Page 1 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 1 1 2 3 1 4 5 6 2 7 8 3 9 10 4 TITLE: Predicting emotional health indicators from linguistic evidence of psychological 11 12 5 distancing 13 14 15 6 16 7 17 8 18 9 SHORT TITLE: Predicting health from linguistic distancing 19 10 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Stress and Health Page 2 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 2 1 2 3 11 Abstract 4 5 6 12 Psychological distancing, a form of cognitive reappraisal, involves construal of emotionally- 7 8 13 valenced stimuli in an objective manner, or with perceived spatial and temporal distance. Prior 9 10 14 work suggests that in appropriate contexts, reappraisal broadly, and distancing specifically, is 11 12 15 related to adaptive mental and physical health outcomes. Additionally, recent research suggests 13 14 15 16 that shifting language to be more distant (i.e., linguistic distancing (LD)) can have adaptive 16 17 17 emotion regulatory effects. The present study addressed whether LD is also associated with 18 19 18 adaptive health indicators.For Participants Review transcribed their Only thoughts while viewing negative or neutral 20 21 22 19 stimuli in one of three ways: 1) by implementing objective language, 2) by implementing spatially 23 24 20 and/or temporally far away language, 3) or by responding naturally. Across psychological 25 26 21 distancing groups, LD was associated with lower negative affectivity (lower perceived stress and 27 28 29 22 depression symptoms), better general well-being (better emotional well-being and energy and 30 31 23 vitality), and better emotion regulation (ER) (greater reappraisal frequency and fewer difficulties 32 33 24 in implementing ER). Participants who used more LD in the objective group had lower negative 34 35 25 affectivity, better general well-being, and better ER, and those in the far group had better ER. The 36 37 38 26 results reveal linguistic mechanisms underlying ER and its relationship to health indicators, 39 40 27 suggesting future examination of LD interventions. 41 42 28 KEYWORDS: emotion regulation, reappraisal, linguistic, distancing, health 43 44 29 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 3 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 3 1 2 3 30 Introduction 4 31 5 6 32 When confronted with a very unpleasant, stressful situation, sometimes the most adaptive 7 8 33 response is to take a step back and coolly appraise what is happening. Indeed, recent empirical 9 10 34 research has shown how distancing oneself from a negative situation, including by appraising the 11 12 35 situation as an objective, impartial observer, or as spatially or temporally far away, can be an 13 14 15 36 adaptive way to regulate one’s emotions (Denny & Ochsner, 2014; Kross, Ayduk, & Mischel, 16 17 37 2005; Ochsner, Silvers, & Buhle, 2012; Trope & Liberman, 2010). More recently, linguistic 18 19 38 evidence of psychologicalFor distancing Review obtained via analysis Only of expressive writing has been shown 20 21 22 39 to be associated with greater emotion regulation efficacy (Nook, Schleider, & Somerville, 2017). 23 24 40 While a growing body of evidence is beginning to coalesce around the beneficial effects of 25 26 41 distancing in a variety of populations and contexts (Kross & Ayduk, 2017), the question of whether 27 28 29 42 psychologically distanced language is associated with adaptive health indicators is less clear and 30 31 43 represented the focus of the current work; such relationships, if observed, may elucidate 32 33 44 dependencies among language, emotion, and health, and may probe the translational value of 34 35 45 linguistic distancing as an emotion regulation intervention. 36 37 38 46 In general, cognitive change targeting appraisal of an emotional situation is a powerful, 39 40 47 flexible mechanism for emotion regulation. Indeed, a growing body of work has examined the 41 42 48 adaptive impacts of cognitive reappraisal, which involves changing one’s cognitive construal of 43 44 45 49 an emotionally evocative stimulus to alter its emotional impact, on experience, behavior, and 46 47 50 physiology (Gross, 1998b, 2015a), including neural activity (Buhle et al., 2014; Ochsner et al., 48 49 51 2012). A sub-form, or tactic (McRae, Ciesielski, & Gross, 2012) for implementing reappraisal is 50 51 52 52 psychological distancing. According to construal level theory (Trope & Liberman, 2010), 53 54 53 distancing can involve interpreting a negative situation as an objective, impartial observer (i.e., by 55 56 57 58 59 60 Stress and Health Page 4 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 4 1 2 3 54 “taking a step back” and employing self-distancing; Kross & Ayduk, 2017), and/or by interpreting 4 5 6 55 a negative situation as spatially or temporally far away (Trope & Liberman, 2010). In addition, 7 8 56 there is evidence that increasing psychological distance in one domain (e.g., self/social) often leads 9 10 57 to increases in distance in other domains as well (e.g., spatial, temporal; Kross & Ayduk, 2017; 11 12 58 Liberman & Trope, 2008; Trope & Liberman, 2010), and individual studies have often 13 14 15 59 operationalized distancing using instructions that manipulate appraisal in multiple distancing 16 17 60 domains at once (e.g., Denny & Ochsner, 2014). 18 19 61 Recent research hasFor demonstrated Review that shifting languageOnly to be more distant helps regulate 20 21 22 62 negative emotions, as distance decreases negative affect and language can encode distance (Gross, 23 24 63 1998b; Kross & Ayduk, 2017; Trope & Liberman, 2010). In addition, psychological distancing 25 26 64 has demonstrated decreases in physiological and neural indices of affective arousal, such as blood 27 28 29 65 pressure and skin conductance (Ayduk & Kross, 2008; Gross, 1998a; Paret et al., 2011) and 30 31 66 amygdala activity (Ochsner & Gross, 2008; Ochsner et al., 2012; Tamir & Mitchell, 2011). 32 33 67 Linguistic Distancing and Emotion Regulation. Recent work has begun to elucidate the linguistic 34 35 68 mechanisms by which psychological distancing impacts emotion regulation (e.g., by speaking in 36 37 38 69 the third person; Kross & Ayduk, 2017; Kross et al., 2014). In addition, Nook and colleagues 39 40 70 (2017) examined whether distancing one’s language spontaneously regulated one’s emotions after 41 42 71 writing about negative images using psychologically “close” or distant” language in social (i.e., 43 44 45 72 self-related), spatial, and temporal domains. The authors used Pennebaker’s Linguistic Inquiry and 46 47 73 Word Count (LIWC; Pennebaker et al., 2015) to analyze the text responses and computed a 48 49 74 linguistic distancing metric that combined several distancing measures (i.e., first-person singular 50 51 52 75 pronouns, present tense verbs, articles, discrepancy words, and words more than six letters). Nook 53 54 76 and colleagues demonstrated that in all three distancing domains, using distanced language 55 56 57 58 59 60 Page 5 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 5 1 2 3 77 reduced negative affect, and that the instructed use of distanced language in one domain (e.g., 4 5 6 78 physical, social, or temporal) had carryover effects in modulating language use in the other 7 8 79 domains as well. 9 10 80 Further work has linked shifts in pronoun usage to reduced negative affect in distressing 11 12 81 situations. For example, participants who use non-first-person pronouns and one’s own name 13 14 15 82 instead of first-person pronouns while introspecting have been shown to report lower negative 16 17 83 affect (Mehl, Robbins, & Holleran, 2012; Nook et al., 2017; Pennebaker & King, 1999). Further, 18 19 84 Kross and colleagues alsoFor demonstrated Review that non-first-person Only language use was shown to lead 20 21 22 85 people to appraise future stressors as more challenging and less threatening (Kross et al., 2014). 23 24 86 However, this prior work has focused on usage of first-person pronouns in particular rather than a 25 26 87 comprehensive linguistic signature more directly reflective of psychological distancing (Cohn, 27 28 29 88 Mehl, & Pennebaker, 2004; Mehl, Robbins, & Holleran, 2012; Nook et al., 2017). 30 31 89 Emotion Regulation and Health Indicators. Extensive research has indicated that individuals who 32 33 90 are better at regulating their emotions have better physical and mental health outcomes (Aldao, 34 35 91 Nolen-Hoeksema, & Schweizer, 2010; Florin, Freudenberg, & Hollaender, 1985; Greer & Watson, 36 37 38 92 1985; Hu et al., 2014; Verzeletti, Zammuner, Galli, & Agnoli, 2016). For example, emotion 39 40 93 dysregulation has been associated with increased risk for cardiovascular disease (Appleton & 41 42 94 Kubzansky, 2014; Berna, Ott, & Nandrino, 2014). 