Upward , Downward , and Horizontal Social Comparisons : Effects on Adjustment , Emotions , and Persistence in Teachers

Corresponding author: Sonia Rahimi, Department of Educational and Counselling Psychology, McGill University, Montreal, QC, Canada E-mail: sonia.rahimi@mail.mcgill.ca Citation: Rahimi, S., Hall, N.C., Wang, H., Maymon, R. (2017). Upward, Downward, and Horizontal Social Comparisons: Effects on Adjustment, Emotions, and Persistence in Teachers. Interdisciplinary Education and Psychology, 1(1):10. Open Access Interdisciplinary Education and Psychology

Concerning research on factors that may help protect teachers from the deleterious effects of teaching-related stress, psychology research suggests that individual differences in personality characteristics (e.g., conscientiousness; Kokkinos, 2007), motivational beliefs (e.g., goals : Butler, 2007; self-efficacy: Wang, Hall, & Rahimi, 2015;causal attributions: Manassero et al., 2006), or coping strategies (e.g., Austin, Shah, & Muncer, 2005;Griffith, Steptoe, & Cropley, 1999) can help to explain vulnerability to stress in teachers.Research on psychological variables in teachers further suggests that teachers' cognitions concerning their social environment may also influence not only their emotions in class (e.g., anger, enjoyment; Frenzel, Goetz, Stephens, & Jacob, 2009) but also their self-reported burnout levels (e.g., depersonalization; Byrne, 1994).However, empirical research on how teachers' beliefs concerning their social environment help them cope with the demands and challenges of their profession are currently lacking (e.g., Sutton, 2004).Moreover, despite a substantial body of research illustrating the potential benefits of strategic social comparisons with one's peers in stressful environments (e.g., Suls, Martin, & Wheeler, 2002;Suls & Wheeler, 2000), specific research on the potential benefits of teachers' beliefs concerning other teachers has yet to be conducted.

Social Comparisons
In contrast to psychological constructs previously examined in research with teachers that have mainly involved persistent motivational beliefs or personality characteristics, motivational strategies (also referred to as "coping strategies" in highly stressful circumstances; Skinner, Edge, Altman, & Sherwood, 2003, or emotional regulation, Gross, 1998) represent how individuals intentionally respond to situational challenges and are thus more dynamic or reflexive in nature.Coping is understood as cognitive or behavioral responses to manage significant stressors (Lazarus & Folkman, 1984), with one specific class of coping strategies involving adaptively comparing oneself with others so as to protect or foster psychological well-being (Suls et al., 2002).These strategies have consistently been found to promote adjustment in response to varied developmental challenges across the life span, as illustrated in empirical research following from both social and developmental psychological traditions.

Social comparison processes
As a prominent theory of motivational self-regulation bridging psychological literatures, Heckhausen, Wrosch, and Schulz's (2010) Motivational Theory of Life-Span Development attempts to explain how individuals regulate their motivation in response to situational opportunities and constraints through three main types of motivational strategies (Hall, Chipperfield, Heckhausen, & Perry, 2010).In educational settings, the most beneficial strategies typically involve goal engagement (effort, help-seeking, reducing distractions) and the most detrimental strategies tend to reflect goal disengagement (lowering expectations; e.g., teachers' intentions to quit; Klassen & Chiu, 2011).Further, self-protective strategies involving the cognitive reinterpretation of setbacks to preserve psychological resources have shown considerable promise for fostering personal and academic development in students (e.g., Hall, 2008;Hall, Chipperfield, Perry, Ruthig, & Goetz, 2006;Hall, Perry, Ruthig, Hladkyj, & Chipperfield, 2006) and teachers (e.g., Totterdell & Parkinson, 1999).Nevertheless, existing research on self-protective strategies in educational populations has focused primarily on one strategy, namely positive reappraisal (i.e., "finding the silver lining"), leaving other potentially beneficial self-protective strategies unexplored.More specifically, whereas research informed by Heckhausen et al.'s (2010) model has examined the benefits of strategic comparisons with others on well-being in adult populations (e.g., elderly adults; Chipperfield & Perry 2006;Hall et al., 2010;Heckhausen, 1999), no studies to date have explored the potential emotional and motivation benefits of this type of motivational strategy for teachers.

