Interpersonal Emotion Regulation in Young Adults With Low and High Psychosis Proneness: A Diary Study

IER is defined as a goal-directed process of using interpersonal contact to regulate the intensity, frequency, and/or duration of positive and negative emotions (Zaki & Williams, 2013). IER involves seeking out others to influence one’s own emotions (i.e., intrinsic IER; Zaki & Williams, 2013) and attempts to regulate other people’s emotions (i.e., extrinsic IER). Thus, contrary to related constructs such as attachment, social support, or emotional contagion, IER involves specific self-focused or other-focused emotion modification goals (Williams et al., 2018; Zaki & Williams, 2013). To achieve these goals, IER uses social means and, therefore, depends on the availability of social contacts and specific interaction dynamics (Petrova & Gross, 2023). In the case of intrinsic IER, these dynamics come in varying levels of feedback from the regulating others, ranging from response-independent processes that require the mere presence of others to response-dependent processes aimed at receiving supportive behaviors or validation from one’s interaction partner (Williams et al., 2018). In contrast, intrapersonal emotion regulation (ER) entails strategies that individuals use on their own to manage their emotions (Lincoln et al., 2022; Petrova & Gross, 2023).

Based on the described IER framework, Hofmann et al. (2016) developed one of the first methods for assessing intrinsic IER, the Interpersonal Emotion Regulation Questionnaire (IERQ). In a series of four studies, the authors demonstrated that the IERQ assesses four distinct intrinsic IER strategies. The first strategy (enhancement of positive affect) targets upregulating positive emotions and includes seeking the presence of others when in a positive affective state (e.g., for sharing the experience with others). The remaining three strategies are aimed at downregulating negative emotions and entail asking others about their opinion (perspective taking), imitating their way of dealing with emotional events (social modeling), or seeking their compassion (soothing). This set of strategies has been validated across various languages and cultures (Ding et al., 2021; Koç et al., 2019; Pruessner et al., 2020).

The relevance of IER strategies in everyday life has been highlighted by recent research revealing that more than two-thirds of daily ER episodes occur within interpersonal contexts, such as among friends, colleagues, family members, or romantic partners (Paul et al., 2023).

Importantly, frequently using IER to regulate one’s emotions has been linked with various mental health outcomes (e.g., Chan & Rawana, 2021; Dixon-Gordon et al., 2018; Hofmann, 2014; Marroquín, 2011). Most of the available studies indicate that higher use of IER to enhance one’s positive emotions has beneficial mental health outcomes. It is related to lower anxiety, depression, stress, and negative affect (Chan & Rawana, 2021; Koç et al., 2019; Pruessner et al., 2020) and to higher well-being, positive affect, and relationship satisfaction (Chan & Rawana, 2021; Mueller et al., 2024; Pruessner et al., 2020).

In contrast, using IER strategies to decrease negative emotions has revealed less consistent links with psychological outcome measures. Higher reliance on others to regulate negative emotions has been proposed to indicate that individuals are less effective in using their own personal regulatory strategies (Hofmann et al., 2016). In line with this assumption, previous studies found that using IER strategies to regulate negative emotions, including soothing, social modeling, and perspective taking, is linked to higher levels of anxiety, depression, and substance abuse (Hofmann et al., 2016) and lower well-being (Chan & Rawana, 2021). However, other studies suggest that employing perspective taking as an IER strategy in the context of negative emotions is linked with lower levels of psychological distress (Akkuş & Peker, 2022; Altan-Atalay & Saritas-Atalar, 2022).

