A
Hope as a Trojan Horse in Sadness and Fear Regulation: Evidence for Emotion Self-Induction as a Regulatory ProcessYasmine Capela1 and Elise S. Dan-Glauser1
1 Institute of Psychology, University of Lausanne, Switzerland
Correspondence concerning this article should be addressed to Yasmine Capela, Institute of Psychology, University of Lausanne, Bâtiment Géopolis, Quartier UNIL-Mouline, CH-1015 Lausanne, Switzerland. Email: yasmine.capelaabdalla@unil.ch
Abstract
Among the many strategies studied for emotion regulation, cognitive approaches such as reappraisal or attentional deployment have received particular emphasis. This focus, however, may have obscured other, more intuitive affective pathways. We explored one such pathway: emotion self-induction (ESI)—the deliberate generation of an emotion to regulate another.
A
In an exploratory online survey (N = 146), participants reported how often they used different types of ESI across contexts involving positive vs. negative and mild vs. intense emotions, and which specific emotions they employed to regulate sadness, fear, guilt, and shame. Results revealed that 90% of participants consciously used ESI at least occasionally. The most common form—inducing positive emotions to down-regulate negative ones—accounted for 83.8% of all reported use and occurred significantly more frequently than any other type (p < .001, AKP = .67). This strategy correlated with adaptive regulation measures, suggesting that it belongs to a broader cluster of effective and well-being-related regulatory styles. Notably, hope emerged as the predominant emotion used to counter sadness (p < .001, AKP = .70) and fear (p < .001, AKP = .34). Based on these findings, we introduce the Oppositional Emotion Regulation Model (OERM), which proposes that regulation can arise from the structured opposition among emotions’ core components (e.g., appraisal, motivational tendency). By leveraging such oppositions, ESI may represent an intuitive, affect-based route to faster emotional recovery. This work invites direct experimental tests of the OERM to clarify the mechanisms linking emotional opposition and adaptive self-regulation.Keywords:
Emotion
Emotion Regulation
Hope
Emotion Self-Induction (ESI)
Oppositional Emotion Regulation Model (OERM)
A
A
A
A
A
The component process model describes emotions as synchronized changes across physiology, action tendencies, expression, and feeling, triggered by goal-relevant appraisals (Scherer, 2005). Their adaptiveness depends on situational fit. Any mismatches in type, frequency, or duration may be maladaptive. To counter this, people often engage in emotion regulation (ER), which includes strategies that modify any aspect of the emotional process (Gross, 2015). ER supports well-being (Kobylińska & Kusev, 2019), whereas dysregulation is linked to psychopathology (Gross, 2015).
Most ER research has focused on a few well-established strategies, mostly focused on cognitive processes, while emotion-driven forms remain less explored. This study examines a new ER strategy, emotion self-induction (ESI), involving the deliberate generation of one emotion to regulate another. ESI process would work by co-activating an emotion’s component patterns (e.g., its appraisals, action tendencies…) that are opposed to the initial emotion, reshaping the ongoing emotional episode.
Setting up ESI as an additional ER strategy comes from evidence indicating that emotions affect one another. Zhan et al. (2015) showed that sadness reduced anger-driven aggression, whereas fear heightened it. Lutz and Krahé (2018) confirmed the sadness–anger effect, as did Winterich et al. (2010), who found reciprocal dampening between sadness and anger, attributing it to appraisal contrasts. Authors explain that the low agency/control for sadness and the high agency/control for anger are the opposing mechanisms that allow them to inhibit each other. This opposing process would be moderated by the Behavioral Activation and Inhibition Systems, underscoring the action-tendency component. These results demonstrate emotional interplay but do not suggest this mechanism for intentional regulation. We propose testing ESI as an ER strategy.
This study had three aims. First (H1), the goal was to determine whether individuals frequently report using ESI as an ER strategy, assuming it's an intuitive route to ER. Second (H2), the study aimed to identify emotions used to regulate sadness, fear, guilt, and shame—unpleasant, low-control emotions typically targeted for change (Sacharin et al., 2012)—with pairings driven by contrasts in their emotional components. Third (H3), the study examined whether individuals prefer specific valence combinations depending on their dispositional tendencies, emotionality, and regulatory profile.
Method
Participants
Because ESI is a novel construct, no prior data were available to conduct an a priori power analysis. We also did not compute post hoc power, as such analyses rely solely on the observed p-value and do not provide meaningful information on sample adequacy. Hence, power was checked after data collection with a sensitivity analysis (see the Results section). The sample size was initially determined pragmatically according to data availability, recruitment feasibility, and resource constraints, resulting in 146 participants (age: M = 22.0, SD = 3.6; 120 women, 24 men, and two individuals who did not disclose their gender). The participants were French-speaking adult undergraduate students, recruited via the internal platform of Lausanne University, and received course credit as compensation for their participation. No additional exclusion criteria were applied.
Measures
To assess the frequency and characteristics of emotion self-induction (ESI) as a deliberate regulation strategy that was never investigated before, a dedicated set of items was developed, tailored to capture its various forms and contexts of use. In addition, we administered a series of validated questionnaires to evaluate potential psychological concomitants of ESI use, including ER strategy use and difficulties, emotional reactivity, personality traits, and mental health indicators.
