time 1 and time 2 in both ECs (GPC: p_bonf <.001; SSPC: p_bonf <.01) and no significant change in the CC. Again, participants in the different conditions did not significantly differ in state rumination before the goal-cueing task, but differences were obtained after the goal-cueing task between GPC and CC (p_bonf <.001) as well as between SSPC and CC (p_bonf <.01). The ECs showed no differences at time 2, supporting the results of the ruminative self-focus. In addition, a significant decrease in the general state rumination rating between time 2 and time 3 was obtained in both ECs (p_bonf <.001) but not in the CC. Again, at time 3, participants in the three conditions no longer differed in their general rumination rating. These results were also underlined by the evaluation of the athletes’ goals in the ECs at the end of the experiment (after SART).

Evaluation of the identified problem after the experimental setting. Athletes did not differ in their difficulty with either the stopping to think about the goal or focusing on negative aspects or bad feelings. Additionally, there were no differences in how the athletes felt after the experiment or in their perception of the severity of the problem. The lower part of Table 2 presents the means, standard deviations and the t-test results of the respective post evaluation ratings separated by condition.

Positive and negative affect. A further 3 × 3 mixed ANOVA was applied to investigate the effect of the goal-cueing task on positive and negative affect as assessed by the PANAS separately. There was a significant main effect of condition on negative affect, F(2,98) = 5.91, p < .01, ηp² = .11, and a significant main effect of time on negative affect, F(2,196) = 74.18, p < .001, ηp² = .43. The effect of the interaction between time and condition was not significant, F(4,196) = 2.44, p = .05, ηp² = .05. With regard to positive affect, there was no significant effect of condition, F(2,98) < 1, and no significant effect of time, F(2,196) = 2.19, p = .11, ηp² = .02, but a significant interaction between time and condition, F(4,196) = 5.24, p < .001, ηp² = .10. However, subsequent pairwise comparisons revealed no significant differences.

Mood. Last, three more 3 × 3 ANOVAs were conducted to also examine the effects of the goal-cueing task on mood assessed by the MDMQ. Regarding the energetic arousal, there was no significant effect of condition, F(2,98) < 1, but a significant effect of time, F(2,196) = 37.44, p[GG] < .001, ηp² = .28, indicating a decrease in athletes’ energetic arousal as time progressed. The interaction between condition and time remained non-significant, F(4,196) = 1.09, p = .36, ηp² = .02. The same pattern emerged for calmness of athletes. There was a significant main effect of time, F(2,196) = 39.33, p < .001, ηp² = .29, but no significant effect of condition, F(2,98) = 1.70, p = .19, ηp² = .03, or of the interaction, F(4,196) = 1.09, p = .36, ηp² = .02. In contrast, results of the 3 × 3 mixed ANOVA revealed a significant main effect of condition on valence, F(2,98) = 4.31, p = .02, ηp² = .08, as well as a significant main effect of time, F(2,196) = 21.11, p < .001, ηp² = .18, which was qualified by a significant main effect of the interaction, F(4,196) = 3.56, p = .01, ηp² = .07. Subsequent post-hoc analyses revealed that athlete’s valence ratings did not significantly change from time1 to time 2, and to time 3 for all three conditions. Only difference between the conditions was obtained after the goal-cueing task, where athletes of the GPC showed significantly lower valence values compared to the CC (p_bonf <.01).

Summary. The results confirm a successful induction of state rumination in that values of both ECs for the general state rumination rating and ruminative self-focus increase from before the goal-cueing task to after the goal-cueing task in contrast to the CC. Furthermore, the results of positive affect provide little interpretable results, while the negative affect of the athletes seems to decrease over time and the CC surprisingly has the highest values. In terms of mood, there was essentially a decrease over time in all three dimensions and no group differences except for valence. Valence was higher in the GPC than in the CC after the goal cueing task.

3.1.1. SART performance

Results of the 3 (condition: general goal, sport-specific goal, control) x 2 (time: SART block 1, SART block 2) mixed ANOVA showed no significant main effects for either errors of commission or the mean RTs. Table 4 illustrates the percentage error rates and mean RTs for each condition during each block of the SART and the respective results of the F-statistic.