Downward social comparisons
Commonly defined as comparing oneself with worse-off others, the motivation behind downward social comparisons is assumed to be self-enhancement of subjective well-being and self-esteem (Hakmiller, 1966;Taylor & Lobel, 1989;Wills, 1981).The psychological benefits of downward comparisons are supported by research showing this strategy to be more beneficial for overall adjustment and mood than upward or horizontal comparisons (Locke, 2003), with individuals often engaging in downward comparisons when they feel threatened or perceive a lack of personal control (Stewart, Chipperfield, Ruthig, & Heckhausen, 2013;Pyszcynski, Greenberg, & Laprelle, 1985;Wheeler & Miyake, 1992;Wills, 1981).Although some studies suggest downward comparisons can negatively affect selfevaluations and emotions under stressful circumstances (Buunk, Collins, Taylor, VanYperen, & Dakof, 1990;Kemmelmeier & Oyserman, 2001), other findings show downward comparisons to help maintain positive self-perceptions and emotional well-being (Bailis & Chipperfield, 2006;Lockwood, Dolderman, Sadler, & Gerchak, 2004), particularly in lowcontrol circumstances (e.g., older adulthood; Bailis et al, 2005;Heckhausen & Schulz, 1995).As such, empirical findings to date on the psychological benefits of downward comparisons are mixed and have yet to be investigated in teachers.

Upward social comparisons
Social comparison theories posit that people are often motivated to seek out positive characteristics they share with more successful others so as to utilize positive role models to inspire accomplishments (Lockwood, Jordan, & Kunda, 2002).The term "role model" is defined as an individual with whom one compares oneself who has achieved a desired goal or status (Holton, 2004) and inspires one to improve one's own requisite competencies (a selfimproving motive, Festinger, 1954;Wood, 1989; see also Lockwood et al., 2002).Accordingly, although some negative effects of this comparison strategy have been reported (e.g., Bailis & Chipperfield, 2006;Buunk et al., 1990;Jagacinski & Nicholls, 1987), upward comparisons have typically been found to foster psychological adjustment across interpersonal, achievement, and health domains (for reviews, see Bailis et al., 2005;Buunk et al., 1990;Collins, 1996;Corcoran, Crusis, & Mussweiler, 2011;Festinger, 1954;Tesser, 1988;Taylor & Lobel, 1989).In educational settings specifically, although comparisons with better-off others have been shown to promote student motivation (e.g., Huguet, Dumas, Monteil, & Genestoux, 2001;Lockwood & Kunda, 1997), the potential psychological benefits of this social comparison strategy for teachers has not been empirically explored.

Social comparisons in teachers
When dealing with instructional challenges, such as student misbehaviour or poor performance, teachers are likely to compare themselves with other teachers that, in many cases, should prove useful for motivating themselves as well as regulating their negative emotions.However, in contrast to existing research examining the effects of social comparisons on adjustment and achievement among adults in interpersonal contexts (e.g., romantic relationships) and health domains (e.g., serious illness), and students in K-12 settings (e.g., achievement outcomes), notably little research to date has examined effects on well-being and motivation in teachers.Concerning existing research with teachers on constructs related to social comparisons, studies have found higher levels of collective selfefficacy, defined as the perceived confidence of an individual teacher in the collective capabilities of teachers as a group (e.g., Tschannen-Moran, Hoy, & Hoy, 1998), to correspond with greater student learning (Parker, Hannah, & Topping, 2006) and lower levels of burnout (Skaalvik & Skaalvik, 2007).Similarly, not unlike the observed emotional benefits of teachers' social goals reflecting a desire to connect emotionally with students (e.g., Butler, 2012;Butler & Shibaz, 2008), teachers who perceive a sense of relatedness to other teachers also report higher levels of instructional engagement and lower negative emotions (Klassen, Perry, & Frenzel, 2012).In addition, teachers are more likely to engage in upward comparisons with individuals whom they perceive as having attained desired achievable goals (Smith, 2000), or to individuals that inspire them (Kitchel et al., 2012).However, this research on general social perceptions in teachers is mixed, with findings also showing perceived relatedness with colleagues to negatively predict work engagement and enjoyment (Klassen et al., 2012), with the findings also limited in that they do not differentiate between social comparison types (downward, horizontal, upward).