Emotion Self-Induction (ESI) Questionnaire
A set of questions concerning the use of ESI as a strategy for ER was presented.
A
These items asked participants whether they deliberately used emotions of the same or opposite valence (positive versus negative) to regulate other emotions of varying intensity (intense versus mild). Four configurations were included, which are listed below with corresponding items:
1.
PESIN — positive ESI for regulating negative emotions (Cronbach’s α = .75).
(i)
“When I feel mild negative emotions, I deliberately evoke other positive emotions to lessen them."
(ii)
“When I feel intense negative emotions, I deliberately evoke other positive emotions to lessen them."
2.
NESIP — negative ESI for regulating positive emotions (Cronbach’s α = .71)
(i)
“When I feel mild positive emotions, I deliberately evoke other negative emotions to lessen them."
(ii)
“When I feel intense positive emotions, I deliberately evoke other negative emotions to lessen them."
3.
NESIN — negative ESI for regulating negative emotions (Cronbach’s α = .80)
(i)
“When I feel mild negative emotions, I deliberately evoke other negative emotions to lessen them."
(ii)
“When I feel intense negative emotions, I deliberately evoke other negative emotions to lessen them."
4.
PESIP — positive ESI for regulating positive emotions (Cronbach’s α = .79)
(i)
“When I feel mild positive emotions, I deliberately evoke other positive emotions to lessen them."
(ii)
“When I feel intense positive emotions, I deliberately evoke other positive emotions to lessen them."
Responses were rated on a 5-point Likert scale ranging from “Almost never” (1) to “Almost always” (5), similar to the response requested for the CERQ (Jermann et al., 2006). We framed ESI questions as an extension of the CERQ (Jermann et al., 2006), a well-established instrument for evaluating cognitive ER. The aim of this parallel is to benefit from the integrity of this scale measurement while expanding its scope to include ESI, an underexplored ER strategy.
A
In addition, participants were asked which emotions they typically use to regulate sadness, fear, guilt, or shame, known as emotions part of the unpleasant and low control quadrant of the Geneva Emotion Wheel (GEW, Sacharin et al., 2012), regardless of their answers to the questions prior. They could select one or multiple responses from 12 counterbalanced options, which are part of the other three quadrants of the GEW and include: joy, pride, happiness, satisfaction, hope, anger, disgust, envy, disdain, curiosity, relief, and “nothing”. For example, the sadness item asked: “When you feel sad, which of the following emotions do you use to lessen that feeling?” This section aimed to identify common emotion pairs used in self-induction, providing a foundation for future experimental investigations. A
The “nothing” category was added to allow participants to indicate if they did not use any of these emotions at all, providing valuable insight into how the use of ESI may vary depending on the target emotion being regulated. Identifying such patterns can clarify whether ESI is a universal strategy or one that is selectively applied depending on the qualia of the emotion experienced.Emotion Regulation Measures
Cognitive Emotion Regulation Questionnaire (CERQ, Jermann et al., 2006). The French version of the CERQ is a validated 36-item questionnaire designed to evaluate cognitive strategies individuals use to regulate emotions following threatening or stressful life experiences. There are nine subscales (nine strategies), each consisting of four items. Targeted ER strategies are acceptance, positive refocusing, refocus on planning, positive reappraisal, putting into perspective, self-blame, rumination, catastrophizing, and blaming others. For example, when deciding upon the feelings and thoughts experienced when confronted to negative situation, one item of the self-blame subscale states: “I feel that I am the one to blame for it.” One item of the acceptance subscale states: “I think that I have to accept that this has happened.” Participants rate each item on a 5-point Likert scale, ranging from 1 (almost never) to 5 (almost always). Published internal reliability coefficients range from Cronbach’s α = .68 to .87, which is consistent with the values obtained in our sample, ranging from .61 to .90.
Difficulties in Emotion Regulation (DERS, Dan-Glauser & Scherer, 2013). The French version of the DERS is a validated 36-item questionnaire designed to evaluate ER difficulties. It has six subscales, each consisting of six items: non-acceptance of emotional response, difficulties in adopting goal-directed behaviors, difficulties in controlling impulsive behaviors, lack of emotional awareness, limited access to ER strategies, and lack of emotional identification or clarity. For example, one item of the non-acceptance of emotional response subscale states: “When I’m upset, I become angry with myself for feeling that way.” One item of the difficulties in adopting goal-directed behaviors subscale states: “When I’m upset, I have difficulty getting work done.” Items are rated on a 5-point Likert scale ranging from 1 (almost never) to 5 (almost always). Published internal reliability coefficients range from Cronbach’s α = .74 to .90, which is consistent with the values obtained in our sample, ranging from .82 to .95.
Emotional Reactivity Measures
Emotion Reactivity Scale (ERS, Lannoy et al., 2014). The French version of ERS is a validated 21-item questionnaire designed to assess emotional sensitivity, intensity, and persistence. For example, one item of the emotional sensitivity subscale states: “When something happens that upsets me‚ it’s all I can think about for a long time.” One item of the emotional persistence subscale states: “If I have a disagreement with someone‚ it takes a long time for me to get over it.” Responses are rated on a 5-point Likert scale ranging from 0 (does not describe me at all) to 4 (describes me very well). Published internal reliability coefficients range from Cronbach’s α = .75 to .94, which is consistent with the values obtained in our sample, ranging from .77 to .94.