Research on how IER relates to psychotic symptoms has been limited. Aspects of IER in psychotic disorders have been analyzed in only a few studies, primarily using ecological momentary assessment (EMA) methodologies (Ludwig et al., 2020; Nittel et al., 2018; Visser et al., 2018). These studies found that using the broadly defined IER strategy of social sharing (i.e., talking with someone about an emotional event) more frequently was related to higher state levels of paranoia (i.e., a strong suspicion that others intend harm to oneself) and negative emotions at any given point in time (Nittel et al., 2018). However, talking with someone about emotional events did not predict changes in paranoia over time (Ludwig et al., 2020; Nittel et al., 2018). This putatively increased use of IER strategies in paranoid states aligns with the generally heightened use of ER strategies that has been noted in psychotic disorders (Lincoln et al., 2022; Visser et al., 2018). The tendency may reflect a greater need for regulation due to the elevated negative affect associated with distressing symptoms such as paranoia (e.g., Nittel et al., 2018). However, these findings also highlight the limited efficacy of IER in alleviating paranoid symptom distress. Additional research has shown that people with elevated levels of positive schizotypal traits (i.e., attenuated trait-level positive symptoms) experience a desire to be alone when with others, mediated by anxiety (Kwapil et al., 2012). This suggests that others may not only fail to alleviate positive psychotic symptom distress but may even exacerbate it. Furthermore, people with high schizotypal traits have been found not to show as much of an increase in positive affect when socializing as those with low schizotypal traits (Minor et al., 2020).

Even less is known about how IER may relate to the negative symptoms associated with psychotic disorders (i.e., amotivation, avolition, and diminished emotional expression). There is reason to assume these symptoms could be related to diminished IER strategy use. Reduced social drive is a core part of the amotivation in psychotic disorders (Marder & Galderisi, 2017), at-risk populations (Strauss et al., 2020), and people with negative schizotypal traits (Kwapil et al., 2012, 2020). This social disinterest is presumably caused by an expectation that social contact will not be beneficial (Jeganathan & Breakspear, 2021; White et al., 2013) or even stressful (Silvia & Kwapil, 2011) and unsurprisingly leads to a diminished initiation of social contact (e.g., Pillny et al., 2020). This likely includes diminished initiation of IER. Moreover, because diminished emotional expression impedes social interactions (e.g., Riehle & Lincoln, 2017; Riehle et al., 2018), people who experience negative symptoms can be expected to perceive the efficacy of IER strategies as relatively low.

In sum, the limited available evidence suggests that psychotic symptoms are likely related to aberrant use and limited efficacy of IER strategies. Underuse of IER may be an important candidate mechanism to explain how social dysfunction leads to psychosis. If an individual cannot access other people as a helpful resource to deal with distressing feelings, they may resort to less functional alternatives which may perpetuate or aggravate attenuated psychotic symptoms. For example, another person might alleviate the distress caused by an initial suspicious thought by offering convincing alternative explanations of a situation (IERQ strategy perspective taking). Without such effective IER, an individual may resort to avoidance behavior that does not challenge the suspicious thought which over time could aggravate into paranoia. However, it is currently unknown whether such difficulties in IER are present in people at risk for psychosis.

It is worth noting that both positive and negative symptoms map onto a continuum from low to high severity in the general population (van Os et al., 2009). Moreover, attenuated positive and negative symptoms are an important risk marker for psychotic disorders, particularly in young adults (Howes & Murray, 2014). We therefore investigated IER strategy use in a sample of young adults (ages 18–29) with low and high psychosis proneness. More specifically, we compared a group with attenuated positive and a group with attenuated negative symptoms to young adults with low psychosis proneness. Since both the frequency and efficacy of IER strategies have been demonstrated to be also modulated by day-to-day contextual factors, such as relationship closeness or emotions to be regulated (Paul et al., 2023; Springstein et al., 2022), we used daily diaries to assess IER strategy use and positive and negative symptoms over one week. We hypothesized that young adults with attenuated negative symptoms would use fewer IER strategies and that young adults with attenuated positive symptoms would use more IER strategies in their daily lives than people with low psychosis proneness. We also hypothesized that young adults with attenuated negative and positive psychotic symptoms would rate using IER strategies in their daily lives as less helpful than people with low psychosis proneness. Furthermore, we expected daily fluctuations in symptoms to predict daily IER use and efficacy in addition to the hypothesized group-level differences.