43 44 45 95 In addition, emotion regulation ability has been shown to predict risk of anxiety, depression, 46 47 96 life satisfaction, and overall general health (Verzeletti et al., 2016). Successful emotion regulation 48 49 97 is associated with increased psychological well-being, including both emotional and social 50 51 52 98 functioning. For example, individuals who use reappraisal habitually show positive health 53 54 99 outcomes, including fewer depressive symptoms, greater self-esteem, and greater life satisfaction 55 56 57 58 59 60 Stress and Health Page 6 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 6 1 2 3 100 (Gross & John, 2003). Further, reappraisal has also been shown to be associated with stronger 4 5 6 101 social relationships (English, John, Srivastava, & Gross, 2012). In addition, for highly stressed 7 8 102 individuals, cognitive reappraisal has been shown to exhibit protective effects, as individuals who 9 10 103 habitually use more reappraisal exhibit less depressive symptoms than individuals who use 11 12 104 reappraisal less (Troy, Wilhelm, Shallcross, & Mauss, 2010). Furthermore, individuals who 13 14 15 105 frequently reappraise report lower levels of stress-related symptoms, compared to other emotion 16 17 106 regulation strategies like suppression (Moore, Zoellner, & Mollenholt, 2008). However, this 18 19 107 literature has not yet exploredFor whether Review these beneficial Only health indicators can be predicted from 20 21 22 108 analysis of expressive writing during emotion regulation alone, which if substantiated may provide 23 24 109 for a relatively quick, unobtrusive means of predicting a variety of health-relevant information. 25 26 110 Emotion Regulation and Expressive Writing. Expressive writing of distressing situations has 27 28 29 111 been shown to be cathartic and lead to sense-making, allowing individuals to gain broader, 30 31 112 objective perspectives, and distance themselves from adverse situations. Recent work has 32 33 113 demonstrated that expressive writing of distressing experiences promotes self-distancing (i.e., 34 35 114 social distancing), as indexed by decreased use of first-person singular pronouns (Park, Ayduk, & 36 37 38 115 Kross, 2016). Relatedly, affective labelling (i.e., putting feelings into words) has been 39 40 116 shown to reduce self-reported negative affect (Lieberman, Inagaki, Tabibnia, & Crockett, 41 42 43 117 2011). Expressive writing demands a certain degree of structure and basic labelling or 44 45 118 acknowledgement of emotions (Lieberman et al., 2007), which may underlie its adaptive emotional 46 47 48 119 effects (Pennebaker & Chung, 2007). 49 50 120 Further, there is evidence of positive health outcomes being associated with expressive 51 52 121 writing about negative situations (Lu & Stanton, 2010; Pennebaker, 1997). For example, 53 54 122 expressive writing is associated with better long-term physical health indexed by illness-related 55 56 57 58 59 60 Page 7 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 7 1 2 3 123 visits to the doctor, blood pressure, lung function, liver function, and number of days in the hospital 4 5 6 124 (Baikie & Wilhelm, 2005). In addition, a rich feature of expressive writing is the translation of 7 8 125 emotional experiences into words. Even in the absence of an instruction to regulate, extensive 9 10 126 verbal processing reduces negative affect and enhances immune function, including T-helper cell 11 12 127 growth, antibody response to Epstein-Barr virus, and antibody response to hepatitis B vaccinations 13 14 15 128 (Pennebaker & Chung, 2007). 16 17 129 Overall, the goal of the present study was to further probe the relationship between linguistic 18 19 For Review Only 20 130 distancing and emotion regulation. In particular, we aimed to examine whether linguistic evidence 21 22 131 of psychological distancing (implemented via either one of two sub-tactics for distancing: 23 24 132 objective, impartial appraisal or via manipulation of spatial and temporal distance) is associated 25 26 133 with adaptive health-related indicators in addition to reduced self-reported negative affect. 27 28 29 134 Addressing these relationships is important in order to further probe the psychological mechanisms 30 31 135 underlying distancing as well as to examine the translational adaptive potential of linguistic 32 33 136 distancing via expressive writing. Notably, we chose to include three groups (one group 34 35 36 137 manipulated to use the objective tactic, the second group manipulated to use the spatial and/or 37 38 138 temporal tactic, and the third group to respond naturally) given prior work that has linked spatial 39 40 139 and temporal aspects of distancing (Nook et al., 2017; Parkinson, Liu, & Wheatley, 2014). We 41 42 43 140 chose to use the standardized linguistic distancing metric to index the degree of language-based 44 45 141 psychological distancing to maximize comparability across groups, given the common linguistic 46 47 142 features underlying the distancing construct. We predicted that self-reported negative affect would 48 49 143 be reduced in both distancing groups and that linguistic distancing would be associated with 50 51 52 144 greater indicators of positive health outside the lab. 53 54 145 Method 55 56 57 58 59 60 Stress and Health Page 8 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 8 1 2 3 146 Participants. Ninety-four healthy adults fluent in English provided written informed consent to 4 5 6 147 participate in the study. Participants were recruited from Rice University’s undergraduate 7 8 148 psychology participant pool and were compensated with academic credit. All 94 participants were 9 10 149 included in the final analyses (67% female; 34% Asian or Pacific Islander, 34% Caucasian, 22% 11 12 150 Hispanic or Latino, 10% African-American, 6% Mixed-Race; age M = 19.37 years, age SD = 1.19 13 14 15 151 years). As described below, participants were randomly assigned to one of three groups (Objective, 16 17 152 Far, Look Only). The final sample included 32 participants in the Objective group (age M = 19.25 18 19 153 years, age SD = 1.01 years),For 31 participants Review in the Far groupOnly (age M = 19.30 years, age SD = 0.99 20 21 22 154 years), and 31 participants in the Look Only group (age M = 18.97 years, age SD = 0.87 years). 23 24 155 All procedures were approved by the Rice University Institutional Review Board. 25 26 156 Materials and Measures 27 28 29 157 Stimuli. Participants viewed neutral or negative digital color images selected from the 30 31 158 International Affective Picture System (IAPS; Lang et al., 1993). Each participant viewed 40 32 33 159 negative pictures that were highly negative (M = 2.15, SD = 1.38, on a scale ranging from 1 to 9 34 35 160 where 9 is completely happy), and highly arousing (M = 6.11, SD = 2.25, on a scale ranging from 36 37 38 161 1 to 9 where 9 is completely aroused). Each participant also viewed 20 neutral images that are 39 40 162 neither negative (M = 6.06, SD = 1.46) nor arousing (M = 3.91, SD = 2.19) from the IAPS database. 41 42 163 Procedure 43 44 45 164 Questionnaire Measures. Following informed consent, all participants first completed 46 47 165 questionnaire measures assessing mental and physical health indicators including negative 48 49 166 affectivity, general health and well-being, and emotion regulation. To assess negative affectivity, 50 51 52 167 participants completed the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 53 54 168 1977), which assesses depressive symptomology during the past week (Cronbach’s � = .85), as 55 56 57 58 59 60 Page 9 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 9 1 2 3 169 well as the Perceived Stress Scale (PSS; Cohen, Kamarck, & Mermelstein, 1983), which assesses 4 5 6 170 the participant’s perceived stress over the course of the past few days (Cronbach’s � = .84). 7 8 171 To assess general health and well-being, participants completed the RAND short-form health 9 10 172 survey (SF-36; Ware & Sherbourne, 1992), which is a 36-item measure containing eight multi- 11 12 13 173 item subscales: 1) physical functioning (Cronbach’s � = .93), 2) role limitations due to physical 14 15 174 health (Cronbach’s � = .84), 3) role limitations due to emotional problems (Cronbach’s � = .83), 16 17 175 4) energy/fatigue (Cronbach’s � = .86), 5) pain (Cronbach’s � = .78), 6) emotional well-being 18 19 176 (Cronbach’s = .90), For 7) social Review functioning (Cronbach’s Only = .85), and 8) general health 20 � � 21 22 177 (Cronbach’s � = .78) (Murdock, Fagundes, Peek, Vohra, & Stowe, 2016; Ware & Sherbourne, 23 24 178 1992). 