The Present Study
Whereas research to date exploring teacher motivation and emotions is consistent with a traditional expectancy-value perspective on achievement motivation (Klassen, Tze, Betts, & Gordon, 2011;Wigfield, Tonks, & Klauda, 2009), research exploring the role of higher-order self-regulation strategies is presently lacking, particularly concerning social comparisons.Given studies underscoring the importance of motivational strategies and volition in educational settings (e.g., Kuhl, 1996;Oettingen & Gollwitzer, 2009;Pintrich, 1999;Wolters, 2003) higher-order, self-regulatory strategies should similarly predict psychological outcomes in teachers.The present study aimed to address this research gap in investigating the effects of specific types of social comparisons (downward, horizontal, upward) on burnout, job satisfaction, intentions to quit, emotions and illness symptoms in teachers.

Hypothesis 1: Downward comparisons
Downward social comparisons should be valuable for teachers in predicting better adjustment and emotional outcomes (e.g., well-being, Locke, 2003;Lockwood & Kunda, 1997).To the extent that the teaching profession is increasingly characterized by stressors beyond the personal control of teachers (e.g., workload, administrative support), the effects of downward comparisons were expected to be consistent with findings showing comparisons with worse-off others to benefit individuals in low-control circumstances (e.g., Bailis et al, 2005;Heckhausen & Schulz, 1995).

Hypothesis 3: Upward comparisons
Upward social comparisons were anticipated to promote better adjustment outcomes and emotions in teachers following from findings with adult populations showing motivational and emotional benefits from contrasts with better-off others (e.g., Huguet et al., 2001;Lockwood & Kunda, 1997).Whereas the psychological and health effects of upward comparisons in prior research is mixed (e.g., Bailis & Chipperfield, 2006;Buunk et al., 1990;Jagacinski & Nicholls, 1987), this strategy was expected to be beneficial for teachers given the present operationalization of upward contrasts as involving role models representing attainable teaching goals (as opposed to simply more successful others, possibly resulting in envy or resentment).

Hypothesis 4: Moderation by experience
Finally, it was expected that the psychological effects of each type of social comparison would be moderated by teachers' years of experience, with new teachers benefitting more from social comparisons than more experienced teachers.This hypothesis is consistent with prior studies showing social comparison effects to be moderated by individual differences (e.g., perceived control, Buunk et al., 1990) and studies based on Heckhausen et al.'s (2010) theory showing beliefs concerning others to have stronger effects for individuals facing control-limiting health challenges (e.g., collective self-efficacy, Bailis & Chipperfield, 2006;downward comparisons, Stewart, Chipperfield, Perry, & Weiner, 2012; for a review, see Heckhausen & Schulz, 1995) or reporting low perceptions of control (e.g., upward comparisons, Bailis & Chipperfield, 2006).To the extent that new teachers experience threatened perceptions of control concerning the profession (e.g., Roth & Gold, 2003) and are logically more likely to engage in certain comparisons (e.g., upward, or senior colleagues), they were expected to experience greater motivational and emotional benefits than their more experienced counterparts.Nevertheless, this interaction hypothesis should be considered speculative in nature given a lack of existing research on the moderating effects of experience or perceived control on the well-being effects of motivational variables for teachers.

Study measures
Descriptive statistics for each study measure including means, standard deviations, ranges, and internal reliabilities are provided in Table 1 below.