After consenting to participate, participants provided demographic data and completed a baseline assessment. Participants who fulfilled the inclusion criteria in the baseline assessment were sent an individualized diary link every day for the seven subsequent days. Diary assessments could be completed between 5 p.m. and midnight each day. We selected a daily diary approach over a more fine-grained EMA design as previous findings suggest a relatively infrequent occurrence of IER in everyday life (Liu et al., 2021; Tran et al., 2023) and as this approach enabled us to use a comprehensive measure of IER (a diary-adapted version of the IERQ) rather than a brief assessment as is often necessary in EMA studies. Moreover, this decision was guided by the consideration that daily diaries increase the accessibility and reduce the intrusiveness of the assessment protocol compared to EMA.

The participants were young adults (age 18–29) who received course credit or partook in a lottery with a chance of winning 25€. Besides age, other inclusion criteria were sufficient German skills (to be able to understand the study material), internet and email access for the one-week participation period, answering correctly to all four attention check items in the baseline assessment, and fulfilling the inclusion criteria of one of the three study groups: psychosis-prone – negative symptoms (PP-NES), psychosis-prone – positive symptoms (PP-POS), and low psychosis proneness (control group, CG). For group allocation, we derived cut-offs for positive and negative psychotic symptom scores on the Community Assessment of Psychic Experiences (CAPE; Stefanis et al., 2002, cape42.homestead.com) a priori from a German community validation sample (Schlier et al., 2015). PP-NES had to score above the 90th percentile in negative symptoms (> 1.40) and PP-POS above the 90th percentile in positive symptoms (> 0.74). Because we were interested in symptom-specific correlates of attenuated negative and positive psychotic symptoms, PP-NES had to score below the 90th percentile in positive symptoms and PP-POS below the 90th percentile in negative symptoms. CG had to score below the median on negative and positive psychotic symptoms (< 0.90 and < 0.38, respectively). These specific cut-off scores on the high end have not been validated for their potential to indicate a psychosis risk state. However, given that they surpass the mean levels found for people with psychotic disorders (Jaya et al., 2021), they can be considered as sufficiently indicative of relevant symptom burden. Additionally, because participants in the two attenuated symptom groups were eligible when scoring up to the 90th percentile of the respective other symptom domains, both groups were expected, on average, to present with risk-carrying symptom burden (Werbeloff et al., 2015).

Participants were recruited via university and other online platforms (such as relevant Facebook groups) and with flyers sent to early psychosis detection centers throughout Germany and Austria.

The baseline assessment was completed by 372 participants, of which 163 (43.8%) fulfilled the inclusion criteria for one of the three study groups, 175 (47.0%) were excluded due to providing symptom ratings above the median in at least one of the two symptom domains (i.e., positive and negative symptoms) but below the 90th percentile in both symptom domains, and 34 (9.9%) were excluded due to providing symptom ratings above the 90th percentile in both symptom domains. The diary study was completed by 127 participants, of which 109 (the final sample; n_PP-NES = 37, n_PP-POS = 20, n_CG = 52) provided at least three diary entries (14% attrition, no group differences, χ2(2) = 1.83, p = 0.401).

The adherence rate for the diaries was high, with 14% missing completely (i.e., 107 diaries). For 22 additional diaries, state paranoia ratings were missing due to participants not completing the diary (the Brief Paranoia Checklist was the last page). Additional missings occurred for the four Diary-IERQ efficacy scales when participants had Diary-IERQ frequency values of 0 on any given frequency scale and thus were not prompted to provide an efficacy rating (9–27% of the diaries depending on the respective Diary-IERQ scale). Regarding group differences in adherence, on average, PP-NES and PP-POS provided roughly one diary entry (thus, significantly) less than CG (PP-NES: M = 5.76, SD = 1.52; PP-POS: M = 5.50, SD = 1.40; CG: M = 6.40, SD = 0.98; F(2,44.41) = 5.09, p = 0.010).