25 26 179 To assess emotion regulation, participants first completed the Emotion Regulation 27 28 29 180 Questionnaire (ERQ; Gross & John, 2003), which assesses how frequently participants report 30 31 181 regulating their emotions via reappraisal and expressive suppression (Cronbach’s � = .79), as well 32 33 182 as the Difficulties in Emotion Regulation Scale (DERS; Gratz & Roemer, 2004), a widely-used 34 35 36 183 assessment of emotion dysregulation (Cronbach’s � = .93). Please see Supplemental Materials for 37 38 184 additional details. 39 40 185 Task Training. Participants were then randomly assigned to one of three groups: 1) Objective, 2) 41 42 43 186 Far or 3) Look Only. Each group was provided different instructions on how to describe the stimuli 44 45 187 depending on the sub-tactic. In the Objective group, participants were instructed to view the 46 47 188 images as an objective, impartial observer and use language that reflected that appraisal. 48 49 189 Specifically, they were told “when you see the instruction “change,” we would like you to write 50 51 52 190 about that picture using impartial and objective words.” For example, a sample interpretation of 53 54 191 an example negative stimulus provided was “A wounded person is lying on the ground with blood 55 56 57 58 59 60 Stress and Health Page 10 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 10 1 2 3 192 clearly visible.” In the Far group, participants were instructed to use spatially and/or temporally 4 5 6 193 distant language to describe the images that they saw. Specifically, they were told “when you see 7 8 194 the instruction “change,” we would like you to write about that picture using words that make the 9 10 195 image seem physically and temporally far away from you.” For example, a sample interpretation 11 12 196 of an example negative stimulus provided was “A wounded person was lying on the ground a long 13 14 15 197 time ago and far away from me.” Participants in the Look Only group saw all of the same images 16 17 198 and were instructed to simply look and respond naturally to all images and write about them 18 19 199 accordingly. Participants Forthen underwent Review the emotion regulationOnly training procedure depending on 20 21 22 200 the group to which they were randomly assigned. The training procedure lasted for 8-10 minutes 23 24 201 and was conducted at a computer in the laboratory by a trained experimenter using procedures 25 26 202 similar to those used in previous research (Denny & Ochsner, 2014). 27 28 29 203 In particular, participants were told that they would see a series of images preceded by one 30 31 204 of two cues: either the word “LOOK” or the word “CHANGE”. When participants viewed the 32 33 205 word "LOOK," they were instructed to look at the stimulus and respond as they naturally 34 35 206 would. When the participants viewed the word “CHANGE,” they were instructed to use objective, 36 37 38 207 impartial language (Objective group), or instructed to describe the image in terms of increased 39 40 208 spatial or temporal distance (Far group), depending on the group to which they were assigned, 41 42 209 during the upcoming expressive writing period. Thus, there were three trial types: Look Neutral 43 44 45 210 (i.e., the LOOK cue paired with a neutral image), Look Negative (i.e., the LOOK cue paired with 46 47 211 a negative image), and Change Negative (i.e., the CHANGE cue paired with a negative image). 48 49 212 After these instructions were given, four “walk-through” example images were then presented (1 50 51 52 213 example Look Negative trial followed by 3 example Change Negative trials in the Objective and 53 54 214 Far groups; 4 Look Negative trials for the Look Only group), with self-paced timing and the 55 56 57 58 59 60 Page 11 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 11 1 2 3 215 experimenter present. While the image was on the screen, participants were instructed to write 4 5 6 216 about their thoughts and feelings in response to the image while applying the cued instruction 7 8 217 (either “LOOK” or “CHANGE”), following procedures established by Nook and colleagues (Nook 9 10 218 et al., 2017). The experimenter ensured that the participant understood the task directions before 11 12 219 proceeding. 13 14 15 220 Reappraisal Task. Participants then completed the reappraisal task. The task session for all groups 16 17 221 consisted of 60 total trials. Each trial consisted of four phases. First, a fixation cross was presented 18 19 222 for 3 to 5 s (M = 4 s). Second,For a cue Review (“LOOK” or “CHANGE”) Only was presented for 4 s. Then, either 20 21 22 223 a negative or neutral image was presented for 30 seconds. During this time, participants were 23 24 224 instructed to write about their thoughts and feelings in response to the image while applying the 25 26 225 cued instruction, per the training procedures above. The image advanced automatically after 30 s. 27 28 29 226 Finally, for each trial, participants were asked during a 4 s rating period to provide a self-reported 30 31 227 negative affect judgement ("How bad do you feel?") on a scale of 1 (not bad at all) to 7 (extremely 32 33 228 bad). 34 35 229 The task consisted of 20 Look Neutral trials, 20 Look Negative trials, and 20 Change 36 37 38 230 Negative trials. In the Look Only group, there were 40 Look Negative and 20 Look Neutral trials. 39 40 231 Three task runs were presented (consisting of 20 trials per condition per run). Order of runs were 41 42 232 counterbalanced across participants. All stimuli were unique during both the training phase and in 43 44 45 233 the actual task, and negative images were counterbalanced in their assignments to condition (i.e., 46 47 234 LOOK or CHANGE) across participants. At the conclusion of the study, participants were orally 48 49 235 debriefed and thanked for their participation. 50 51 52 236 Data Analysis. To quantify the degree of linguistic distancing employed by each participant on 53 54 237 each trial, we used Pennebaker's Linguistic Inquiry and Word Count (LIWC; Pennebaker et al., 55 56 57 58 59 60 Stress and Health Page 12 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 12 1 2 3 238 2015), which computes percentages of words falling within particular word categories. LIWC has 4 5 6 239 been used extensively in empirical research exploring similar constructs, ranging from 7 8 240 examinations of the adaptiveness of social feedback (Doré, Morris, Burr, Picard, & Ochsner, 9 10 241 2017), to emotion regulation strategies such as self-distancing (Kross & Ayduk, 2008). The LIWC 11 12 242 analyses focused on positive affect words, negative affect words, and a composite linguistic 13 14 15 243 measure of psychological distancing (i.e., using the methodology of Mehl et al., 2012; Nook et al., 16 17 244 2017; Pennebaker & King, 1999). Please see Supplemental Materials for analyses on negative and 18 19 245 positive affect word usage.For In particular, Review the composite Only psychological distancing score was 20 21 22 246 calculated by z-scoring first-person singular pronouns (e.g., I, my), present-tense verbs, articles 23 24 247 (e.g., the, a), discrepancy words (e.g., would, should), and words of more than six letters, which 25 26 248 were then averaged as indicated below. Previous research has indicated that these lexical 27 28 29 249 components best track verbal immediacy and have been used to index linguistic distancing in 30 31 250 previous literature (Mehl, Robbins, & Holleran, 2012; Cohn et al., 2004; Nook et al., 2017; 32 33 251 Pennebaker & King, 1999). Accordingly, the z-scored frequencies of first-person singular 34 35 252 pronouns, present-tense verbs, and discrepancy words were reverse-scored (multiplied by -1) and 36 37 38 253 then averaged with the z-scored frequencies of articles and words of more than six letters for each 39 40 254 trial, per participant and per group. Linguistic distancing was then calculated as the difference 41 42 255 between Change Negative and Look Negative using this metric (i.e., score for Change Negative – 43 44 45 256 score for Look Negative). Lower linguistic distancing scores suggest that writing is personal and 46 47 257 focused on the present, whereas higher linguistic distancing scores show that writing is impersonal 48 49 258 and not focused on the present. 50 51 52 259 We first aimed to conduct manipulation checks to examine whether self-reported negative 53 54 260 affect ratings and composite linguistic distancing scores differed between the Look Negative and 55 56 57 58 59 60 Page 13 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 13 1 2 3 261 Change Negative trials in the Objective and Far groups. We used repeated-measures analyses of 4 5 6 262 variance (ANOVAs) to test for significant differences among the three trial types (Change 7 8 263 Negative, Look Negative, and Look Neutral) and among the groups. Post-hoc t-tests were 9 10 264 conducted to further assess pair-wise differences between the trial types and groups. Then, we 11 12 265 examined correlations between linguistic distancing and health indicators including symptoms of 13 14 15 266 depression (CES-D), perceived stress (PSS), general well-being (all eight subscales of the SF- 16 17 267 36), reappraisal frequency (a subscale of the ERQ), and difficulties in emotion regulation 18 19 268 (DERS). For Review Only 20 21 22 269 Three standardized health composite metrics were created: 1) Negative Affectivity, 2) 23 24 270 General Well-Being, and 3) Emotion Regulation. For each scale, we first created z-scores for 25 26 271 each participant, and then we averaged standardized scores across the following scales for each 27 28 29 272 participant: 1) The Negative Affectivity composite score averaged the standardized Perceived 30 31 273 Stress scores and Symptoms of Depression scores. 