Social comparisons
Three multi-item measures adapted from the Optimization of Primary and Secondary Control Scale by Heckhausen and colleagues (OPS; see Haase, Heckhausen, & Koeller, 2008;Heckhausen, Schults, & Wrosch, 1998; 1 = strongly disagree to 5 = strongly agree) were developed for this study to evaluate downward, horizontal, and upward social comparisons in teachers, with two items measuring intentional interpersonal contrasts with respect to downward comparisons (M = 3.03, SD = 1.00, r(457) = .51,p < .001;e.g., "When I experience teaching setbacks, I remind myself that I am better-off than other teachers in many ways"), two items measuring horizontal comparisons (M = 3.34, SD = .89,r(465) = .33,p < .001;e.g., "When I have difficulties with my students, I keep in mind that other teachers are struggling too"), and two items measuring upward comparisons (M = 3.68, SD = .93,r(457) = .63,p < .001;e.g., "When I experience teaching difficulties, I remind myself of successful teachers who overcame similar setbacks").

Job satisfaction
A 5-item, 7-point measure was used to assess job satisfaction (Moe, Pazzaglia, & Ronconi, 2010; 1 = strongly disagree to 7 = strongly agree).Sample items included "In most ways my job is close to my ideal" and "If I could live my life over, I would not change the choices made in my job" (M = 5.07, SD = 1.38, α = .90;cf.Moe et al., 2010, α = .84).

Illness symptoms
A 3-item, 5-point scale was adapted from Cohen and Hoberman (1983; 1 = not at all in a week to 5 = five or more times a week) to assess teachers' perceived frequency of illness

Preliminary analyses
Independent-samples t-tests were conducted to examine potential initial differences in our study measures as a function of gender.The results showed a weak gender effect such that female teachers (M = .92,SD = .90)reported higher depersonalization than male teachers (M = 1.29,SD = 1.13; t(77.61)= 2.53, p = .013,d = .25).Zero-order correlations (see Table 2) additionally showed positive correlations between each type of social comparison (downward, horizontal, and upward), and an expected pattern of positive relations among similarly valenced dependent measures (e.g., job satisfaction, enjoyment) as well as negative relations between oppositely valenced outcomes (e.g., job satisfaction, exhaustion).Moreover, age was negatively correlated with horizontal comparisons, depersonalization, anger, and anxiety, as well as positively correlated with enjoyment and years of experience.Grade level was positively correlated with depersonalization and negatively correlated with personal accomplishment, warranting the inclusion of both variables as covariates in the main analyses.Finally, although education level was not correlated with any outcome measure, it was confounded with level of instruction and similarly included as a covariate consistent with previous literature (e.g., Kokkinos, 2007;Pas, Bradshaw, & Hershfledt, 2012).

Main analysis
Hierarchical linear regressions evaluated the effects of social comparisons on burnout (exhaustion, depersonalization, accomplishment), job satisfaction, quitting intentions, teaching-related emotions (enjoyment, anger, anxiety), and illness symptoms, and the moderating effects of teaching experience.Step 1 included demographic variables evaluated as covariates including teachers' gender, age, highest level of education achieved (postsecondary diploma, bachelor's degree, master's degree, doctoral degree), and grade level of instruction (primary/elementary, secondary/high school, post-secondary).
Step 2 evaluated the main effects for each social comparison measure (Hypotheses 1-3) as well as years of experience.Finally, Step 3 assessed the interactions between each social comparison measure and years of experience (Hypothesis 4).All main and interaction effects were meancentered prior to analysis with simple slopes analyses evaluated at one standard deviation above or below the sample mean (Meyers, Gamst, & Guarino, 2006).