Further sample characteristics and results of the baseline assessment are presented in Table 1. As can be seen there, the three groups did not differ on any demographic variables assessed. The group differences in CAPE negative and positive psychotic symptoms reflected the group allocation procedure. Additionally, PP-NES scored significantly higher on CAPE depression than both other groups, whereas PP-POS scored higher on CAPE depression than CG. Moreover, the PP-NES group reported less habitual use of the IER strategies enhancement of positive affect and soothing than the CG. Furthermore, we found a main effect of group on the habitual use of IERQ perspective taking, but the pair-wise comparisons did not yield significant group differences. PP-NES also reported higher levels of emotional suppression in the ERQ than both other groups.

Our finding of a lower frequency of using IER strategies to enhance or sustain positive emotions in negative symptoms aligns with these two previous studies (Gunn & Donahue, 2022; Schwartz et al., 2022). The asociality (i.e., low social motivation and social withdrawal) that is an essential aspect of negative symptoms in daily life (e.g., Kwapil et al., 2012; Pillny et al., 2020) appears to be an obvious explanation for these findings. Young adults with negative symptoms would not engage in as many social interactions due to a lack of social interest and, therefore, report a lower frequency of using IER. Another explanation, however, may be the diminished experience of positive affect (i.e., anhedonia) in this group that we found in line with previous research (e.g., Kwapil et al., 2012; Schwartz et al., 2022). Because the diminished positive affect explained away the group difference in IER strategy use between the attenuated negative symptoms and the other two groups, we cannot rule out that the reduced use of IER strategies for positive emotions in negative symptoms is primarily contingent on a diminished regulatory need. However, our design did not allow for an analysis of the causal direction of the associations of affect and IER strategy use, leaving this an interesting avenue for future research.

Concerning the frequency of using IER strategies for negative emotions, our results indicated that young adults who experience attenuated negative symptoms appear to have someone guide them on how to deal with distressing emotions less frequently (social modeling) and seem to less frequently seek out others for comfort and sympathy (soothing). An important specification of these effects was that they were the strongest when we controlled for the elevated negative affect levels reported by this group. Under the assumption that negative affect precedes IER strategy use, this could be interpreted as a mismatch between the regulatory need for negative affect (high) and frequency of IER strategy use (moderate) in negative symptoms. Such a mismatch has recently been found for intrapersonal ER strategies in a clinical psychosis sample (Raugh & Strauss, 2022). Gunn and Donahue (2022) found an association in the habitual use of the two IER strategies, social modeling and soothing, with negative schizotypal traits when other schizotypal traits were accounted for in the analysis, but did not find associations when these other schizotypal traits were not accounted for. These findings could be consistent with the mismatch hypothesis when one assumes that controlling other schizotypal traits may be a proxy for controlling for elevated negative affect levels. However, replication of these findings focusing on this question is warranted – preferably in an ecological momentary assessment study where the causal direction of this relationship can be tested (i.e., people with attenuated negative symptoms respond to elevations of negative affect less likely with IER).

From a theoretical point of view, the reluctance of people with negative symptoms to engage in IER for distressing emotions could be explained by low expectations that others will react empathic or sympathetic to their displays of vulnerability (Jeganathan & Breakspear, 2021; Riehle, (in press); White et al., 2013). Further fueling this reluctance could be self-ascribed difficulties in describing emotions to others that have been found at an equal level both in schizophrenia and clinical high-risk samples (e.g., Kimhy et al., 2016). Taking IER strategy use into consideration, therefore, sheds new light on several previous findings, replicated by our study, showing that using emotional suppression more is associated with negative symptoms (Kimhy et al., 2016; Larson et al., 2020; Osborne et al., 2017; van der Velde et al., 2015; van Rijn et al., 2011). Negative experiences with using IER and/or disclosing feelings to others and self-ascribed difficulties in verbalizing emotional states each could form a vicious cycle with emotional suppression. Inquiring these processes further may open up an interesting avenue to better understand etiological paths to diminished emotional expression as a negative psychotic symptom.