2) The General Well-Being composite score 32 33 274 averaged the standardized Energy and Vitality scores, Emotional Well-Being scores, and Social 34 35 275 Functioning scores. 3) The Emotion Regulation composite score averaged the standardized 36 37 38 276 Reappraisal Frequency scores and standardized reverse-scored Difficulties in Emotion 39 40 277 Regulation scores. Once the three new composite scores were created, we computed the 41 42 278 correlations with the composite linguistic distancing metric. 43 44 45 279 Results 46 280 Self-reported negative affect ratings. As a manipulation check, we first assessed the self-reported 47 48 281 negative affect data. There was a main effect of trial type on self-reported negative affect reports, 49 50 282 F(2, 156) = 585.03, p < .001, Cohen’s f = 2.70, and a main effect of group on self-reported negative 51 52 53 283 affect reports, F(2, 98) = 3.25, p = .043, Cohen’s f = 0.25 (Figure S1 and Table S1). As expected, 54 55 284 across the psychological distancing groups, participants reported feeling less negative in the 56 57 58 59 60 Stress and Health Page 14 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 14 1 2 3 285 Change Negative trials relative to the Look Negative trials, t(62) = 3.57, p < .001, Cohen’s d = 4 5 6 286 0.44. To unpack the main effect of group, we did post-hoc pairwise comparisons among the groups; 7 8 287 however, no pairwise group differences were observed (p > .3). Overall, manipulation checks were 9 10 288 successful. Please see Supplemental Materials for additional details. 11 12 289 Composite Linguistic Distancing. Table S1 shows descriptive statistics for each trial type for each 13 14 15 290 group on all lexical component measures as well as composite linguistic distancing scores. There 16 17 291 was a main effect of trial type on composite linguistic distancing scores, F(2, 151) = 35.42, p < 18 19 292 .001, Cohen’s f = 0.68, andFor a main Revieweffect of group on compositeOnly linguistic distancing scores, F(2, 20 21 22 293 94) = 8.54, p < .001, Cohen’s f = 0.43 (Figure 1 and Table S1). As expected, across the 23 24 294 psychological distancing groups (Objective and Far), participants implemented more linguistic 25 26 295 distancing during Change Negative trials relative to Look Negative trials, t(62) = 8.59, p < .001, 27 28 29 296 Cohen’s d = 1.08. To unpack the main effect of group, participants in the Objective group 30 31 297 implemented more linguistic distancing overall relative to the Look Only group, t(61) = 5.15, p < 32 33 298 .001, Cohen’s d = 1.30, and among Look Negative trials in particular (t(61) = 3.25, p = .002, 34 35 299 Cohen’s d = 0.82). Participants in the Far group also implemented more linguistic distancing 36 37 38 300 overall relative to the Look Only group, t(60) = 4.62, p < .001, Cohen’s d = 1.17, and among Look 39 40 301 Negative trials in particular (t(60) = 2.80, p = .007, Cohen’s d = 0.71). However, there were no 41 42 302 significant differences in composite linguistic distancing between the Objective and Far groups for 43 44 45 303 any trial type (all p > .9). Further, composite linguistic distancing did not significantly predict 46 47 304 reappraisal success (i.e. mean self-reported negative affect score for Change Negative relative to 48 49 305 Look Negative trials; Table S2). 50 51 52 306 Association between linguistic distancing and health indicators. 53 54 55 56 57 58 59 60 Page 15 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 15 1 2 3 307 Table S3 shows descriptive statistics by group for each health indicator. There were no 4 5 6 308 pair-wise significant differences among the groups for any of the health indicator questionnaire 7 8 309 measures. No outliers—quantified as more than three IQRs from the hinges—were present in any 9 10 310 variable (i.e., composite linguistic distancing and all health indicator scores). 11 12 311 Before examining the association between linguistic distancing and health-relevant 13 14 15 312 indicators, as a manipulation check, we assessed the degree of correlation of the health indicators 16 17 313 with each other. Many, though not all, of the health indicators above were correlated with each 18 19 314 other in expected directionsFor (Table Review S4). We then examinedOnly associations between linguistic 20 21 22 315 distancing and health indicators in three ways: collapsing across the Objective and Far groups in 23 24 316 order to assess associations with distancing in general, and then within each group alone. Raw data 25 26 317 pertaining to analyses performed in the present study are available at the following link: 27 28 29 318 https://osf.io/8bmsp/?view_only=6574b8ca7c544a8eadc52d7d046fe4cd. 30 31 319 Association between linguistic distancing and health indicators: Collapsing across the Objective 32 33 320 and Far groups. 34 35 321 Negative Affectivity. Collapsing across distancing groups (i.e., Objective and Far), participants 36 37 38 322 with greater linguistic distancing in their transcriptions during reappraisal, indexed by subtracting 39 40 323 the Look Negative linguistic distancing metric from the Change Negative linguistic distancing 41 42 324 metric, reported diminished amounts of negative affectivity (Figure 2 and Table 1). Consistent 43 44 45 325 with hypotheses, there was a significant negative association between linguistic distancing and the 46 47 326 negative affectivity composite score, r(61) = –.30, p = .017, 95% CI [–.511, –.057]. There was a 48 49 327 significant negative association between linguistic distancing and perceived stress as measured by 50 51 52 328 the PSS (M = 17.5, SD = 7.7), r(61) = –.31, p = .013, 95% CI [–.518, –.068] (Figure 2A), and 53 54 55 56 57 58 59 60 Stress and Health Page 16 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 16 1 2 3 329 symptoms of depression as measured by the CES-D (M = 15.0, SD = 11.3), r(61) = –.27, p = .035, 4 5 6 330 95% CI [–.482, –.019] (Figure 2B). 7 8 331 General Well-Being. In addition, individuals with greater linguistic distancing reported greater 9 10 332 general well-being as measured by the SF-36. There was a significant positive association between 11 12 333 linguistic distancing and the general well-being composite score, r(61) = .28, p = .029, 95% CI 13 14 15 334 [.030, .490]. There was a significant positive association of linguistic distancing and emotional 16 17 335 well-being (M = 67.2, SD = 17.6), r(61) = .29, p = .021, 95% CI [.046, .502] (Figure 2C), and a 18 19 336 marginal positive associationFor of linguistic Review distancing and Only energy and vitality (M = 49.2, SD = 15.5), 20 21 22 337 r(61) = .23, p = .066, 95% CI [–.016, .454] (Figure S2). 23 24 338 Emotion Regulation. Finally, individuals with greater linguistic distancing reported better 25 26 339 emotion regulation. There was a significant positive association between linguistic distancing and 27 28 29 340 the emotion regulation composite score, r(61) = .33, p = .009, 95% CI [.085, .531]. There was a 30 31 341 significant negative association of linguistic distancing and difficulties in emotion regulation (i.e., 32 33 342 DERS scores) (M = 76.9, SD = 19.4), r(61) = –.29, p = .022, 95% CI [–.500, –.043] (Figure 2D), 34 35 343 and a positive association with reappraisal frequency (i.e., ERQ reappraisal scores) (M = 30.6, SD 36 37 38 344 = 5.9), r(61) = .27, p = .031, 95% CI [.026, .487] (Figure S2), respectively. 39 40 345 Results were broadly consistent when using a regressed change approach, which involved 41 42 346 implementing general linear models with linguistic distancing during Change Negative trials as 43 44 45 347 the independent variable, linguistic distancing during Look Negative trials as the covariate, and 46 47 348 each health index as the dependent variable, although these analyses did not reach significance 48 49 349 thresholds for negative affectivity and emotion regulation measures. Please see Supplemental 50 51 52 350 Materials for more details. 53 54 351 Association between linguistic distancing and health indicators: Objective Group. 