Emotional Style Questionnaire (ESQ, Kesebir et al., 2019). The ESQ is a validated 24-item questionnaire designed to assess six dimensions of emotional functioning: outlook, resilience, social intuition, self-awareness, sensitivity to context, and attention. For example, one item of the outlook subscale states: “I am very good at seeing the positive side of things.” One item of the resilience subscale states: “When I experience a setback, I do not stay upset for very long.” Participants respond on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). A French translation that was obtained through a translation/back-translation procedure in our lab was used. Preliminary (unpublished) French version validation obtained with 326 participants from the same population as in the present study yield satisfactory indices. A general Cronbach alpha of .77 for the full scale (ranging from .66 to .80 for the different subscales) was obtained. Confirmatory structure analysis showed that the underlying structure satisfactorily aligns with the original version of Kesebir et al. (2019), χ²(237) = 475.07, p < .001, RMSEA = .056, 90% CI [.048, .063], CFI = .893, TLI = .875. In the present sample, the observed internal reliability coefficients ranged from Cronbach’s α = .72 to .85, consistent with the above.
Reactivity Intensity Polarity and Stability questionnaire (RIPoSt-40, Brancati et al., 2019). The RIPoSt-40 is a validated 40-item questionnaire assessing reactivity, intensity, and affective stability in five subscales: affective instability (AI), positive emotionality (P), negative emotionality (N), emotional impulsivity (EI), and negative ER (NED). For example, one item of the EI subscale states: “I have quick, brutal, emotional reactions, almost impulsive ones.” One item of the N subscale states: “I easily feel stress over unexpected changes, even when they are of little importance.” The French version of the questionnaire was used. Participants rated each item on a 6-point Likert scale, ranging from 1 (never) to 6 (always). Published internal reliability coefficients range from Cronbach’s α = .72 to .95, which is consistent with the values obtained in our sample, ranging from .85 to .91.
Personality Trait Measures
NEO Five-Factor Inventory (NEO-FFI, Rolland et al., 1998). The French version of the NEO-FFI is a validated 60-item instrument that assesses the five major dimensions of personality: neuroticism, extraversion, openness, agreeableness, and conscientiousness. For example, one item of the neuroticism subscale states: “I am not a worrier.” One item of the conscientiousness subscale states: “I keep my belongings clean and neat.” Participants rate their agreement with each item on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Published internal reliability coefficients range from Cronbach’s α = .62 to .84, which is consistent with the values obtained in our sample, ranging from .68 to .88.
Mental Health Indicators
Perth Alexithymia Questionnaire (PAQ, Luminet et al., 2021). The French version of the PAQ is a validated 24-item assessment of alexithymia. It has five subscales: Negative-difficulty identifying feelings (N-DIF), Positive-difficulty identifying feelings (P-DIF), Negative-difficulty describing feelings (N-DDF), Positive-difficulty describing feelings (P-DDF), and General-externally orientated thinking (G-EOT). These subscales form composite indices: General-difficulty identifying feelings (G-DIF = N-DIF + P-DIF), General-difficulty describing feelings (G-DDF = N-DDF + P-DDF), Negative-difficulty appraising feelings (N-DAF = N-DIF + N-DDF), Positive-difficulty appraising feelings (P-DAF = P-DIF + P-DDF), General-difficulty appraising feelings (G-DAF = N-DAF + P-DAF), and an overall alexithymia score. For example, one item of the N-DIF subscale states: “When I’m feeling bad, I can’t tell whether I’m sad, angry, or scared.” One item of the G-EOT subscale states: “I tend to ignore how I feel.” Items are rated on a 7-point Likert scale, ranging from 1 (strongly disagree) to 7 (strongly agree). Published internal reliability coefficients range from Cronbach’s α = .89 to .96, which is consistent with the values obtained in our sample, ranging from .83 to .95.
Patient Health Questionnaire for Depression and Anxiety (PHQ-4, Kroenke et al., 2009). The PHQ-4 is a very brief, validated 4-item screening tool for anxiety and depressive symptoms, asking participant “In the last two weeks, how often have you been bothered by the following problems?". For example, one item of the anxiety subscale states: “Feeling nervous, anxious, or on edge.” One item of the depression subscale states: “Feeling down, depressed, or hopeless.” A French version whose items were taken from two French-validated questionnaires: the PHQ-9 (Carballeira et al., 2007) and the GAD-7 (Micoulaud-Franchi et al., 2016) was used. Responses are rated on a 4-point Likert scale ranging from 0 (never) to 3 (almost every day). Published internal reliability coefficients range from Cronbach’s α = .82 to .90, our sample obtained slightly lower values, ranging from .63 to .82.
Procedure
The procedure consisted of an online survey administered via Qualtrics (Qualtrics, Provo, UT). Participants were first presented with an information sheet and consent form, followed by demographic questions (age and gender). Participants then completed a series of questionnaires assessing emotionality, emotion regulation (ER), dysregulation, and personality, administered in a counterbalanced order. As part of these questionnaires, they reported on their use of ESI as an ER strategy, included at the end of the CERQ (Jermann et al., 2006). In addition, the four emotion-pairing questions were counterbalanced across participants, and the 12 response options (described in Measures) provided for each question were also presented in a randomized order. The testing ended with a short debriefing. Participants wanting more information could contact the researchers at any time during the process.