Main effects
As outlined in Table 3, results showed that whereas downward comparisons predicted higher levels of job satisfaction in teachers, they also predicted higher levels of anger.In contrast, horizontal comparisons predicted higher levels of burnout with respect to emotional exhaustion and depersonalization, as well as lower levels of personal accomplishment.Additionally, horizontal comparisons predicted lower job satisfaction, stronger quitting intentions, lower enjoyment, as well as higher levels of anger, anxiety, and illness symptoms.Finally, upward comparisons were found to predict better levels on each outcome assessed including emotional exhaustion, depersonalization, personal accomplishment, job satisfaction, intentions to quit, enjoyment, anger, anxiety, and illness symptoms.

Discussion
Despite long-standing empirical evidence showing social comparisons to be an effective strategy for coping with stressors in interpersonal, health, and educational settings (e.g., Buunk et al., 1990;Collins, 1996;Corcoran et al., 2011;Festinger, 1954), and related psychosocial variables showing significant benefits for teachers (e.g., relatedness, collective self-efficacy; Klassen et al., 2012, Goddard et al. 2000), social comparisons have not previously been explored as a motivational strategy among teachers.To address this research gap, the present study investigated the psychological effects of social comparisons as a motivational self-regulation strategy among teachers on critical occupational measures of burnout, job satisfaction, and intentions to quit, as well as teaching-related emotions and illness symptoms.Overall, the consistent pattern of results observed provide empirical support for further exploring social comparisons in this population in showing teachers' efforts to compare themselves with other teachers to have both beneficial and detrimental effects depending on the type of comparison made, with these effects being further moderated by teaching experience.

Hypothesis 1: Downward comparisons
The first study hypothesis proposed that despite mixed findings in other domains on the psychological effects of downward comparisons, teachers who compared themselves to worse-off others should report better adjustment and motivation levels.The present results provide partial support for this hypothesis in showing ambivalent effects of this type of social comparison strategy.Whereas teachers who more frequently compared themselves with worse-off teachers reported greater job satisfaction, they also reported significantly higher levels of anger.As such, these findings are consistent with studies showing downward comparisons to both contribute to better adjustment (e.g., Locke, 2003;Pohlmann, & Möller, 2009) as well as to more negative emotions toward others (e.g., Buunk et al., 1990).However, given that the mixed effects for downward comparisons were notably weak in magnitude (βs = .11,.14),and that no significant effects of this strategy were found for any other measures, the primary implication from these results is that downward comparisons appear to be a largely ineffective predictor of psychological well-being in teachers.

Hypothesis 2: Horizontal comparisons
As outlined in the second study hypothesis, horizontal social comparisons were similarly expected to predict better well-being outcomes, based in part on previous work showing related socially-oriented variables to predict lower levels of burnout and higher job satisfaction in teachers (e.g., perceived relatedness, collective self-efficacy; Bailis & Chipperfield, 2006;Klassen et al., 2012;Skaalvik & Skaalvik, 2007).However, a pattern of results directly opposite of Hypothesis 2 was observed, with comparisons made by teachers to similar others (i.e., other teachers who were similarly struggling to achieve their instructional goals) proving particularly maladaptive for teachers in predicting poorer levels on all outcomes assessed.More specifically, horizontal comparisons predicted not only poorer levels on each burnout measure (emotional exhaustion, depersonalization, personal accomplishment), but also lower job satisfaction, stronger intentions to quit, a more negative profile of teaching-related emotions (lower enjoyment, higher anger and anxiety), and higher levels of reported illness symptoms.
Despite these results being directly contrary to existing research on similar constructs (e.g., positive effects of teacher relatedness on job satisfaction; Skaalvik & Skaalvik, 2007), they are nonetheless consistent with findings showing teachers who report higher levels of relatedness with colleagues to also experience lower work engagement and enjoyment (Klassen et al., 2012).According to Klassen et al. (2012), teachers' efforts to compare themselves with other teachers who they perceive as similarly unsuccessful should not be equated with efforts to build relationships in the classroom (e.g., social goals, Butler &Shibaz, 2008), with social comparison research further showing perceptions of similarity with others on undesirable traits can to lead to negative emotions and feelings of insecurity (Locke, 2005).As the present measure of horizontal comparisons included statements such as "If I don't reach my teaching goals, I will tell myself that many other teachers are in the same situation," it is possible that the clearly failure-oriented, commiserative nature of this scale may have contributed to this negative pattern of results (i.e., representing the opposite of collective selfefficacy).

