Appraisal self-respect: Scale validation and construct implications

Abstract

Despite the widely accepted recognition of the notion of self-respect and its importance for emotional well-being, it has received scant attention in the psychological literature. We report on the development and validation of a scale to measure trait (character-based) appraisal self-respect (ASR), conceptualised as a disposition to perceive or appraise oneself as being a respectworthy honourable person. We tested the factor structure, reliability, convergent, discriminant and criterion validity of the ASR scale in samples of adult individuals (combined N = 1910 across samples). The resulting ASR scale was found to be essentially unidimensional and showed good internal and acceptable test-retest reliability. Trait ASR was correlated with (yet distinct from) theoretically related measures of global self-esteem, moral self and principledness, and was distinct from other self-esteem facets not based on honourable character traits. Importantly, it related to well-being and prosocial behaviour over-and-above self-esteem. The validation work served to consolidate the theoretical boundaries and utility of this important concept.

Introduction

The concept of self-respect is widely regarded to be fundamental to emotional and psychological well-being. It is recognised as a feature of self-esteem (e.g., Crocker et al., 2003; Rosenberg, 1965), and it is believed to be critical to living a life that brings satisfaction and promotes flourishing (Dillon, 2010). However, little research has been directed at the concept of self-respect in its own right. In this paper, we demonstrate the value of a particular type of self-respect named appraisal self-respect for social and psychological outcomes.

We define trait (character-based) appraisal self-respect (ASR) as a disposition to perceive or appraise oneself as being a respectworthy honourable person (Kumashiro et al., 2002; Dillon, 2010). Respect is a specific type of attitude that, unlike liking, is directed towards a target (the self in the case of self-respect) when perceived to possess attributes which command recognition and proper consideration or regard, regardless of personal affinities and needs (Clucas, 2019; Dillon, 2010; Frei & Shaver, 2002; Prestwich & Lalljee, 2009). In this way, self-respect is a more specific self-evaluation than global self-esteem (Clucas, 2019), which is defined as a generalised attitude of favourableness or unfavourableness towards the self as a whole (Rosenberg et al., 1995). Respect-commanding or respectworthy qualities that have received the most attention in the literature in relation to self-respect include one’s achieved admirable honourable character¹ traits and one’s inherent worth as a human being (Clucas, 2019; Renger, 2018). In this paper, we focus on self-respect in the form of perceived respectworthiness of one’s character as honourable, as opposed to self-respect in the form of appreciation or recognition of one’s personhood as a rational, autonomous and equal agent (Dillon, 2010; Kristjansson, 2007; Kumashiro et al., 2002; Renger, 2018). We are specifically interested in perception of one’s respectworthiness grounded in an appraisal of one’s character as honourable and thus worthy of regard, as opposed to a more global self-respect assessment. We focus on ASR as a trait, that is an individual difference variable with people having levels of ASR that are relatively stable over time and across situations (Kumashiro et al., 2002; Kristjansson, 2007), as opposed to a momentary state (Luchies et al., 2010).

High trait ASR individuals perceive themselves as respectworthy by virtue of their honourable character, which commands the regard of themselves and others by conferring them value and standing (Cross et al., 2014; Kristjansson, 2007). According to the Oxford English Dictionary, being honourable is “a quality of character entitling the person to great respect” and to “hold in honour” is to feel “great respect” towards a person or thing (“honour”, 2021, para. 3, 4). Which character traits are considered honourable and thus worthy of regard is related to the honour code internalised within one’s social environment. In Western cultures, being an honourable person involves behaving in moral, principled and prosocial ways, and demonstrating dignified behaviour in the form of affirming and defending one’s worth (Cross et al., 2014; Uskul et al., 2012). ASR is also dependent on showing fortitude of character by refusing to submit to “contemptible, degrading, or otherwise immoral” behaviours that can undermine personal integrity and/or social reputation (see Kristjansson, 2007, p. 229–230; Schlenker, 2008; Uskul et al., 2012), and standing up for oneself and one’s convictions, thereby affirming one’s worth, and demonstrating moral courage (Kristjansson, 2007; Luchies et al., 2010; Telfer, 1968).