A
The procedure was approved by the ethical committee of the University of Lausanne (C_SSP_112020_00006).Data Analyses
As a preliminary step, we conducted a sensitivity analysis to evaluate whether the available sample size (N = 146) was adequate to detect effects of small-to-moderate magnitude.
Subsequently, our analyses focused on five primary indicators: (1) prevalence of ESI strategy use (0–100%), and (2) mean frequency of ESI use at the response level (1–5) to respond to our H1, (3) emotion-pairing prevalence (0–100%) of each emotion option used to regulate sadness, fear, shame, and guilt to respond to our H2, (4) correlations between frequency of ESI use and other measures to answer our H3, and (5) robustness analyses of the main results.
General data treatment rules. All analyses followed common data treatment procedures. Because moderate to strong correlations were observed between mild and intense opposite-valenced ESI (rs = .55–.60, p < .001) and no main effect of intensity emerged (p = .818, AKP = –.01; mild: M = 1.8, SD = .6; intense: M = 1.8, SD = .6), subsequent analyses were conducted on aggregated data combining mild and intense conditions for each ESI type. Prevalence estimates were then computed at the subject level by taking the ordinal maximum across relevant items (Almost never < … < Almost always), thereby assigning one label per participant. Proportions were calculated across participants and are reported with binomial standard errors, while composite scores represent means across conditions and thus use conventional standard errors instead of binomial ones. Frequency outcomes are summarized with means and standard deviations. Missing data were handled using available cases for stand-alone tests (i.e., participants were included if they had data on the variables in that specific test), whereas listwise deletion was applied within each family of analyses when Westfall–Young corrections were required, in order to ensure a common subject set across all contrasts. Composite variables were calculated as row-wise means of available items, and participants with all items missing on a composite were excluded from analyses involving that composite.
Comparative analyses. For within-subject contrasts, we applied sign-flip permutation tests on subject-wise differences (two-sided; 50,000 flips). When multiple related pairwise contrasts were prespecified, we controlled the family-wise error rate using the Westfall–Young maxT procedure (20,000 joint permutations on a common subject set). Effect sizes for paired contrasts were indexed by the Algina–Keselman–Penfield (AKP) robust standardized mean difference, based on 20% trimmed means and 20% winsorized standard deviations, scaled to Cohen’s d. When AKP estimates could not be computed (e.g., due to zero winsorized variance or very small n), Cohen’s dz was reported as a fallback.
Correlational analyses. Correlational analyses used Pearson’s r with permutation-based two-sided p-values. When multiple correlations were grouped within the same family (e.g., questionnaire × ER family), maxT corrections were again applied.
Robustness and replicability. To evaluate the robustness and replicability of our findings, we complemented classical statistics with four approaches: (a) jackknife leave-one-out, (b) bootstrap, (c) multiverse analyses, and (d) split-half replicability. These analyses focused on our key results and provided converging evidence for the reliability of the results. More details about these analyses are in the supplementary material.
All analyses were conducted in R, with AI-based assistance, under the supervision, adaptation, and validation of the authors.
Results
Preliminary results
We conducted a sensitivity analysis with N = 146 (α = .05, power = .80). For paired mean differences on 5-point scales, the minimum detectable effects were .14, .19, and .23 points, assuming within-person SDs of .6, .8, and 1.0, respectively. For paired proportions, the minimum detectable differences were 12.7, 16.4, and 18.0 percentage points for discordant-pair rates of q = .30, .50, and .60, respectively. These results indicate that the design was sufficiently powered to detect effects of small-to-moderate magnitude in usage frequencies and prevalence across strategies.
Prevalence and Frequency of ESI Strategy Use
Regarding prevalence, at the subject level, the proportion of participants reporting at least occasional use (which we can call prevalence) was 90% (SE = 2.6%) overall, 87.7% (SE = 2.9%) for negative ER (i.e., the target emotion to regulate is negative; PESIN or NESIN), and 56.9% (SE = 4.3%) for positive ER (i.e., the target emotion to be regulated is positive; NESIP or PESIP). Among the different ESI types, PESIN showed a prevalence of 83.8% (SE = 3.2%), NESIN 46.2% (SE = 4.4%), PESIP 46.2% (SE = 4.4%), and NESIP 36.9% (SE = 4.2%).
Regarding the distribution of participants across the different frequency categories for each ESI type, results showed that the majority of participants (79.2%) explicitly reported using PESIN from sometimes to often. Indeed, PESIN was almost never used by 16.2% (SE = 3.2%), sometimes by 27.7% (SE = 3.9%), regularly by 29.2% (SE = 4.0%), often by 22.3% (SE = 3.7%), and almost always by 4.6% (SE = 1.8%) of participants. NESIN was almost never used by 53.8% (SE = 4.4%), sometimes by 22.3% (SE = 3.7%), regularly by 11.5% (SE = 2.8%), often by 11.5% (SE = 2.8%), and almost always by .8% (SE = .8%) of participants. PESIP was almost never used by 53.9% (SE = 4.4%), sometimes by 23.9% (SE = 3.7%), regularly by 13.1% (SE = 3.0%), often by 8.5% (SE = 2.4%), and almost always by .8% (SE = .8%) of participants. NESIP was almost never used by 63.1% (SE = 4.2%), sometimes by 19.2% (SE = 3.5%), regularly by 11.5% (SE = 2.8%), often by 5.4% (SE = 1.9%), and almost always by .8% (SE = .8%) of participants.