55 56 57 58 59 60 Page 17 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 17 1 2 3 352 Negative Affectivity. We next assessed these relationships within each group in order to probe 4 5 6 353 differences according to different distancing sub-tactics (Figure S3). Within the Objective group, 7 8 354 participants with greater linguistic distancing reported overall greater mental health. There was a 9 10 355 there was a significant negative association between linguistic distancing and the negative 11 12 356 affectivity composite score, r(30) = –.45, p = .010, 95% CI [–.690, –.120]. Greater linguistic 13 14 15 357 distancing within this group showed a significant negative association with symptoms of 16 17 358 depression (i.e., CES-D scores) (M = 14.1, SD = 12.8), r(30) = –.44, p = .011, 95% CI [–.111, – 18 19 359 .685], and perceived stressFor (i.e., PSS Review scores) (M = 16.7, Only SD = 8.5), r(30) = –.43, p = .015, 95% CI 20 21 22 360 [–.675, –.092]. 23 24 361 General Well-Being. Individuals with greater linguistic distancing reported greater general well- 25 26 362 being (i.e., SF-36 scores). There was a significant positive association between linguistic 27 28 29 363 distancing and the general well-being composite score, r(30) = .41, p = .018, 95% CI [.076, .667]. 30 31 364 There was a significant positive association with emotional well-being (M = 66.8, SD = 19.7), 32 33 365 r(30) = .41, p = .019, 95% CI [.073, .665], a significant positive association with social functioning 34 35 366 (M = 82.0, SD = 16.2), r(30) = .36, p = .046, 95% CI [.007, .626], and a significant positive 36 37 38 367 association with energy and vitality (M = 51.3, SD = 17.5), r(30) = .35, p = .048, 95% CI [.004, 39 40 368 .624]. 41 42 369 Emotion Regulation. There was no significant association between linguistic distancing and the 43 44 45 370 emotion regulation composite score, r(30) = .25, p = .17, 95% CI [–.108, .551]. Further, there was 46 47 371 a marginal negative association between linguistic distancing and difficulties in emotion regulation 48 49 372 (i.e., DERS scores) (M = 75.3, SD = 20.4), r(30) = –.32, p = .079, 95% CI [–.598, .038]. 50 51 52 373 Association between linguistic distancing and health indicators: Far Group. 53 54 55 56 57 58 59 60 Stress and Health Page 18 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 18 1 2 3 374 Emotion Regulation. Within the Far group (Figure S4), there was a significant positive association 4 5 6 375 between linguistic distancing and the emotion regulation composite score, r(29) = .44, p = .013, 7 8 376 95% CI [.101, .687]. Participants with greater linguistic distancing showed a significant positive 9 10 377 association with reappraisal frequency as measured by the ERQ (M = 29.9, SD = 6.1), r(29) = .45, 11 12 378 p = .010, 95% CI [.118, .696], and a marginal negative association with difficulties in emotion 13 14 15 379 regulation (i.e., DERS scores) (M = 78.6, SD = 18.5), r(29) = –.31, p = .091, 95% CI [–.598, .051]. 16 17 380 No other relationships were significant (Table 1). 18 19 381 We further quantitativelyFor Review examined differences Only in correlation magnitude between the 20 21 22 382 Objective and Far groups between linguistic distancing and each health indicator, and no 23 24 383 correlation differences were significant. There were also no significant relationships between 25 26 384 linguistic distancing and each of the remaining subscales of the SF-36 (i.e., physical functioning, 27 28 29 385 role limitations due to physical health, etc.) in any of the groups. 30 31 386 Discussion 32 33 387 Summary of Results. 34 35 388 In this study, we sought to examine the relationship between expressive writing, emotion 36 37 38 389 regulation, and validated questionnaire measures of overall health. We found that when individuals 39 40 390 are manipulated to use language reflective of psychological distancing, positive relationships were 41 42 391 present between linguistic distancing and several indicators of health and wellness. Across both 43 44 45 392 psychological distancing groups, linguistic distancing was associated with lower negative 46 47 393 affectivity (lower perceived stress and symptoms of depression), better general well-being (better 48 49 394 emotional well-being and energy and vitality), and better emotion regulation (greater reappraisal 50 51 52 395 frequency and fewer difficulties in emotion regulation). When individuals were manipulated to 53 54 396 specifically use objective language, linguistic distancing was associated with lower negative 55 56 57 58 59 60 Page 19 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 19 1 2 3 397 affectivity (lower symptoms of depression and perceived stress), better general well-being (better 4 5 6 398 emotional well-being, social functioning, and energy and vitality), and better emotion regulation 7 8 399 as well (fewer difficulties in emotion regulation). When individuals were manipulated to 9 10 400 specifically use far language, linguistic distancing was associated with better emotion regulation 11 12 401 (greater reappraisal frequency and fewer difficulties in emotion regulation). 13 14 15 402 Though correlational, the present study provides additional insight into the mechanisms 16 17 403 connecting emotion regulation, language, and psychological and health-related indicators. Nook 18 19 404 and colleagues showed thatFor spontaneously Review engaging inOnly linguistic distancing during reappraisal 20 21 22 405 (i.e., after having been given instructions to reinterpret the meaning of a negative image rather than 23 24 406 engage in distancing) was positively associated with increasing reappraisal success (defined as 25 26 407 greater reductions in self-reported negative affect due to reappraisal). While self-reported negative 27 28 29 408 affect ratings did indicate significant reappraisal success overall in the present study, the lack of 30 31 409 significant relationships between linguistic distancing and reappraisal success in the current results 32 33 410 may be due to several factors, including potentially greater sensitivity and less restrictive range in 34 35 411 the measured health indicators relative to reappraisal success. Future work with larger samples 36 37 38 412 may further examine relationships between linguistic distancing and reappraisal success. Notably, 39 40 413 the data presented are primarily informative about the existence and direction of the relationships 41 42 414 more so than the magnitudes of the correlations. Future work involving larger samples may provide 43 44 45 415 more reliable estimates of correlation effect sizes. 46 47 416 Links to Previous Research. 48 49 417 In addition, the present results suggest that incorporating lexical shifts reflective of 50 51 52 418 distancing may be adaptive. While previous research has elucidated linkages among explicit 53 54 419 reappraisal frequency across multiple tactics and some adaptive health indicators, such as 55 56 57 58 59 60 Stress and Health Page 20 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 20 1 2 3 420 depression, life-satisfaction, and well-being (Gross & John, 2003), our results indicate for the first 4 5 6 421 time that overall linguistic distancing in particular is associated with several adaptive health 7 8 422 indicators. Thus, the present work builds on an emerging body of research on the adaptiveness of 9 10 423 distancing as a tool in one’s emotion regulation repertoire (Ayduk & Kross, 2008; Kross & Ayduk, 11 12 424 2008, 2017). 13 14 15 425 The present results are also relevant to prior work on expressive writing. Substantiating the 16 17 426 promise of expressive writing as a means to regulate emotion, a meta-analysis of 146 studies on 18 19 427 experimental disclosure foundFor that disclosingReview thoughts andOnly feelings about personal and meaningful 20 21 22 428 topics is beneficial for psychological and physical health (Frattaroli, 2006). Other examinations of 23 24 429 the efficacy of expressive writing interventions on improving health outcomes have yielded 25 26 430 divergent conclusions, however. A meta-analysis of 64 randomized studies on the effect 27 28 29 431 of expressive writing on depressive symptoms found that brief, self-directed expressive writing is 30 31 432 not efficacious in reducing depressive symptoms among healthy adults with varying levels 32 33 433 of stress (Reinhold, Bürkner, & Holling, 2018). However, the present results suggest that focused, 34 35 434 task-driven expressive writing specifically reflective of psychological distancing may be 36 37 38 435 associated with positive health outcomes. Overall, these results suggest that expressive writing, 39 40 436 and the emotional granularity and complexity that this entails, may be an important linguistic 41 42 437 mechanism underlying some of the adaptive effects of emotion regulation (and reappraisal, in 43 44 45 438 particular) that drive health benefits in general (Kashdan, Barrett, & McKnight, 2015).” 