Table 4. Mean values (M) and standard deviations (SD) of errors of commission (in percent) and mean reaction times (mean RTs) for correct go-trials during the Sustained Attention to Response task (SART)

	Block 1	Block 2	Results F-statistic
	M (SD)	M (SD)	Errors of commission	Mean RT
General problem condition			FC(2,93) < 1 FT(1,93) = 3.84, p = .05, ηp^2 =.04 FI(2,93) = 2.83, p = .06, ηp^2 =.06	FC(2,93) = 1.09, p = .34, ηp^2 = .02 FT(1,93) = 3.88, p = .05, ηp^2 = .04 FI (2,93) = 1.35, p = .26, ηp^2 = .03
Errors of commission	47.12 (23.54)	46.19 (26.32)
Mean RT	387.21 (78.37)	405.36 (91.20)
Sport-specific problem condition
Errors of commission	43.45 (16.21)	51.57 (20.48)
Mean RT	373.08 (56.21)	373.49 (63.73)
Control condition
Errors of commission	46.38 (20.40)	48.58 (24.45)
Mean RT	370.57 (72.69)	378.35 (75.58)

Note. F_c = F-statistic in terms of condition, F_T = F-statistic in terms of time, and F_I = F-statistic in terms of the interaction condition x time. Data of five athletes were removed before data analysis because of non-serious reactions during the SART (valid go-trials = 0).

4. Discussion

The aim of the present experiment was to test the modified goal-cueing task for the induction of rumination for the specific context of sports. There are two main findings: First, the results of the mixed ANOVA showed that the level of state rumination in both experimental conditions increased significantly after the goal-cueing task compared to the control condition, which was consistent with our hypothesis. Second, regardless of the context of the goal (general, sport-specific), there were no differences between athletes in the experimental conditions, suggesting that the goal-cueing task can be reliably applied to induce rumination in the sport context. Additional questions characterizing the goals did not show any differences in the severity of the problems, but they did show differences in priority and importance of the goals. Stratified randomization of athletes into one of the three groups may not have allowed them to choose their most important goal (if it was not in the assigned context). However, as the results showed that both contexts are equally likely to trigger state rumination in athletes, in future studies they could freely choose in which context they identify a problem in their goal achievement process and thus focus on their most important goal. Furthermore, because people are more likely to ruminate about goals that are personally relevant and central to their well-being (Martin & Tesser, 1996), the free choice of context could further increase participants’ state rumination through the goal-cueing task.

Regarding the further time course after the goal-cueing task until the end of the experiment (after the SART), a significant decrease of state rumination was observed in both experimental conditions, which even fell below the initial level. This can have two reasons: First, the SART could serve as a welcome distraction to switch from a self-referential focus to an external focus. Athletes in the experimental conditions might have used self-regulation strategies (e.g., positive refocusing, relativization, or reappraisal) to keep their focus on the task and perform well. Especially athletes with several years of competitive experience represent a target group with experience in self-regulation or at least a good knowledge of it. Thus, it cannot be ruled out that theoretical and practical knowledge of self-regulation influenced state rumination during the SART. Second, in our and previous studies (see Roberts et al., 2013, 2020), participants defined a personally relevant goal but were then confronted with a task that was unrelated to the defined goal and thus did not seem goal enhancing or particularly personally relevant. Consequences of induced rumination might be stronger or last longer when a personally relevant task was used, as might occur in a sports context. However, the null results are consistent with previous findings (Michel-Kröhler, Wessa & Berti, 2023; Roberts et al., 2020) and future studies should therefore combine the goal-cueing task with a more relevant task in the laboratory or directly transfer the goal-cueing task into the field studies.

Surprisingly, the analyses related to affect showed that negative affect decreased over time in all conditions, and athletes in the control condition reported the highest levels of negative affect during the experiment compared to both experimental conditions. In contrast, no significant main effects for condition or time were observed for positive affect. However, the descriptive values showed that the positive affect in both experimental conditions increased after the goal-cueing task, compared to the control condition, in which the positive affect tended to decrease. This may be due to three reasons: (1) Participants were already told in the instructions immediately before the experimental part that they have to identify a personal concern in their individual goal achievement process, which could already trigger higher negative affect and lower positive affect in the experimental conditions before the goal-cueing task. (2) The affective outcomes could also be related to the specificity of rumination during the goal-cuing task. Roberts et al. (2020) indicated that when rumination was high in specificity, there was no evidence that state rumination impaired affect regulation. Accordingly, future studies should consider the specificity of rumination, as well as other characteristics of rumination (e.g., repetitiveness, intrusiveness) as appropriate, and examine how these relate to athletes’ affect regulation. (3) Martin and Stoner (1995) stated that negative affect does not always suggest goal non-achievement, nor does positive affect always suggest goal achievement. Sometimes the opposite is also true, depending in part on the context in which these feelings occur (Martin & Tesser, 1996).