Regarding our frequency of use, i.e., the mean intensity with which participants reported employing each strategy (ranging from “almost never” to “almost always”), our permutation tests showed a main effect of valence, p < .001, AKP = .72, with participants reporting greater frequency of use of ESI for negative ER (PESIN, NESIN; M = 2.0, SD = .7) than for positive ER (NESIP, PESIP; M = 1.5, SD = .6). PESIN frequency of use, M = 2.4, SD = 1.0, significantly differed from any other ESI strategies (M = 1.6, SD = .6), p < .001, AKP = .67.
Prevalence of Emotion Pairings in ESI
A
Hope emerged as the most used emotion in ESI to regulate sadness (M = 62.7%, SE = 4.3%), p < .001, AKP = .70, significantly more than all the other 11 options (Fig. 1; M = 19.6%, SE = 1.2%). It was followed by joy (M = 41.3%, SE = 4.4%), which differed significantly from hope, p = .012, AKP = .29, and happiness (M = 35.7%, SE = 4.3%), which also differed from hope, p = .001, AKP = .32. Detailed p-values and effect sizes for all other pairing comparisons in sadness regulation are provided in the Supplementary Material (Table S1).Fig. 1
Prevalence of use of each of the 11 emotions (or none) to regulate sadness, fear, shame, and guilt.
Note. Error bars represent the standard error (SE) of the proportion for each response option. Bars sharing a letter do not differ significantly; bars with different letters differ significantly (FWER, two-sided, α = .05). For clarity, letters are assigned so that “a” corresponds to the group with the highest mean prevalence, “b” to the next, and so on. In sadness and fear regulation, Hope appears in red where its selection rate exceeded that of sometimes all other emotions at a statistically significant level.
Hope was the most frequently used emotion to regulate fear (M = 44.4%, SE = 4.4%), significantly more than all other options, p < .001, AKP = .34 (Fig. 1; M = 15.4%, SE = 1.2%). Curiosity (M = 29.4%, SE = 4.1%) followed, showing a non-significant difference with hope, p = .117, AKP = .22, yet showing a noticeable tendency to be less frequently used. Relief use (M = 27.0%, SE = 4.0%) was significantly different than hope, p = .035, AKP = .22. Detailed p-values and effect sizes for all other pairing comparisons in fear regulation are provided in the Supplementary Material (Table S1).
Emotion choices for ESI use in shame were largely inconclusive (Fig. 1), with few and mostly non-significant differences across options, usage ranging from envy (M = 4.8%, SE = 1.9%) to pride (M = 23.8%, SE = 3.8%), ps ranged from .007 to 1.000. The only significant difference was between pride and envy, p = .006, AKP = .08. Detailed p-values and effect sizes for all other pairing comparisons in shame regulation are provided in the Supplementary Material (Table S1).
A similar pattern emerged for guilt regulation (Fig. 1), with usage ranging from joy (M = 4.0%, SE = 1.7%) to reporting no consciously used emotion (none; M = 25.4%, SE = 3.9%), and p-values spanning .003 to 1.000. Detailed p-values and effect sizes for all pairing comparisons in shame regulation are provided in the Supplementary Material (Table S1).
Correlations Between Frequency of ESI Use and Other Emotion and Personality Assessments
Table 1 presents all statistically significant and non-significant Pearson’s correlations between PESIN, NESIP, NESIN, and PESIP with each subscale of the questionnaires: CERQ, DERS, ERS, ESQ, RIPoSt-40, NEO-FFI, PAQ, and PHQ-4.