Empirical research supports honourable character traits as being respectworthy in the eyes of others and in one’s own eyes. Research on interpersonal respect shows respect towards someone to rest on the perception of that person’s honourable character traits (e.g., moral integrity, concern for others’ welfare, mental toughness) (see Frei & Shaver, 2002; Prestwich & Lalljee, 2009). Similarly, research supports that one’s honourable character traits, such as moral integrity and strength of character, influence one’s perceived respectworthiness and respect for oneself. Indeed, adherence to moral standards significantly influenced direct global momentary ratings of self-respect over-and-above self-esteem (see Clucas, 2019). Luchies et al. (2010) also found a person’s self-respect (as measured by a single global item) to diminish when they acted as a “doormat” and failed to show strength of character by forgiving a partner who had not made amends.

A self-perception as a moral person is an important basis for ASR but ASR is a broader construct than moral self-appraisal; other character traits also contribute to ASR, such as adherence to broader non-moral principles, aspects of strength of character such as moral courage and standing up for oneself, and dignified behaviour. These additional character traits which characterise ASR are not captured by moral self-concept or self-esteem measures (e.g., the Self-description Questionnaire (SDQ) III Honesty/Reliability subscale (Marsh & O’Neill, 1984) or the Multidimensional Self-esteem Inventory (MSEI) moral self-esteem subscale (O'Brien, 1980)) that specifically assess self-perception as a moral person who behaves in accordance with their moral values. It is, for instance, possible to behave morally, but not stand up for oneself and, therefore, not feel honourable and worthy of regard (i.e. respectworthy) (Luchies et al., 2010; Telfer, 1968). It is also possible to experience ASR from engaging in prosocial behaviours that do not entail perceived moral obligation but demonstrate admirable personal qualities (De Groot & Steg, 2009; Uskul et al., 2012). Trait ASR is likely to play an important role in helping connect the disparate moral self and global self-esteem bodies of literature (Power, 2004). Moral individuals may experience self-esteem from a broader self-evaluation as honourable and respectworthy, even when not strongly invested in being moral per se (see Clucas, 2019).

Despite the concept being commonly referred to in everyday discourse and its likely important implications for well-being and prosocial behaviour, little research has investigated trait ASR. A notable exception is the study by Kumashiro et al. (2002) that showed trait ASR to predict personal well-being and pro-relationship behaviour in marital relationships independently of self-esteem. However, Kumashiro et al.’s measure of ASR was only a preliminary attempt to measure the construct and had not undergone a thorough validation process. Moreover, research is needed to consolidate the theoretical boundaries of trait ASR with the related concepts of global self-esteem and moral self-appraisal.

Since it is a broad self-evaluation as having respectworthy honourable character traits, we expect ASR to be a relatively stable self-evaluation. High ASR individuals are likely to experience secure feelings of self-worth from a confident self-perception as having honourable character traits that warrant regard and respect from oneself and others, regardless of personal desires, affinities or needs (Dillon, 2010), with positive implications for well-being (Paradise & Kernis, 2002). ASR may also be more amenable to intervention than other self-esteem facets (e.g., appearance or social approval), which are less under one’s personal control. In addition, we expect trait ASR to benefit society by relating to moral and prosocial behaviour, more so than trait global self-esteem (Rosenberg et al., 1995) since this domain-specific self-evaluation is more directly relevant to these outcomes (Marsh & Craven, 2006), and evidence of a relationship between self-esteem and moral and prosocial behaviour has been inconclusive (Baumeister et al., 2003). Our purpose is 1) to provide a measurement scale of trait ASR, which is valid and reliable, 2) to demonstrate the utility of the concept in predicting positive psychological and social outcomes over-and-above global self-esteem, and 3) to empirically consolidate the theoretical boundaries of the construct.