With respect to mood, apart from the time effects, indicating a decrease in the athletes’ energetic arousal, calmness, and valence over the course of the experiment, no significant differences were found between the two experimental conditions. Thus, it can be assumed that athletes in the two experimental conditions did not differ significantly in their emotional response to the goal-cueing task, which is in line with previous findings (Michel-Kröhler, Wessa & Berti, 2023; Roberts et al., 2013). The only difference between conditions was obtained after the goal-cueing task, where athletes of the GPC showed significantly lower valence values compared to the control condition. It does not seem surprising that athletes feel rather unwell and discontent after identifying and evaluating a problem in their individual goal achievement process compared to an athlete who reports on his or her previous daily routine. However, it is interesting to note that this is only the case for athletes in GPC and not in SSPC. This could be due to the different perceived importance and priority of the goals in the two experimental conditions. This also fits with the statement of Martin and Tesser (1996), who stated that although there are different modes of rumination, they are all based on the same underlying mechanism, namely a problematic goal process. The only difference between the modes is the content (e.g., valence, time orientation, etc.).

Finally, this experiment was limited by being conducted as online experiment, which makes it much more difficult to control conditions and potential confounding variables compared to conducting the experiments in the laboratory (Hussy et al., 2013). However, we took some steps to increase the quality of the implementation of the online experiments (for more details see also Michel-Kröhler, Wessa & Berti, 2023 using the same approach). We additionally asked the participants for their honest evaluation of their performance quality. The results suggested that the implementation and the environment was appropriate for participation in an experiment. Moreover, existing comparative studies between laboratory and online experiments showed that the results agree surprisingly well (Musch & Klauer, 2002; Reips & Lengler, 2005).

In summary, our results support previous findings that rumination is related to individual goal achievement and thus extend the scope of Goal Progress Theory (Martin & Tesser, 1996) to sport. For example, it would be valuable to examine the relationship between the dynamics of state rumination and relevant performance parameters in sports (i.e., objectively measurable performance in the respective sports). To verify this, the paradigm needs to be applied in future studies in a natural setting such as training, where daily influencing factors such as personal relationships and social interactions (with coaches, peers, opponents, families, etc.) take place (see also Healy et al., 2018). Further, practical advice can be derived for coaches and athletes that can be used to address athletes’ perceived discrepancies between a desired goal and current state. First, coaches should be aware that it is not only sport-specific problems that can trigger increased rumination in their athletes, but also problems in other areas of the athletes’ lives that should be considered. Second, it is important to bear in mind that goal achievement is subjective, that means that the objective completion of a task is not the same as the subjective achievement of a goal (Martin & Tesser, 1996). According to this, an athlete may have achieved his or her personal best time, but not his or her goal of achieving a medal rank. Therefore, rumination may also be an indication to the coach of “hidden” goals of the athlete that were not considered at the beginning of the process or have gained importance during the pursuit. This means that the coach can use a perceived discrepancy in the athlete’s goal achievement process as an opportunity to discuss and, if necessary, adjust previously defined goals. In the context of Goal Progress Theory (Martin & Tesser, 1996), a particularly useful hint for athletes seems to be that ruminating ends only when (1) goals are achieved, (2) alternative ways to achieve goals are generated, (3) alternative thoughts can be used to evaluate the individual’s goal achievement process, or (4) when there is a willingness to give up on a goal (D’Zurilla & Goldfried, 1971). In order to better internalize these thought processes, Ntoumanis and Sedikides (2018) suggest mental contrasting. This is a method that can help athletes identify goal unattainability early in their individual goal achievement process, facilitating goal exit or goal re-entry at a later time to promote adaptive affective and cognitive responses rather than potential negative consequences. This should enable athletes to better differentiate between achievable and unachievable goals and increase their goal pursuit by increasing their commitment to realistic goals (for more details see Ntoumanis & Sedikides, 2018; see also Healy et al., 2018 for further recommendations for applied practice).

5. Conclusion

In conclusion, the current experiment has shown that the goal-cueing task can significantly induce state rumination in athletes. Therefore, this procedure can be used in future (sport-specific) studies to experimentally investigate questions about the influence of rumination on different (performance) domains. In addition, the results further support the strong correlation between rumination and individual goal achievement in the face of unresolved goals in the context of sport.

Declaration of conflicting interests

The authors report there are no competing interests to declare.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data that support the findings of this study are openly available in OSF at https://osf.io/xc6yg/

Additional information

Notes on contributors

Alena Michel-Kröhler

Alena Michel-Kröhler is a scientist at the Institute for Psychology at the Johannes Gutenberg-University Mainz, Germany. Her main research interest is the examination of the influence of functional and dysfunctional thoughts on athletic performance using different methodological approaches (e.g., questionnaire studies, laboratory and field experiments).

Stefan Berti

Stefan Berti is an associate professor at the Institute for Psychology at the Johannes Gutenberg-University Mainz, Germany. He has a degree in psychology and sports science and uses mainly experimental and psychophysiological method for his empirical research.