|
Questionnaires
|
PESIN
r; p
|
NESIN
r; p
|
PESIP
r; p
|
NESIP
r; p
|
|
|---|---|---|---|---|---|
|
CERQ (N = 130)
|
Adaptive regulation
|
.55; <.001
|
.01; .985
|
.16; .333
|
− .07; .896
|
|
Non-adaptive regulation
|
.08; .765
|
.26; .021
|
.25; .032
|
.34; .001
|
|
|
Acceptance
|
.45; <.001
|
.03; .968
|
.19; .174
|
− .04; .955
|
|
|
Positive refocusing
|
.52; <.001
|
.09; .830
|
.16; .378
|
.00; 1.000
|
|
|
Refocus on planning
|
.37; <.001
|
− .04; .968
|
.05; .784
|
− .12; .624
|
|
|
Positive reappraisal
|
.43; <.001
|
− .06; .938
|
.14; .420
|
− .11; .673
|
|
|
Putting into perspective
|
.28; .007
|
.01; .985
|
.08; .762
|
.00; 1.000
|
|
|
Self-blame
|
− .07; .765
|
.16; .402
|
.10; .695
|
.31; .004
|
|
|
Rumination
|
.13; .461
|
.11; .750
|
.15; .419
|
.06; .919
|
|
|
Catastrophizing
|
.05; .765
|
.31; .004
|
.33; .001
|
.31; .004
|
|
|
Other blame
|
.08; .765
|
.06; .938
|
.04; .784
|
.13; .566
|
|
|
DERS (N = 120)
|
Total score
|
− .36; <.001
|
.10; .531
|
.07; .700
|
.43; <.001
|
|
Non-acceptance of emotional response
|
− .22; .029
|
.27; .016
|
.15; .322
|
.47; <.001
|
|
|
Difficulties in adopting goal-directed behaviors
|
− .19; .036
|
− .02; .949
|
− .02; .911
|
.13; .147
|
|
|
Difficulties in controlling impulsive behaviors
|
− .26; .019
|
.12; .531
|
.17; .243
|
.34; .001
|
|
|
Lack of emotional awareness
|
− .33; .001
|
− .11; .531
|
− .12; .461
|
.21; .044
|
|
|
Limited access to emotion regulation strategies
|
− .32; .001
|
.27; .016
|
.09; .646
|
.36; <.001
|
|
|
Lack of emotional identification or clarity
|
− .23; .029
|
.00; .966
|
.03; .911
|
.36; <.001
|
|
|
ERS (N = 118)
|
Total score
|
− .11; .284
|
.21; .044
|
.13; .256
|
.21; .034
|
|
Emotional sensitivity
|
− .13; .229
|
.20; .051
|
.08; .532
|
.19; .056
|
|
|
Emotional intensity
|
.01; .946
|
.21; .050
|
.20; .069
|
.17; .059
|
|
|
Emotional persistence
|
− .24; .020
|
.18; .051
|
.07; .532
|
.24; .017
|
|
|
ESQ (N = 122)
|
Outlook
|
.36; <.001
|
− .18; .205
|
.04; .988
|
− .32; .002
|
|
Resilience
|
.28; .010
|
− .06; .891
|
.00; 1.000
|
− .20; .089
|
|
|
Social intuition
|
.14; .213
|
− .02; .891
|
− .01; 1.000
|
− .13; .280
|
|
|
Self-awareness
|
.16; .195
|
− .09; .774
|
.02; .999
|
− .32; .002
|
|
|
Sensitivity to context
|
.01; .874
|
− .21; .119
|
− .11; .738
|
− .26; .016
|
|
|
Attention
|
.20; .092
|
− .04; .891
|
.00; 1.000
|
− .13; .280
|
|
|
Ripost-40 (N = 121)
|
Affective instability (AI)
|
− .12; .172
|
.23; .039
|
.22; .066
|
.25; .019
|
|
Positive emotionality (P)
|
.21; .053
|
.01; .920
|
.12; .468
|
− .12; .356
|
|
|
Negative emotionality (N)
|
− .25; .020
|
.17; .173
|
− .03; .925
|
.37; <.001
|
|
|
Emotional impulsivity (EI)
|
− .26; .018
|
.06; .761
|
.03; .925
|
.11; .356
|
|
|
Negative emotion regulation (NED)
|
− .16; .125
|
.12; .398
|
.06; .826
|
.18; .105
|
|
|
NEO-FFI (N = 119)
|
Neuroticism
|
− .25; .016
|
.03; .848
|
− .11; .625
|
.22; .073
|
|
Extraversion
|
.14; .232
|
− .23; .049
|
− .09; .709
|
− .17; .202
|
|
|
Openness
|
.31; .002
|
.05; .848
|
.04; .859
|
.04; .887
|
|
|
Agreeableness
|
.09; .342
|
− .14; .416
|
− .13; .535
|
− .08; .740
|
|
|
Conscientiousness
|
.29; .004
|
− .13; .416
|
− .01; .952
|
− .03; .887
|
|
|
PAQ (N = 113)
|
Total score
|
− .15; .262
|
.10; .583
|
.05; .906
|
.38; <.001
|
|
General difficulty appraising feelings (G-DAF)
|
− .13; .302
|
.07; .753
|
.04; .906
|
.35; .001
|
|
|
General difficulty describing feelings (G-DDF)
|
− .15; .262
|
.05; .789
|
.00; .986
|
.31; .002
|
|
|
General difficulty identifying feelings (G-DIF)
|
− .09; .524
|
.10; .581
|
.09; .729
|
.37; <.001
|
|
|
Externally oriented thinking (G-EOT)
|
− .16; .262
|
.12; .511
|
.05; .906
|
.34; .001
|
|
|
Difficulty appraising negative feelings (N-DAF)
|
− .18; .188
|
.07; .786
|
− .05; .902
|
.23; .018
|
|
|
Difficulty describing negative feelings (N-DDF)
|
− .18; .171
|
.07; .789
|
− .06; .848
|
.20; .033
|
|
|
Difficulty identifying negative feelings (N-DIF)
|
− .15; .262
|
.06; .789
|
− .03; .906
|
.24; .016
|
|
|
Difficulty appraising positive feelings (P-DAF)
|
− .05; .641
|
.07; .789
|
.13; .414
|
.41; <.001
|
|
|
Difficulty describing positive feelings (P-DDF)
|
− .09; .524
|
.02; .874
|
.07; .830
|
.37; <.001
|
|
|
Difficulty identifying positive feelings (P-DIF)
|
− .01; .896
|
.12; .511
|
.19; .137
|
.42; <.001
|
|
|
PHQ4 (N = 109)
|
Anxiety
|
− .23; .032
|
.09; .366
|
.05; .595
|
.24; .020
|
|
Depression
|
− .15; .107
|
.28; .009
|
.13; .286
|
.22; .024
|
|
Note. P values were adjusted within each family using the Westfall–Young maxT procedure (FWER, two-sided, α = .05). Bold = significant, regular = not significant. PESIN stands for the use of positive emotions for regulating negative emotions; NESIP for the use of negative emotions for regulating positive emotions; NESIN for the use of negative emotions for regulating negative emotions; PESIP for the use of positive emotions for regulating positive emotions.