It is also worth noting that the associations between negative symptoms and reduced IER strategy use to alleviate negative emotions are not necessarily indicative of impaired ER. For example, there have been mixed findings as to whether increased use of emotional suppression is related to lower social functioning levels in schizophrenia and high-risk samples (cf., Chapman et al., 2020; Kimhy et al., 2016; Lincoln et al., 2022). In addition, the results of our analyses of day-to-day changes in symptoms and IER strategy use (within-person effects) suggested that irrespective of the participant group, everyone was less inclined to use IER strategies on days where they experienced higher negative symptoms (i.e., motivational impairments). It could be considered particularly functional not to rely on IER on days with low motivation to seek out contact with others. Here, flexibility in strategy selection may be a key issue (Aldao et al., 2015; Bonanno & Burton, 2013; Lincoln et al., 2022). Interestingly, one study (Bortolon et al., 2022) found that attenuated negative symptoms were related to difficulties in enhancing emotional expression to provide extrinsic IER (i.e., accommodating the feelings of others). The ER flexibility framework has yet to be applied to the relationship between psychotic symptoms and the intrinsic IER strategies considered in our study.

An unresolved question is why negative symptoms specifically seem to affect IER strategies such as social modeling and soothing, but not perspective taking. Given that perspective taking was reported with the lowest frequency of all IER strategies, we cannot rule out that a floor effect may have prevented relevant group differences or symptom associations from showing up. However, since Gunn and Donahue (2022) even identified a positive association between negative schizotypy and perspective taking, future research should examine these contrasting effects.

The attenuated negative symptoms group also reported a lower perceived efficacy of using IER strategies for positive and negative emotions. An important implication of this finding is that the low perceived efficacy likely fuels a vicious cycle of negatively appraised social interactions, social disinterest, and withdrawal (Jaya et al., 2022; Jeganathan & Breakspear, 2021; Riehle, in press; White et al., 2013). Additionally, negative symptoms, particularly diminished expression, appear to be specifically related to more negative interpersonal experiences (Jeganathan & Breakspear, 2021; Riehle, in press; Riehle et al., 2018), and this also at subclinical symptom levels (Riehle & Lincoln, 2017). Accordingly, difficulties with expressing emotions could lead to misunderstandings or limited empathic concern in others, who then may not provide helpful responses. This could explain why attempts at using IER are rated as less helpful by those with negative symptoms. Interestingly, the study by Schwartz et al. (2022) suggested intact efficacy of IER for positive emotions in negative schizotypal traits. Of note, whereas Schwartz and colleagues (2022) assessed time-lagged changes in affect from one assessment to the next in their EMA (prospective indirect change measure), we asked for the perceived efficacy of the specific IER strategies that the participants had used throughout the day (retrospective direct change measure). It could, therefore, be that Schwartz and colleagues (2022) detected more short-term, momentary changes in affect that are not impaired by negative symptoms. The diminished efficacy of using IER strategies reported by the attenuated negative symptoms group could then be interpreted as negatively biased by negative beliefs about the pleasantness of social interactions. Such beliefs are common in negative symptoms (Pillny et al., 2018; Yang et al., 2018) and an important aspect of the cognitive conceptualization of negative symptoms (Beck et al., 2011). A more fine-grained EMA approach would be helpful to address these mixed findings further, as would be studies that assess subjective (e.g., perceived IER efficacy, affect changes) and objective (e.g., psychophysiological changes) markers of using IER.