Overview of Present Research

We present the results of a series of studies designed to provide reliability and validation evidence for the Appraisal Self-respect Scale (ASRS) and demonstrate the utility of the construct. First, we present information on the development and finalisation of ASR scale items, confirm the scale’s unidimensional factor structure, establish trait ASR as distinct from global self-esteem and demonstrate test-retest reliability (Phase 1). Next, we demonstrate the scale’s convergent and discriminant validity using measures of moral self, principledness, personality traits, and other self-esteem facets (Phase 2), and show the ASRS to relate to prosocial behaviour and subjective well-being over-and-above self-esteem (Phase 3).

All studies obtained ethical approval from University Ethics committees. Informed consent was obtained by providing participants with an information sheet detailing study procedures and risks, and explaining that questionnaire completion assumed informed consent, or in some cases electronically (part of samples 4B and 6) or in writing (subsidiary investigation in Phase 2). Participation was voluntary and anonymous. A power calculation indicated that a minimum of 150 participants was needed to detect a typical effect size in individual differences research of r = .20 at 80% power; we aimed to recruit a minimum of 250 participants in each study investigating correlates of ASR to achieve stable estimates (Schönbrodt & Perugini, 2013).

Phase 1: Scale Development, Factor Structure, Reliability and Relationship with Self-Esteem

In this phase, we 1) present information on the development and factor structure of the ASRS, 2) demonstrate internal and test-retest reliability and 3) model the relationship between trait ASR and global self-esteem using factor analysis and structural equation model (SEM) techniques.

We aimed to develop a unidimensional trait ASR scale since we were interested in capturing the global dimension of (character-based) appraisal self-respect (i.e., global self-appraisal as having a respectworthy honourable character). A focus on the global dimension (as opposed to the individual underpinning character traits) is theoretically justified because it is generally agreed that ASR makes most sense as a holistic concept (see Kristjansson, 2007). We aimed to develop a measure for use in the general adult population that was brief for ease of administration in basic and applied research contexts.

As a specific self-evaluation as having respectworthy honourable character traits, ASR is narrower in bandwidth than global self-esteem (see Clucas, 2019). In line with multidimensional models of self-esteem (see Marsh & Craven, 2006), we expected ASR to be distinct from, yet strongly related to global self-esteem. This was also expected based on prior research on state self-respect and ASR (Clucas, 2019; Kumashiro et al., 2002). Like other domain-specific self-evaluations, ASR was expected to contribute to global self-esteem though a bottom-up process, and also to be influenced by global self-esteem in a top-down fashion (see Rosenberg et al., 1995). Indeed, feelings of self-worth developed early in life support the use of self-enhancing strategies or biases to promote and protect feelings of self-worth in people with high self-esteem, such as engaging in selective social comparison processes, taking credit for success and excusing failure, and minimising self-descriptiveness of undesirable traits (Brown et al., 2001). High global self-esteem can, therefore, lead to inflated and non-realistic self-views, particularly with respect to global domain self-assessments (Baumeister et al., 2003), which are also more likely to be inflated by overall feelings of positivity (vs. negativity) associated with high self-esteem. Self-reported ASR is, therefore, likely to reflect global self-esteem in addition to grounded or “realistic” self-perception as having respectworthy honourable character traits.

Method

Samples

Demographic sample details on age, gender and country of residence can be found in Table 1. Sample 1 (pilot study sample) was used to provide additional support for honourable character traits being perceived as respectworthy prior to the development of the ASR item pool and consisted of 80 university students recruited face-to-face across university campus contacts. Sample 2 was used to explore the factor structure of the ASRS (exploratory factor analysis (EFA) sample) and consisted of 219 university students and acquaintances of student researchers,² recruited through the psychology department online research participation system (RPS) as well as via face-to-face contacts on and off campus. Sample 3 was used to confirm the factor structure of the ASRS (confirmatory factor analysis (CFA) 1 sample) and consisted of a different sample of 371 university students recruited through RPS. Sample 4 (CFA 2 sample) was used to provide additional support for the ASRS’s factor structure in a more diverse general population sample of 731 non-student adults from Western English-speaking countries recruited through Prolific Academic, a crowdsourcing research platform structure (paid between £0.90 and £2.50 for their participation depending on questionnaire length – rate of around £6/h). Sample 4 was also used to model the relationship between ASR and global self-esteem. Seven participants were excluded from Sample 3 because they did not show engagement with the questions based on their response to negatively worded items (see Online Resource Appendix D) and 29 participants (4%) were excluded from Sample 4 because they failed the attention check on the questionnaire (see Online Resource Appendix D for supporting literature).