Robustness and Internal Replicability Results
A
Robustness analyses supported the reliability of our main findings (see Table S2 in the Supplementary Material). Overall, 53.3% of key results demonstrated high robustness, 14.3% showed moderate-to-high robustness, 29.9% were moderately robust, and 2.6% were moderate-to-low. The latter category pertains to weaker secondary correlations, such as the positive association between NESIP and the Difficulty of Describing Negative Feelings (N-DDF; PAQ) or NESIP and depression score (PHQ4).Regarding internal replicability, 46.8% of results exhibited high replicability, 14.3% were moderate-to-high, 11.7% moderate, 9.1% moderate-to-low, and 18.8% low. Although a subset of secondary correlations showed lower replicability—e.g., the negative association between NESIN and Extraversion (NEO-FFI), or between PESIN and Non-acceptance of Emotional Responses (DERS)—the overall pattern indicates that most findings are both statistically reliable and internally consistent, supporting confidence in the validity of the primary results.
Discussion
This study examined a newly defined ER strategy: ESI. We measured its frequency in daily life, its links to broader ER processes, and the emotions used to regulate specific targets. Our key finding is that hope emerges as a central regulatory emotion.
Most participants reported ESI use, predominantly by engaging positive emotions to regulate negative ones (PESIN), which was more common than other types. ESI thus functions as an ER strategy, with subcategories associated with adaptive versus maladaptive regulation, and which specificities are reflected in emotionality, regulation, dysregulation, and personality patterns.
Hope was the emotion most often recruited to regulate sadness and fear. This aligns with evidence that hope links reappraisal to better mental health (Peh et al., 2017), and extends it by showing that people actively self-induce hope to regulate emotions. Because hope acknowledges difficulty while remaining oriented toward possibility, it pairs well with sadness or fear. In contrast, no consistent regulatory emotion emerged for shame or guilt, suggesting that explicit ESI is less common for self-conscious emotions.
PESIN correlated positively with adaptive indicators (e.g., use of positive refocusing) and negatively with maladaptive ones (e.g., difficulties in emotion regulation). NESIP, although less frequent, showed associations with depression and anxiety, maladaptive regulation, and alexithymia. NESIN was similarly infrequent and weakly maladaptive. PESIP showed only links to catastrophizing and global non-adaptive scores, suggesting a more neutral profile. Personality results align with meta-analytic findings linking personality to ER strategies (Barańczuk, 2019): PESIN correlated positively with conscientiousness and openness and negatively with neuroticism, while NESIN was negatively related to extraversion. PESIN and NESIP also showed opposite links to emotional persistence, indicating that PESIN may shorten negative episodes, whereas NESIP may prolong them.
Based on hope-based PESIN, the Oppositional Emotion Regulation Model (OERM; see supplementary material Table S3) proposes that people induce emotions whose components oppose those of the target emotion. Grounded in the component process model (Scherer, 2005), it draws on contrasts between hope and sadness in appraisal, valence, action tendencies, and physiology, as well as fear–hope and sadness–anger (Zhan et al., 2015; Lutz & Krahé, 2018), supported by emotional blunting (Winterich et al., 2010) and appraisal-based transitions theories (Ellsworth, 1991).
Correlational design and student sampling limit generalizability, and self-reports may introduce bias. Still, findings identify ESI, especially hope-based PESIN, as an intuitive ER route and introduce OERM as a framework for regulation through emotion-opposition processes.
Declarations
Section
No conflicts of interest to disclose.
A
This study received ethics approval from the ethical committee of the University of Lausanne (C_SSP_112020_00006). A
Informed consent to participate and to publish was obtained from all participants. The data and codes supporting this study are provided in the supplementary materials. Authors’ contributions were as follows: Yasmine Capela contributed to Conceptualization, Methodology, Investigation, Formal Analysis, Data Curation, and Writing – Original Draft, while Elise Dan Glauser contributed to Conceptualization, Methodology, Supervision, and Writing – Review & Editing.Electronic Supplementary Material
Below is the link to the electronic supplementary material
A
Author Contribution
Authors’ contributions were as follows: Yasmine Capela contributed to Conceptualization, Methodology, Investigation, Formal Analysis, Data Curation, and Writing – Original Draft, while Elise Dan Glauser contributed to Conceptualization, Methodology, Supervision, and Writing – Review & Editing.