46 47 439 Further, a recent quantitative meta-analysis on distancing proposes a taxonomy of the 48 49 440 subordinate distancing tactics; the authors outline how distancing can be in the form of (1) space, 50 51 52 441 (2) time, or (3) objectivity (Powers & LaBar, in press). In the present study, we chose to 53 54 442 investigate space and time within one group and objectivity in the second group given prior work 55 56 57 58 59 60 Page 21 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 21 1 2 3 443 that has linked spatial and temporal aspects of distancing. In particular, previous research that 4 5 6 444 separately manipulated spatial and temporal distance (Nook et al., 2017; Parkinson, Liu, & 7 8 445 Wheatley, 2014) found evidence for “spillage” or “bleed over” between the two tactics; thus, the 9 10 446 authors have argued that there is evidence that spatial and temporal distancing reflect a more 11 12 447 common underlying core distancing construct (cf. Kross & Ayduk, 2016). Other research has 13 14 15 448 also investigated spatial and temporal distance and found evidence of a common underlying 16 17 449 construct for distancing (Doré, Ort, Braverman, & Ochsner, 2015). Therefore, while construal 18 19 450 level theory posits that indeedFor there Review is a common distancing Only construct underlying all of spatial, 20 21 22 451 temporal, and objective distancing (Trope & Liberman, 2010), for efficiency’s sake, in this 23 24 452 study, we combined spatial and temporal distancing into one group to compare versus objective 25 26 453 distancing. In the present study, while spatial/temporal and objective distancing did not result in 27 28 29 454 significantly different reductions in self-reported negative affect nor significantly different 30 31 455 degrees of linguistic distancing, the objective group did, indeed, have a greater number of 32 33 456 significant correlations with positive health indicators than the spatial/temporal distance group, 34 35 457 alluding to potentially different mechanisms underlying these tactics. Future work may continue 36 37 38 458 to explore potentially divergent consequences, however, among spatial, temporal, and objective 39 40 459 distancing separately. 41 42 460 We were primarily interested in looking at relationships between linguistic distancing and 43 44 45 461 negative affectivity, general health and well-being, and emotion regulation. In the present study, 46 47 462 individuals who implemented more linguistic distancing reported having lower perceived stress. 48 49 463 Reductions in perceived stress have been reported as a function of receiving explicit longitudinal 50 51 52 464 training in psychological distancing (Denny & Ochsner, 2014). The perceived stress measure used 53 54 465 in the present study has been shown to reliably predict depressive symptoms and vulnerability to 55 56 57 58 59 60 Stress and Health Page 22 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 22 1 2 3 466 depression in otherwise healthy adults (Cohen et al., 1983; Hewitt, Flett, & Mosher, 1992), and 4 5 6 467 thus the present work substantiates the relevance of linguistic distancing to clinically-meaningful 7 8 468 health indicators, including perceived stress, self-reported depressive symptoms, and difficulties 9 10 469 using emotion regulation. 11 12 470 The present study also indexed general health including energy and vitality, emotional 13 14 15 471 well-being, and social functioning. Previous literature supports the relationship between emotion 16 17 472 regulation and vitality. For example, a study on 6,025 participants who completed the General 18 19 473 Well-being schedule foundFor that individuals Review with high levelsOnly of emotional vitality – a measure of 20 21 22 474 energy, positive well-being, and effective emotion regulation – had lower risk of coronary heart 23 24 475 disease relative to individuals with low levels of emotional vitality (Kubzansky & Thurston, 2007). 25 26 476 The present study elucidates connections between linguistic distancing and emotional well-being 27 28 29 477 and builds upon prior work; for example, in one study, individuals who report using reappraisal 30 31 478 frequently in their day-to-day life showed fewer depressive symptoms, more life satisfaction, more 32 33 479 optimism, had better self-esteem, and higher levels of personal growth, self-acceptance, 34 35 480 environmental mastery, and clear purpose in life (Gross & John, 2003). The present work extends 36 37 38 481 some of these benefits to the degree of linguistic distancing in particular during appraisal of 39 40 482 negative stimuli. In addition, the present study shows links between linguistic distancing and social 41 42 483 functioning. Reappraisal has been shown to predict stronger social connections and favorable 43 44 45 484 sociometric standing in college students (English et al., 2012). Thus, the present results, while 46 47 485 correlational, suggest that linguistic distancing may play a role in fostering such psychological 48 49 486 well-being and strong social relationships. 50 51 52 487 Future Directions & Limitations. 53 54 55 56 57 58 59 60 Page 23 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 23 1 2 3 488 Overall, the present study may have implications in translational and clinical domains 4 5 6 489 (Kross & Ayduk, 2017). The present results substantiate the idea that language may be used as a 7 8 490 target for efficacious psychotherapies. While no significant differences in correlation strength 9 10 491 between linguistic distancing and health indicators were found between the Objective and Far 11 12 492 groups, intriguingly, there were more health indicators significantly associated with objective 13 14 15 493 distancing relative to far distancing. It is important to note, however, that the difference between 16 17 494 “significant” and “not significant” is not itself statistically significant (Gelman & Stern, 2006). 18 19 495 While speculative given Forthe lack ofReview between-group differences, Only this may suggest that distancing 20 21 22 496 using objective means (i.e., appraising negative situations as an objective, rational, impartial 23 24 497 observer) may be more promising as a therapeutic “main ingredient” relative to distancing using 25 26 498 modulations of perceived spatial and temporal distance, per se. Future work may continue to 27 28 29 499 examine this possibility, as well as relevant individual difference factors that may predispose some 30 31 500 individuals to be more receptive to one reappraisal sub-tactic relative to another. 32 33 501 It is important to note that the present findings are correlational in nature. Hence, causality 34 35 502 cannot be established by the current data; it could be that psychological distancing leads to better 36 37 38 503 health outcomes, or it could be that better health leads to more effective use of distancing when 39 40 504 language is manipulated to incorporate lexical shifts. Causality may be explicitly tested in future 41 42 505 work via longitudinal designs examining the long-term health consequences of linguistic emotion 43 44 45 506 regulation using different tactics (e.g., distancing versus other forms of reappraisal) and between 46 47 507 distancing sub-tactics (e.g., objective versus far distancing). Relatedly, future work may further 48 49 508 unpack the correlates of linguistic distancing using additional sub-tactics, including separately 50 51 52 509 examining spatial and temporal distancing. If individuals are trained to implement more linguistic 53 54 510 distancing, they may show decreases in their perceived stress and increases in their general health 55 56 57 58 59 60 Stress and Health Page 24 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 24 1 2 3 511 over time. Consistent with this idea, Kross and colleagues suggest that individuals with major 4 5 6 512 depression and social anxiety may benefit from linguistic distancing, as self-distancing strategies 7 8 513 facilitate coping with distressing situations (Kross et al., 2014). 9 10 514 Indeed, psychological distancing is a cognitive reappraisal tactic with significant potential 11 12 515 to be used in a therapeutic setting. For example, clinical trials for depression and PTSD indicate 13 14 15 516 that practicing the ability to distance oneself is a flexible strategy that may be implemented in 16 17 517 clinical practice (Ayduk & Kross, 2010; Kross & Ayduk, 2017). Therefore, language, and in 18 19 518 particular language incorporatingFor psychologicalReview distancing, Only may represent a particularly effective 20 21 22 519 target for mental health interventions designed to foster adaptive emotion regulation. 