Table 1 Sample (S) characteristics

A subgroup of 255 participants from Sample 4 (Sample 4B) was invited to complete the ASR scale three months later to provide test-retest reliability information on the ASR scale. Test-retest reliability data was obtained from 196 participants (77% follow-up rate) of which 180 passed the attention check.

Measures

We aimed to obtain a pool of items from which to construct a unidimensional measure of trait ASR. Prior to generating the item pool based on theorising and existing research on respect, self-respect and honour, we carried out a pilot study to provide additional support for honourable character traits being perceived as respectworthy. We aimed to demonstrate that honourable character traits are perceived to be more closely related to self-respect than the more general construct of self-esteem, using vignettes, and gain a fuller picture of characteristics more closely associated with self-respect than with self-esteem through an inductive analysis of answers to open-ended questions.

Sample 1 completed a study in two parts. In Part 1, they completed two open-ended questions asking them about the personal attributes/characteristics that contribute to their self-esteem and self-respect. An inductive thematic analysis was performed to look for patterns in the open-ended responses, followed by McNemar tests to compare frequencies of mention of each characteristic for self-respect and self-esteem. In Part 2, participants were presented with ten scenarios describing a person as having or lacking an honourable character trait postulated or shown in the literature to be related to self-respect: Adherence to morals, criminal behaviour, dignified behaviour, magnanimity, moral courage, personal care (taking care of one’s appearance), respecting the environment and hard work - or self-esteem: Confidence in skills/abilities and competence (in sports) (see Online Resource Appendix B for supporting literature). Participants rated the scenarios on the extent to which they agreed that the person described had (a) high/low self-respect and (b) high/low self-esteem on 5-point Likert scales. Self-respect and self-esteem scenario ratings were compared using paired samples t-tests.

In accordance with prior literature (Clucas, 2019; Kristjansson, 2007; Kumashiro et al., 2002; Luchies et al., 2010), analysis of the open-ended questions showed participants listed honourable character traits when describing characteristics contributing to their self-respect, namely morals (standards and being law-abiding), strength of character (discipline and willpower), and dignity, which were mentioned significantly more frequently for self-respect than for self-esteem, supporting honourable character traits as being experienced as important respectworthy qualities. Similarly, paired samples t-tests supported the portrayed character traits of adherence to morals, criminal behaviour, not having dignified behaviour, magnanimity, moral courage, not showing personal care and not respecting the environment as influencing how respectworthy a person is by showing these traits to be more strongly related to self-respect than to self-esteem in scenario ratings (see Online Resource Appendix B for a fuller description of the results).

In contrast, the identified characteristics of attractiveness, confidence, positive attitude towards self and emotional state (e.g., anxiety and well-being) were mentioned significantly more frequently for self-esteem than for self-respect, and competence and confidence in skills/abilities were more strongly related to self-esteem in scenario ratings, reflecting the contribution of physical, emotional and performance self-esteem domains and the definition of global self-esteem as a person’s attitude towards the self as a whole (Marsh & Craven, 2006). These findings support self-respect as being a distinct and more specific self-evaluation than global self-esteem in line with theorising and prior research (Clucas, 2019).