A
Data Availability
All data supporting the findings of this study are available within the paper, its Supplementary Material and at the following URL : https://osf.io/yxwu7/overview?view_only=ed00efad152f473cbdbe3cc0bd26513d
References
Barańczuk, U. (2019). The five factor model of personality and emotion regulation: A meta-analysis. Personality and individual differences, 139, 217–227. https://doi.org/10.1016/j.paid.2018.11.025
Brancati, G. E., Barbuti, M., Pallucchini, A., Cotugno, B., Schiavi, E., Hantouche, E. G., & Perugi, G. (2019). Reactivity, Intensity, Polarity and Stability questionnaire (RIPoSt-40) assessing emotional dysregulation: Development, reliability and validity. Journal of Affective Disorders, 257, 187–194. https://doi.org/10.1016/j.jad.2019.07.028
Carballeira, Y., Dumont, P., Borgacci, S., Rentsch, D., De Tonnac, N., Archinard, M., & Andreoli, A. (2007). Criterion validity of the French version of Patient Health Questionnaire (PHQ) in a hospital department of internal medicine. Psychology and Psychotherapy: Theory Research and Practice, 80(1), 69–77. https://doi.org/10.1348/147608306X103641
Dan-Glauser, E. S., & Scherer, K. R. (2013). The Difficulties in Emotion Regulation Scale (DERS): Factor structure and consistency of a French translation. Swiss Journal of Psychology, 72(1), 5–11.
Ellsworth, P. C. (1991). Some implications of cognitive appraisal theories of emotion. In K. T. Strongman (Ed.), International review of studies on emotion (Vol. 1, pp. 143–161). Wiley.
Gross, J. J. (2015). Emotion regulation: Current status and future prospects. Psychological inquiry, 26(1), 1–26. https://doi.org/10.1080/1047840X.2014.940781
Jermann, F., Van der Linden, M., d'Acremont, M., & Zermatten, A. (2006). Cognitive emotion regulation questionnaire (CERQ). European Journal of Psychological Assessment, 22(2), 126–131. https://doi.org/10.1027/1015-5759.22.2.126
Kesebir, P., Gasiorowska, A., Goldman, R., Hirshberg, M. J., & Davidson, R. J. (2019). Emotional Style Questionnaire: a multidimensional measure of healthy emotionality. Psychological assessment, 31(10), 1234. https://doi.org/10.1037/pas0000745
Kobylińska, D., & Kusev, P. (2019). Flexible emotion regulation: How situational demands and individual differences influence the effectiveness of regulatory strategies. Frontiers in psychology, 10, 72. https://doi.org/10.3389/fpsyg.2019.0007
Kroenke, K., Spitzer, R. L., Williams, J. B., & Löwe, B. (2009). An ultra-brief screening scale for anxiety and depression: the PHQ–4. Psychosomatics, 50(6), 613–621. https://doi.org/10.1016/S0033-3182(09)70864-3
Lannoy, S., Heeren, A., Rochat, L., Rossignol, M., Van der Linden, M., & Billieux, J. (2014). Is there an all-embracing construct of emotion reactivity? Adaptation and validation of the emotion reactivity scale among a French-speaking community sample. Comprehensive Psychiatry, 55(8), 1960–1967. https://doi.org/10.1016/j.comppsych.2014.07.023
Luminet, O., Fantini-Hauwel, C., Banse, E., Bigot, A., & Preece, D. (2021). French Translation of the Perth Alexithymia Questionnaire. https://doi.org/10.31234/osf.io/r7dft
Lutz, J., & Krahé, B. (2018). Inducing sadness reduces anger-driven aggressive behavior: A situational approach to aggression control. Psychology of Violence, 8(3), 358. https://doi.org/10.1037/vio0000167
Micoulaud-Franchi, J. A., Lagarde, S., Barkate, G., Dufournet, B., Besancon, C., Trébuchon-Da Fonseca, A., Gavaret, M., Bartolomei, F., Bonini, F., & McGonigal, A. (2016). Rapid detection of generalized anxiety disorder and major depression in epilepsy: validation of the GAD-7 as a complementary tool to the NDDI-E in a French sample. Epilepsy & Behavior, 57, 211–216. https://doi.org/10.1016/j.yebeh.2016.02.015
Peh, C. X., Liu, J., Bishop, G. D., Chan, H. Y., Chua, S. M., Kua, E. H., & Mahendran, R. (2017). Emotion regulation and emotional distress: The mediating role of hope on reappraisal and anxiety/depression in newly diagnosed cancer patients. Psycho-Oncology, 26(8), 1191–1197. https://doi.org/10.1002/pon.4297
Rolland, J. P., Parker, W. D., & Stumpf, H. (1998). A psychometric examination of the French translations of NEO-PI-R and NEO-FFI. Journal of Personality Assessment, 71(2), 269–291. https://doi.org/10.1207/s15327752jpa7102_13
Sacharin, V., Schlegel, K., & Scherer, K. R. (2012). Geneva emotion wheel rating study. Center for Person, Kommunikation, Aalborg University, NCCR Affective Sciences. Aalborg University, Aalborg.
Scherer, K. R. (2005). What are emotions? And how can they be measured? Social science information, 44(4), 695–729. https://doi.org/10.1177/0539018405058216
Winterich, P. K., Han, S., & Lerner, S. J. (2010). Now That I’m Sad, It’s Hard to Be Mad: The Role of Cognitive Appraisals in Emotional Blunting. Personality and Social Psychology Bulletin, 36(11), 1467–1483. https://doi.org/10.1177/0146167210384710
Zhan, J., Ren, J., Fan, J., & Luo, J. (2015). Distinctive effects of fear and sadness induction on anger and aggressive behavior. Frontiers in Psychology, 6, 725. https://doi.org/10.3389/fpsyg.2015.00725