Hypothesis 3: Upward comparisons
The third hypothesis proposed that upward social comparisons would be beneficial for teachers, following from existing research showing a focus on positive role models, or individuals who have overcome similar challenges, to enhance motivation and well-being (Corcoran et al., 2011;Huguet et al, 2001;Lockwood & Kunda, 1997;Taylor & Lobel, 1989).In clear support of this hypothesis, the present findings showed upward social comparisons to be by far the most effective self-regulation strategy across all outcomes, and the only comparison strategy to predict lower quitting intentions.More specifically, teachers who reported more upward comparisons were found to experience, not only lower levels of burnout (exhaustion, depersonalization, accomplishment), but also higher job satisfaction and optimal teaching emotions (enjoyment, anger, anxiety), as well as less frequent illness symptoms and thoughts about quitting the profession.These results are consistent with theoretical assertions that teacher self-efficacy can indeed be facilitated through the vicarious experience of observing demonstrations of skillful teaching (Tschannen-Moran et al., 1998; see also Kunter et al., 2011), as well as research in adult and student populations showing comparisons with others who have successfully overcome similar challenges to contribute to positive affect and motivation (e.g., Bailis et al., 1995;Huguet et al., 2001).The present findings thus contribute to research on the potential psychological benefits of upward social comparisons in showing this strategy to most strongly and consistently predict, not only physical and psychological health outcomes in teachers, but also their stated intentions to quit the profession.

Hypothesis 4: Moderation by experience
Consistent with existing social comparison research showing the effects of upward social comparisons to contribute to well-being particularly for individuals who are struggling to make social connections (e.g., Bailis & Chipperfield, 2006) or reporting low perceptions of personal control (e.g., Bailis et al, 2005;Buunk et al., 1990), our results also suggested that upward social comparisons were more beneficial for less experienced teachers.More specifically, whereas comparisons with relevant role models were of benefit to more experienced teachers with respect to job satisfaction, intentions to quit, and teaching-related enjoyment, these effects were noticeably stronger for new teachers.These moderation effects underscore the importance of educational research examining motivation and self-regulation early in the teaching profession (e.g., within the first two years; Roness, 2011) as well as continued efforts in social comparison research to examine how individual differences on psychosocial factors might moderate the effects of these comparisons on adjustment (e.g., Buunk et al., 1990;Lockwood et al., 2004).