Subsequently, we used existing theorising and research on respect, self-respect and honour (Clucas, 2019; Cross et al., 2014; Dillon, 2010; Kristjansson, 2007; Kumashiro et al., 2002; Luchies et al., 2010; Frei & Shaver, 2002; Prestwich & Lalljee, 2009; Uskul et al., 2012), complemented by the results of the pilot study, to develop items to capture self-perception as having key respectworthy honourable character traits, notably: Adherence to morals/standards, strength of character (including moral courage) and having dignified behaviour which in combination support trait ASR (see Kristjansson, 2007). We aimed to capture self-appraisal as having honourable character merits, as opposed to recognition of inherent human worth (Dillon, 2010; Renger, 2018). We started with 24 items but agreed on a pool of 18 items for further testing that most directly captured the concept of ASR (see Online Resource Appendix C for a list of the items). The items had simple phrasing and face as well as content validity, enhanced by having an expert in the field from outside the research team comment on the items. Both positively and negatively worded items were included to reduce acquiescence bias. Three of the items were taken directly from Kumashiro et al.’s preliminary short ASR measure: “I give in too easily to others’ wishes or requests” (RC), “I should treat myself better than I do” (RC) and “I have a lot of respect for myself”. The items were scored using 7-point Likert scales (1 = strongly disagree, 2 = disagree, 3 = slightly disagree, 4 = neither agree nor disagree, 5 = slightly agree, 6 = agree, 7 = strongly agree).

To examine the ASRS’s relationship with self-esteem, Sample 4 also completed the ten item Rosenberg (1965) Self-esteem Scale (RSES) (α = .92).

Statistical Analysis

To explore the factor structure of the 18 initial ASR items, an EFA was conducted in Sample 2 using Robust Maximum Likelihood Extraction (RML) in FACTOR 10.8.04 (Lorenzo-Seva & Ferrando, 2013). Two CFAs were carried out in Mplus version 6 using Maximum Likelihood Robust estimation in Samples 3 and 4 to confirm the unidimensional factor structure of the final ASR scale. Model fit was evaluated using various indices: Tucker–Lewis Index (TLI) and Comparative Fit Index (CFI) values ≥ .90 and Root Mean Square Error of Approximation (RMSEA) values < .08 indicate acceptable model fit whilst TLI and CFI values ≥ .95, RMSEA values ≤ .06 and Standardized Root Mean Square Residual (SRMR) values ≤ .05 indicate good model fit (Byrne, 2013). For the EFA and CFA analyses, data were first checked for extreme univariate and multivariate outliers. Unidimensionality assessment I-ECV (Item Explained Common Variance) and ECV (Explained Common Variance) indices were obtained for all three samples from EFAs using RML in FACTOR.

To examine the test-retest reliability of the ASRS scale, an Intraclass Correlation Coefficient (ICC) was used. To examine the relationship of ASR with global self-esteem, an EFA of the 7 ASRS items and 10 RSES items was initially conducted in FACTOR using Robust Diagonally Weighted Least Squares (RDWLS) with polychoric correlations (since the RSES used a shorter 4-point rating scale) and Promin rotation in Sample 4. This was followed by full bifactor and correlated ASR self-esteem bifactor CFA analyses in Mplus 6 using Diagonally Weighted Least Squares (WLSMV) estimation to model ASR as a distinct subdomain of self-esteem and then estimate the correlation between the ASR factor and the general self-esteem factor.

Results and Discussion

Exploratory Factor Analysis (EFA)

The EFA identified one main factor based on parallel analysis (eigenvalue of 5.46, explained variance = 30.3%) in line with our theorising. Eight extreme outliers were removed prior to carrying out the EFA. With the exception of three items with very low factor loadings, factor loadings ranged between .30 and .82 (see Online Resource Appendix C). We then selected items most closely associated with the main factor, with factor loadings above .55 (Comery & Lee, 1992). After redundancy concerns were considered, we were left with seven items (see Table 2 for the list). Most of the items that were removed because of factor loadings below .55 were negatively worded, which is consistent with research showing that introducing negatively worded items can create a method artefact linked to differences in response style to positively and negatively worded items that compromises the unidimensionality of scales such as self-esteem and self-concept measures (Marsh, 1996). Moreover, participants often find it harder to respond to negatively worded items, increasing the risk of confusion and inaccurate responses (Marsh, 1996) (see Online Resource Appendix C for additional analyses supporting this explanation, and Online Resource Appendix D for other methods that we used to check for acquiescence bias or careless responding).