Our results concerning the attenuated positive symptoms group showed that this group turned to others for advice in distressing situations (perspective taking) more frequently than both other groups and also reported higher efficacy of this strategy. We found group differences for the habitual IERQ and the Diary-IERQ. Furthermore, irrespective of group, within-person effects showed that all IER strategies for alleviating negative emotions (i.e., soothing, social modeling, perspective taking) were used more frequently on days on which people felt more paranoid. In line with this, Gunn and Donahue (2022) found significant positive associations between positive schizotypal traits and those same three IER strategies assessed on a habitual level. Our results extend these findings to the everyday context of our diary study. This suggests that people in paranoid states frequently turn to others to deal with distressing emotions. Due to the strong association between negative affect and paranoia (e.g., Kramer et al., 2013; Krkovic et al., 2019; Kwapil et al., 2012, 2020), paranoia may come along with an additional need to regulate negative affect that people meet with increased IER just as they meet it with increased intrapersonal ER (Lincoln et al., 2022; Visser et al., 2018). Indeed, when controlling for negative affect levels, those associations largely disappeared or were notably reduced (but see Gunn & Donahue, 2022, where these effects remained stable after controlling for other schizotypal traits). Whereas it may seem counterintuitive that elevated levels of suspiciousness and paranoia would lead to increased IER, it is important to consider that the IERQ does not assess whether someone engages in IER with many different people, just a few, or even with only a single trusted person. Previous studies have suggested that most IER takes place with close others (Geiger et al., 2023; Liu et al., 2021), and it is a viable hypothesis to test whether the circle of people that are turned to for IER is smaller in people with positive psychotic symptoms.

We also found that daily paranoia levels were associated with a lower perceived efficacy of using IER to enhance positive affect. As for negative symptoms, this may be due to specific biases in the retrospective report potentially driven by paranoid rumination. However, we cannot disentangle whether this explains our findings, whether attempts at using IER for positive emotions when in a more paranoid state fail more readily, or whether experiencing failed attempts at using IER for positive emotions increases paranoia levels. Nevertheless, our study identifies these questions as important avenues for future research.

The contrasting use of IER strategies in attenuated negative and positive psychotic symptoms highlights the importance of considering interpersonal regulation processes as specific markers not only of different subtypes of psychosis proneness but also of different aspects of psychopathology per se. A growing body of research suggests that IER is affected differently in different forms of psychopathology (Barthel et al., 2018; Dixon-Gordon et al., 2018; Hofmann, 2014; Hofmann & Doan, 2018). For instance, depression has been found to be associated with decreased IER, whereas anxiety shows an opposite pattern with increased reassurance and support-seeking (Beesdo-Baum et al., 2012; Dixon-Gordon et al., 2018; Rector et al., 2019). One may suspect that the discrepancy between these findings and other research suggesting that depression and anxiety are both related to increased use of IER (e.g., Hofmann et al., 2016) may be attributable to variations in the definitions and measurements of IER constructs. In this study, we derived the Diary-IERQ from the IERQ (Hofmann et al., 2016). Nevertheless, we did not find significant effects of depressive symptoms on IER. Our results suggest that amotivational (negative psychotic) but not depressive symptoms are related to decreased use of IER. Future research should aim to disentangle the differential effects of IER use associated with these two symptom domains. For heightened levels of anxiety (e.g., worrying, paranoid thoughts), the data seem to consistently suggest increased use of IER aimed at alleviating negative affective states (e.g., reassurance seeking, seeking advice).

The effect of increased IER use in more anxious participants may also explain why corona-related social distancing was related to more frequent and effective use of perspective taking to downregulate negative emotions. Throughout our data collection (October 2021 – July 2022), meeting friends was largely unrestricted in Germany (Steinmetz et al., 2022). It is possible that at this late stage of the pandemic, more anxious individuals adhered to the remaining social restrictions more (e.g., Matthews et al., 2021). Following these restrictions may have increased contact with close others (e.g., partner, family) compared to non-close others (e.g., colleagues, acquaintances) and thus created additional opportunities for IER (Paul et al., 2023).