Limitations and future directions
Concerning potential limitations of the present study to be addressed in future research, five main issues warrant consideration.First, the study employed a newly developed scale adapted from the Optimization of Primary and Secondary Control measure (OPS; Haase et al., 2008) consisting of only two items per subscale, with the two items for the horizontal comparisons measure not being highly correlated (r = .33).Whereas the first item addressed teachers bonding with others who similarly failed to meet personal teaching goals, the second focused on connecting with other teachers over difficulties with students.Additional analyses further suggest that teachers' attempts to bond over shared personal failures may be primarily responsible for the negative effects of horizontal comparisons, with post-hoc single-item correlations showing the personal failure item to be more strongly correlated with burnout than the student-focused item (e.g.r = -.12 vs. -.04 for personal accomplishment).
Accordingly, in addition to the moderate internal reliability estimates found for the upward and downward comparison measures (.63 and .51),further research to develop more in-depth and reliable self-report measures of social comparisons in teachers is needed.Further, selfreport or interview methods inquiring as to variety of specific teaching experiences and the types of social comparisons made by teachers in response would provide a more ecologically valid means of exploring the prevalence and effectiveness of each social comparison strategy.Additionally, given the importance of social comparison processes for responding to not only setbacks but also success experiences in educational settings (e.g., outperforming others, Exline & Lobel, 1999), future research should explore how teachers compare amongst themselves in response to positive events is recommended (e.g., student successes, teaching recognition).
The second limitation concerns mediating and moderating variables not investigated in this study.More specifically, as evidenced by prior research showing other psychosocial variables such as self-efficacy, achievement goals, and perceived control to moderate the effects of social comparisons (e.g., Butler, 1992;Buunk et al., 1990;Lockwood et al., 2004), it is possible that similar moderation effects may account for the mixed effects observed (e.g., downward comparisons).Additionally, further examination of the moderating effects of additional variables not explored in this study, such as school culture, cohort effects, leadership style, and personality-related factors, could also help to ascertain how and when these effects occur in real-world teaching contexts.Similarly, it is possible that other, more directly overlapping socio-cognitive variables not assessed in this study may to some extent account for the observed relationships, for example, the negative effects of horizontal comparisons on adjustment outcomes (e.g., competitiveness, belongingness, social goals; Bergin & Bergin, 2009;Butler, 2012;Van Ryzin et al., 2009).As such, it is anticipated that future research in which additional potential moderators or mediators of the relations between social comparisons and adjustment in teachers are examined may shed more light on the mixed findings in this study.
Third, given that the comparisons examined in this study were from teachers to specific others, future research to investigate comparisons made by teachers with larger reference groups or between teacher cohorts (e.g., two groups of teachers) is also needed for greater insight into the varied ways in which teachers compare themselves on a collective level.Furthermore, to better understand the positive implications of upward comparisons on teacher well-being, further study is recommended to also examine other types of upward comparisons that may not be motivating for teachers.For example, upward comparisons would not be expected to be beneficial for teachers who focus on the disparities between themselves and role models, or for whom the success of the role model is perceived as unattainable.In addition, future studies would benefit from exploring domain specificity in regards to the types of social comparisons preferred by teachers in response to varied classroom stressors (i.e., dealing with setbacks regarding student misbehavior as compared to setbacks regarding interactions with colleagues or administrators).
A fourth limitation pertains to the cross-sectional and self-report nature of the study data.Given that the long-term relationships between these social comparisons strategies and adjustment outcomes may differ from those concurrently observed, future longitudinal research is warranted to replicate these findings over a longer duration (e.g., six-month lag).Likewise, whereas the present methodology consisted exclusively of self-report measures, further examination of teachers' social comparison strategies utilizing more objective methodologies that circumvent self-report biases (e.g., common method variance) is encouraged.For example, experience sampling methods could better inform our understanding of how teachers compare themselves to others in real-life classroom settings (cf.Goetz, Frenzel, Stoeger, & Hall, 2010), with objective health indicators also serving as more ecologically valid outcome measures (e.g., teacher self-efficacy and cortisol; Schwerdtfeger, Konermann, & Schonhofen, 2008).
Finally, whereas the present study findings suggest clear potential psychological benefits for teachers who focus on role models, as well as possible negative implications of seeking social comparisons with similar others over shared teaching failures, further research to experimentally evaluate these assertions is needed.More specifically, interventions promoting a focus on role models among teachers, and explicitly discouraging teachers to seek out similarly failing others (e.g., "teaching misery does not love company"), are recommended to more directly evaluate the causal implications of the present findings.For example, teachers could be provided an opportunity to reflect on a teacher they admire as part of a professional development seminar, or dialogue with an experienced teacher who has overcome setbacks, so as to promote adjustment and persistence in new teachers.In sum, it is anticipated that future research on the types of social comparisons made by teachers could help to meaningfully inform the development of first-year orientation and ongoing professional development programs for teachers aimed at both reducing attrition and improving psychological well-being.

Figure 1 .
Figure 1.Upward comparisons predicting job satisfaction as a function of years of experience

Table 2 .
Zero-order correlations among study variables

Table 3 .
Hierarchical multiple regression analyses