23 24 520 That said, as with any emotion regulation strategy, it will be crucial to investigate for whom 25 26 521 and under which circumstances distancing may be most and least appropriate (Doré, Silvers, & 27 28 29 522 Ochsner, 2016), and how to flexibly select when to implement distancing and/or other emotion 30 31 523 regulation strategies (Gross, 2015b), which represents an important skill of its own and a key area 32 33 524 of future study. 34 35 525 Overall, the findings of this study illustrate relationships among language, emotion 36 37 38 526 regulation, and validated health indicators. Based on these results, future work may further 39 40 527 examine whether and how, when implemented in an appropriate context, linguistic distancing 41 42 528 manipulations may be an effective tool to foster adaptive health outcomes. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 25 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 25 1 2 3 530 References 4 5 6 531 Aldao, A., Nolen-Hoeksema, S., & Schweizer, S. (2010). Emotion-regulation strategies across 7 8 532 psychopathology: A meta-analytic review. Clinical Psychology Review, 30(2), 217–237. 9 10 533 Appleton, A. A., & Kubzansky, L.D. (2014) Emotion regulation and cardiovascular disease risk. 11 12 534 In: Gross JJ, editor. Handbook of Emotion Regulation. 2. New York: The Guilford Press. 13 14 15 535 Ayduk, Ö., & Kross, E. (2010). From a distance: Implications of spontaneous self-distancing for 16 17 536 adaptive self-reflection. Journal of Personality and Social Psychology, 98(5), 809–829. 18 19 537 Ayduk, Ö. & Kross, E. (2008).For Enhancing Review the Pace of Recovery:Only Self-Distanced Analysis of 20 21 22 538 Negative Experiences Reduces Blood Pressure Reactivity. Psychological Science, 19(3), 23 24 539 229–231. 25 26 540 Baikie, K. A., & Wilhelm, K. (2005). Emotional and physical health benefits of expressive 27 28 29 541 writing. Advances in Psychiatric Treatment, 11(5), 338–346. 30 31 542 https://doi.org/10.1192/apt.11.5.338 32 33 543 Berna, G., Ott, L., & Nandrino, J.-L. (2014). Effects of Emotion Regulation Difficulties on the 34 35 544 Tonic and Phasic Cardiac Autonomic Response. PLOS ONE, 9(7), e102971. 36 37 38 545 Buhle, J. T., Silvers, J. A., Wager, T. D., Lopez, R., Onyemekwu, C., Kober, H., … Ochsner, K. 39 40 546 N. (2014). Cognitive reappraisal of emotion: a meta-analysis of human neuroimaging 41 42 547 studies. Cerebral Cortex (New York, N.Y.: 1991), 24(11), 2981–2990. 43 44 45 548 Cohn, M. A., Mehl, M. R., & Pennebaker, J. W. (2004). Linguistic Markers of Psychological 46 47 549 Change Surrounding September 11, 2001. Psychological Science, 15(10), 687–693. 48 49 550 Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. 50 51 52 551 Journal of Health and Social Behavior, 24(4), 385–396. 53 54 55 56 57 58 59 60 Stress and Health Page 26 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 26 1 2 3 552 Denny, B. T., & Ochsner, K. N. (2014). Behavioral effects of longitudinal training in cognitive 4 5 6 553 reappraisal. Emotion (Washington, D.C.), 14(2), 425–433. 7 8 554 https://doi.org/10.1037/a0035276 9 10 555 Doré, B. P., Morris, R. R., Burr, D. A., Picard, R. P., & Ochsner, K. N. (2017). Helping Others 11 12 556 Regulate Emotion Predicts Increased Regulation of One’s Own Emotions and Decreased 13 14 15 557 Symptoms of Depression. Personality and Social Psychology Bulletin, 43(5), 729–739. 16 17 558 Doré, B. P., Silvers, J. A., & Ochsner, K. N. (2016). Toward a Personalized Science of Emotion 18 19 559 Regulation. Social andFor Personality Review Psychology CompassOnly, 10(4), 171–187. 20 21 22 560 Doré, B., Ort, L., Braverman, O., & Ochsner, K. N. (2015). Sadness shifts to anxiety over time 23 24 561 and distance from the national tragedy in Newtown, Connecticut. Psychological Science, 25 26 562 26(4), 363–373. https://doi.org/10.1177/0956797614562218 27 28 29 563 English, T., John, O. P., Srivastava, S., & Gross, J. J. (2012). Emotion Regulation and Peer- 30 31 564 Rated Social Functioning: A Four-Year Longitudinal Study. Journal of Research in 32 33 565 Personality, 46(6), 780–784. https://doi.org/10.1016/j.jrp.2012.09.006 34 35 566 Florin, I., Freudenberg, G., & Hollaender, J. (1985). Facial expressions of emotion and 36 37 38 567 physiologic reactions in children with bronchial asthma. Psychosomatic Medicine, 47(4), 39 40 568 382–393. 41 42 569 Frattaroli, J. (2006). Experimental disclosure and its moderators: a meta-analysis. Psychological 43 44 45 570 Bulletin, 132(6), 823–865. https://doi.org/10.1037/0033-2909.132.6.823 46 47 571 Gelman, A., & Stern, H. (2006). The Difference Between “Significant” and “Not Significant” is 48 49 572 not Itself Statistically Significant. The American Statistician, 60(4), 328–331. 50 51 52 573 https://doi.org/10.1198/000313006X152649 53 54 55 56 57 58 59 60 Page 27 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 27 1 2 3 574 Gratz, K. L., & Roemer, L. (2004). Multidimensional Assessment of Emotion Regulation and 4 5 6 575 Dysregulation: Development, Factor Structure, and Initial Validation of the Difficulties in 7 8 576 Emotion Regulation Scale. Journal of Psychopathology and Behavioral Assessment, 26(1), 9 10 577 41–54. https://doi.org/10.1023/B:JOBA.0000007455.08539.94 11 12 578 Greer, S., & Watson, M. (1985). Towards a psychobiological model of cancer: Psychological 13 14 15 579 considerations. Social Science & Medicine, 20(8), 773–777. 16 17 580 Gross, J. J. (1998a). The emerging field of emotion regulation: An integrative review. Review of 18 19 581 General PsychologyFor, 2(3), 271–299. Review https://doi.org/10.1037/1089-2680.2.3.271 Only 20 21 22 582 Gross, J. J. (1998b). Antecedent- and response-focused emotion regulation: divergent 23 24 583 consequences for experience, expression, and physiology. Journal of Personality and 25 26 584 Social Psychology, 74(1), 224–237. 27 28 29 585 Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes: 30 31 586 implications for affect, relationships, and well-being. Journal of Personality and Social 32 33 587 Psychology, 85(2), 348–362. 34 35 588 Gross, J. J. (2015a). Emotion Regulation: Current Status and Future Prospects (Vol. 26). 36 37 38 589 Gross, J. J. (2015b). The Extended Process Model of Emotion Regulation: Elaborations, 39 40 590 Applications, and Future Directions: Psychological Inquiry, 26(1). 41 42 591 Hewitt, P. L., Flett, G. L., & Mosher, S. W. (1992). The Perceived Stress Scale: Factor structure 43 44 45 592 and relation to depression symptoms in a psychiatric sample. Journal of Psychopathology 46 47 593 and Behavioral Assessment, 14(3), 247–257. https://doi.org/10.1007/BF00962631 48 49 594 Hu, T., Zhang, D., Wang, J., Mistry, R., Ran, G., & Wang, X. (2014). Relation between Emotion 50 51 52 595 Regulation and Mental Health: A Meta-Analysis Review. Psychological Reports, 114(2), 53 54 596 341–362. https://doi.org/10.2466/03.20.PR0.114k22w4 55 56 57 58 59 60 Stress and Health Page 28 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 28 1 2 3 597 Kashdan, T. B., Barrett, L. F., & McKnight, P. E. (2015). Unpacking Emotion Differentiation: 4 5 6 598 Transforming Unpleasant Experience by Perceiving Distinctions in Negativity. Current 7 8 599 Directions in Psychological Science, 24(1), 10–16. 9 10 600 Kross, E., Ayduk, Ö., & Mischel, W. (2005). When asking “why” does not hurt. Distinguishing 11 12 601 rumination from reflective processing of negative emotions. Psychological Science, 16(9), 13 14 15 602 709–715. https://doi.org/10.1111/j.1467-9280.2005.01600.x 16 17 603 Kross, E., Bruehlman-Senecal, E., Park, J., Burson, A., Dougherty, A., Shablack, H., … Ayduk, 18 19 604 O. (2014). Self-talkFor as a regulatory Review mechanism: howOnly you do it matters. Journal of 20 21 22 605 Personality and Social Psychology, 106(2), 304–324. https://doi.org/10.1037/a0035173 23 24 606 Kross, E., & Ayduk, Ö. (2017). Chapter Two - Self-Distancing: Theory, Research, and Current 25 26 607 Directions. In J. M. Olson (Ed.), Advances in Experimental Social Psychology (Vol. 55, 27 28 29 608 pp. 81–136). Academic Press. https://doi.org/10.1016/bs.aesp.2016.10.002 30 31 609 Kross, E. & Ayduk, Ö. (2008). Facilitating Adaptive Emotional Analysis: Distinguishing 32 33 610 Distanced-Analysis of Depressive Experiences From Immersed-Analysis and Distraction. 34 35 611 Personality and Social Psychology Bulletin, 34(7), 924–938. 36 37 38 612 Kubzansky, L. D., & Thurston, R. C. (2007). Emotional vitality and incident coronary heart 39 40 613 disease: benefits of healthy psychological functioning. Archives of General Psychiatry, 41 42 614 64(12), 1393–1401. https://doi.org/10.1001/archpsyc.64.12.1393 43 44 45 615 Lang, P. J., Greenwald, M. K., Bradley, M. M., & Hamm, A. O. (1993). Looking at pictures: 46 47 616 affective, facial, visceral, and behavioral reactions. Psychophysiology, 30(3), 261-273. 48 49 617 Liberman, N., & Trope, Y. (2008). The psychology of transcending the here and now. Science, 50 51 52 618 322(5905), 1201–1205. http://dx.doi.org/10.1126/science.1161958. 