How frequently and effectively IER is used is likely a relevant transdiagnostic risk and maintenance factor. Further understanding the role of IER in psychopathology, thus, will likely help develop transdiagnostically effective preventive and therapeutic strategies. Regarding measurement, our diary adaptation of the IERQ showed significant associations with the habitual IERQ. Interindividual differences in habitual IERQ strategies explained 10–46% of the between-subject variation in the Diary-IERQ frequency scores of the same strategy, constituting moderate to large effects. Nevertheless, 53–67% of the total variation in the Diary-IERQ scores was due to within-person (i.e., day-to-day) variation. This suggests that the frequency of IER use is not an entirely stable construct, even though it shows a certain level of intraindividual stability. In our study, such residual within-person variance needed to be expected given that a one-week ecological momentary assessment can include non-representative weeks for some individuals. Moreover, a considerable portion of this within-person variation was explained by day-to-day fluctuations in other measured constructs (i.e., negative symptoms and paranoia), suggesting systematic in addition to random variation. Furthermore, we cannot rule out that rephrasing some of the IERQ items to fit the diary adaptation and to separate frequency from efficacy of strategy use also added to measurement inequivalence.

A limitation of our study design is that it necessitated time-locked analyses, which prevented us from analyzing the causal direction of our effects and introduced potential bias due to retrospective assessments. More fine-grained EMA approaches and experimental investigations are needed to answer some of the questions raised by our study. Another limitation is that our assessment of time spent with others (part of the diary negative symptom assessment) did not include an evaluation of the types of motivation for spending time with others. We cannot disentangle whether our findings, particularly for PP-NES and negative symptoms, suggest that spending little time with others resulted in fewer opportunities for IER or whether not being motivated to use IER resulted in less time spent with others. Additionally, whereas restricting the sample to young adults elevates our findings’ relevance for the understanding of psychosis-risk states, our recruitment strategy attracted a sample for which the generalizability towards non-Western cultures, other age groups, and the lower socio-economic status portion of the population is limited (Henrich et al., 2010). The validity of our approach of preselecting groups of participants with high/low psychotic symptoms is limited as this approach prevented us from investigating the phenomena in question for the entire continuum of psychotic experiences. Moreover, whereas our extreme group approach helped to identify symptom-specific associations, this came at the cost of excluding participants who scored high in both symptom domains, which limits the generalizability of the findings to those with a more severe symptom profile. Additionally, our approach was limited based on our use of a self-report symptom measure (CAPE). Although this approach aligns with the other studies in the field (Gunn & Donahue, 2022; Schwartz et al., 2022), future studies may benefit from interview-based symptom assessments. These are the gold standard, particularly for negative symptoms, both in people at risk for psychosis (Strauss et al., 2020) and in people with schizophrenia (Kirkpatrick et al., 2010; Kring et al., 2013). We also did not assess disorganized symptoms, which could have unique associations with some aspects of IER (Gunn & Donahue, 2022; Kwapil et al., 2020). Another limitation is that we followed an IER definition that assumes an explicit emotion regulation goal (Petrova & Gross, 2023; Zaki & Williams, 2013). This definition also underlies the IERQ (Hofmann et al., 2016) and helps to distinguish IER from related concepts such as attachment, social support, or emotional contagion (Williams et al., 2018; Zaki & Williams, 2013). However, it falls short of potential implicit regulation goals or situations where a social context passively results in emotion regulation but was not actively sought or even intended. Moreover, it would be interesting to explore additional IER processes not encompassed in the current study, such as interpersonal regulatory difficulties (Dixon-Gordon et al., 2018; Williams et al., 2018) or extrinsic interpersonal strategy use (Niven et al., 2011). Finally, it should be considered that the sample size of the psychosis-prone positive symptoms group was relatively small, which may have limited our ability to detect group differences.