53 54 55 56 57 58 59 60 Page 29 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 29 1 2 3 619 Lieberman, M. D., Inagaki, T. K., Tabibnia, G., & Crockett, M. J. (2011). Subjective responses 4 5 6 620 to emotional stimuli during labeling, reappraisal, and distraction. Emotion (Washington, 7 8 621 D.C.), 11(3), 468–480. https://doi.org/10.1037/a0023503 9 10 622 Lieberman, M. D., Eisenberger, N. I., Crockett, M. J., Tom, S. M., Pfeifer, J. H., & Way, B. M. 11 12 623 (2007). Putting feelings into words: affect labeling disrupts amygdala activity in response 13 14 15 624 to affective stimuli. Psychological Science, 18(5), 421–428. https://doi.org/10.1111/j.1467- 16 17 625 9280.2007.01916.x 18 19 626 Lu, Q., & Stanton, A. L. (2010).For How Review benefits of expressive Only writing vary as a function of writing 20 21 22 627 instructions, ethnicity and ambivalence over emotional expression. Psychology & Health, 23 24 628 25(6), 669–684. https://doi.org/10.1080/08870440902883196 25 26 629 McRae, K., Ciesielski, B., & Gross, J. J. (2012). Unpacking cognitive reappraisal: goals, tactics, 27 28 29 630 and outcomes. Emotion (Washington, D.C.), 12(2), 250–255. 30 31 631 Mehl, M. R., Robbins, M. L., & Holleran, S. E. (2012). How Taking a Word for a Word Can Be 32 33 632 Problematic: Context-Dependent Linguistic Markers of Extraversion and Neuroticism. 34 35 633 Journal of Methods and Measurement in the Social Sciences, 3(2), 30–50. 36 37 38 634 Moore, S. A., Zoellner, L. A., & Mollenholt, N. (2008). Are Expressive Suppression and 39 40 635 Cognitive Reappraisal Associated with Stress-Related Symptoms? Behaviour Research 41 42 636 and Therapy, 46(9), 993–1000. https://doi.org/10.1016/j.brat.2008.05.001 43 44 45 637 Murdock, K., Fagundes, C., Kristen Peek, M., Vohra, V., & P. Stowe, R. (2016). The Effect of 46 47 638 Self-Reported Health on Latent Herpesvirus Reactivation and Inflammation in an 48 49 639 Ethnically Diverse Sample (Vol. 72). https://doi.org/10.1016/j.psyneuen.2016.06.014 50 51 52 53 54 55 56 57 58 59 60 Stress and Health Page 30 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 30 1 2 3 640 Nook, E. C., Schleider, J. L., & Somerville, L. H. (2017). A linguistic signature of psychological 4 5 6 641 distancing in emotion regulation. Journal of Experimental Psychology. General, 146(3), 7 8 642 337–46. 9 10 643 Ochsner, K. N., & Gross, J. J. (2008). Cognitive Emotion Regulation: Insights from Social 11 12 13 644 Cognitive and Affective Neuroscience. Current Directions in Psychological Science, 14 15 645 17(2), 153–158. 16 17 18 646 Ochsner, K. N., Silvers, J. A., & Buhle, J. T. (2012). Functional imaging studies of emotion 19 For Review Only 20 647 regulation: a synthetic review and evolving model of the cognitive control of emotion. 21 22 648 Annals of the New York Academy of Sciences, 1251, E1-24. 23 24 25 649 Paret, C., Brenninkmeyer, J., Meyer, B., Yuen, K. S. L., Gartmann, N., Mechias, M.-L., & 26 27 650 Kalisch, R. (2011). A Test for the Implementation–Maintenance Model of 28 29 30 651 Reappraisal. Frontiers in Psychology, 2. https://doi.org/10.3389/fpsyg.2011.00216 31 32 652 Park, J., Ayduk, Ö., & Kross, E. (2016). Stepping back to move forward: Expressive writing 33 34 653 promotes self-distancing. Emotion (Washington, D.C.), 16(3), 349–364. 35 36 37 654 Parkinson, C., Liu, S., & Wheatley, T. (2014). A Common Cortical Metric for Spatial, Temporal, 38 39 655 and Social Distance. Journal of Neuroscience, 34(5), 1979–1987. 40 41 656 https://doi.org/10.1523/JNEUROSCI.2159-13.2014 42 43 44 657 Pennebaker, J. W., & King, L. A. (1999). Linguistic styles: language use as an individual 45 46 658 difference. Journal of Personality and Social Psychology, 77(6), 1296–1312. 47 48 659 Pennebaker, J. W., & Chung, C. K. (2011). Expressive writing and its links to mental and 49 50 660 physical health. In H. S. Friedman (Ed.), Oxford handbook of health psychology. New 51 52 53 661 York, NY: Oxford University Press. 54 55 56 57 58 59 60 Page 31 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 31 1 2 3 662 Pennebaker, J. W., & Chung, C. K. (2007). Expressive Writing, Emotional Upheavals, and 4 5 6 663 Health. In Foundations of Health Psychology (pp. 263–284). Oxford University Press. 7 8 664 Pennebaker, J.W., Booth, R.J., Boyd, R.L., & Francis, M.E. (2015). Linguistic Inquiry and Word 9 10 665 Count: LIWC2015. Austin, TX: Pennebaker Conglomerates. 11 12 666 Pennebaker, J. W. (1997). Writing About Emotional Experiences as a Therapeutic Process. 13 14 15 667 Psychological Science, 8(3), 162–166. https://doi.org/10.1111/j.1467-9280.1997.tb00403.x 16 17 668 Powers, J. P. & LaBar, K. L. (in press). Regulating Emotion Through Distancing: A Taxonomy, 18 19 669 Neurocognitive Model,For and SupportingReview Meta-Analysis Only. Neuroscience and Biobehavioral 20 21 22 670 Reviews. 23 24 671 Radloff, L. S. (1977). The CES-D Scale: A Self-Report Depression Scale for Research in the 25 26 672 General Population. Applied Psychological Measurement, 1(3), 385–401. 27 28 29 673 Reinhold, M., Bürkner, P.-C., & Holling, H. (2018). Effects of expressive writing on depressive 30 31 674 symptoms—A meta-analysis. Clinical Psychology: Science and Practice, 25(1), e12224. 32 33 675 https://doi.org/10.1111/cpsp.12224 34 35 36 676 Tamir, D. I., & Mitchell, J. P. (2011). The default network distinguishes construals of 37 38 677 proximal versus distal events. Journal of Cognitive Neuroscience, 23(10), 2945–2955. 39 40 678 Trope, Y., & Liberman, N. (2010). Construal-Level Theory of Psychological Distance. 41 42 43 679 Psychological Review, 117(2), 440–463. https://doi.org/10.1037/a0018963 44 45 680 Troy, A. S., Wilhelm, F. H., Shallcross, A. J., & Mauss, I. B. (2010). Seeing the Silver Lining: 46 47 681 Cognitive Reappraisal Ability Moderates the Relationship Between Stress and Depressive 48 49 50 682 Symptoms. Emotion (Washington, D.C.), 10(6), 783–795. 51 52 683 Verzeletti, C., Zammuner, V. L., Galli, C., & Agnoli, S. (2016). Emotion regulation strategies 53 54 684 and psychosocial well-being in adolescence. Cogent Psychology, 3(1), 1199294. 55 56 57 58 59 60 Stress and Health Page 32 of 44

LINGUISTIC DISTANCING HEALTH INDICATORS 32 1 2 3 685 Ware, J. E., & Sherbourne, C. D. (1992). The MOS 36-item short-form health survey (SF-36). I. 4 5 6 686 Conceptual framework and item selection. Medical Care, 30(6), 473–483. 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 33 of 44 Stress and Health

LINGUISTIC DISTANCING HEALTH INDICATORS 33 1 2 3 Table 1. Correlations of Linguistic Distancing Metric Collapsed Across Psychological Distancing Groups and Within 4 Each Group 5 Group 6 Health Indicator Collapsed Across 7 Objective Far 8 Objective and Far 9 Negative Affectivity (Composite) r = -.301*, p = .017 r = -.450*, p = .010 r = -.200, p = .281 10 Depression Symptoms r = -.266*, p = .035 r = -.442*, p = .011 r = -.124, p = .887 11 * * 12 Perceived Stress r = -.311 , p = .013 r = -.427 , p = .015 r = -.252, p = .171 13 14 15 General Well-Being (Composite) r = .276*, p = .029 r = .414*, p = .018 r = .175, p = .345 16 Energy and Vitality r = .232†, p = .066 r = .352*, p = .048 r = .176, p = .343 17 Emotional Well-Being r = .291*, p = .021 r = .411*, p = .019 r = .172, p = .354 18 19 Social Functioning For Reviewr = .194, p = .127 Onlyr = .355*, p =.046 r = .087, p = .640 20 21 22 Emotion Regulation (Composite) r = .326*, p = .009 r = .250, p = .168 r = .440*, p = .013 23 Reappraisal Frequency r = .272*, p = .031 r = .108, p = .556 r = .454*, p = .010 24 * † † 25 Difficulties in Emotion Regulation r = -.287 , p = .022 r = -.315 , p = .079 r = -.309 , p = .091 26 687 Notes: *indicates significance at the .05 level (2-tailed), †indicates marginal significance at the .10 level (2-tailed). 27 688 For the Objective and Far groups, the linguistic distancing metric was calculated by subtracting the Look Negative 28 689 linguistic distancing score from the Change Negative linguistic distancing score. 29 690 30 691 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Stress and Health Page 34 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 Figure 1. Linguistic Distancing composite metric within each trial type within each group. Error bars 27 represent ± one standard error. 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 35 of 44 Stress and Health

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 Figure 2. Collapsing across the psychological distancing groups, there was a significant association between 27 the linguistic distancing composite metric and (A) perceived stress (r(61) = –.31, p = .013), (B) symptoms 28 of depression (r(61) = –.27, p = .035), (C) emotional well-being, (r(61) = .29, p = .021), and (D) 29 difficulties in emotion regulation, (r(61) = –.29, p = .022). 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60