Table 2 Descriptive statistics and standardised factor loadings for the retained seven ASRS items in the EFA and CFA analyses

The factor with the retained seven items explained 51% of the variance in the items (eigenvalue of 3.58) (see Table 2 for factor loadings). The 7-item ASR scale (ASRS) correlated significantly with the deleted items’ average composite score, r(207) = .58, p < .001. Although fewer in number, the items still provided a good representation of the key character traits as described earlier, underpinning ASR.³ The items were also broad rather than referring to specific experiences, making them more widely applicable.

Inspection of unidimensionality assessment item-level indices showed all seven items to satisfactorily represent the principal latent dimension (i.e., global dimension of appraisal self-respect) with I-ECV values over .77, see Table 2); I-ECV values represent the proportion of common variance explained by the first principal factor at the item level. An overall scale ECV (proportion of all common variance in the ASRS items explained by the first principal factor) of .90 (above the threshold of .70) also supported the final 7-item scale to be essentially unidimensional despite some heterogeneity in item content (Ferrando & Navarro-González, 2018). The factor was also strong and well-defined and, therefore, likely replicable with a Generalized G-H Index value of .87 (> .80, see Ferrando & Navarro-González, 2018).

Inter-item correlations for the seven items ranged from .28 to .63 (M = .43). Corrected item-total correlations ranged from .49 to .75 (α = .84, Omega total = .89, Omega H = .77⁴). Descriptive statistics for the ASR items in the EFA and CFA samples can be found in Table 2.

CFAs

Twelve extreme univariate and/or multivariate outliers were removed for CFA 1 in Sample 3 and 16 for CFA 2 in Sample 4. CFA 1 in Sample 3 confirmed the unidimensionality of the 7-item ASRS, showing that the one-factor model fit the data well with fit indexes within acceptable to good ranges: χ2(14, N = 359) = 36.11, p = .001, RMSEA = .067, TLI = .94, CF I = .96, SRMR = .04 (see Table 2 for factor loadings). Corrected item-total correlations ranged from .46 to .71 (α = .83, Omega total = .88, Omega H = .75). CFA 2 in Sample 4 also showed the one-factor model to fit the data reasonably well: χ2(14, N = 715) = 64.39, p < .001, RMSEA = .070, TLI = .94, CF I = .96, SRMR = .03, but the fit improved after correlating the residuals for the items “I take pride living according to my moral code” and “I will stick to my principles even if asked to do otherwise” as recommended by the modification indices (this suggestion makes theoretical sense since the two items both directly assess adherence to principles): χ2(13, N = 715) = 40.60, p < .001, RMSEA = .056, TLI = .97, CF I = .98, SRMR = .03 (see Table 2 for factor loadings). Corrected item-total correlations ranged from .58 to .71 (α = .87, Omega total = .90, Omega H = .82). Returning to CFA sample 1, we found the residuals for the two items to also be significantly correlated (p = .016), with the model fit also improving after allowing for the residuals to correlate (χ2(13, N = 359) = 30.32, p = .004, RMSEA = .062, TLI = .95, CFI = .97, SRMR = .03).

In both CFA samples, I-ECV values were all above .72 (CFA sample 1) or .76 (CFA sample 2, see Table 2) and the overall ECV was high (see Table 2), supporting the ASRS to be essentially unidimensional. The factor was also strong and well-defined (see Table 2 for G-H values). Samples 5/7 provided further evidence of a unidimensional factor structure for the ASRS⁵ (see Online Resource Appendix G).

Test-Retest Reliability

The ASRS showed acceptable test-retest reliability over a three-month period with an ICC of .76 [.69, .82] (N = 176). This supports the measure as tapping a relatively stable trait.

Relationship between ASR and Global Self-Esteem

The manifest correlation between the ASRS and RSES in Sample 4 was .58. In accordance with our predictions, an EFA of the 7 ASRS items and 10 RSES items in FACTOR supported ASR and global self-esteem as being distinct. Indeed, parallel analysis identified two factors, although an acceptable fit was only achieved with a 3-factor model (RMSEA = .072, CI_95% = [.05, .08], vs. RMSEA (2-factors) = .088, CI_95% = [.07, .10] and RMSEA (1-factor) = .135, CI_95% = [.12, .15]): An ASR factor, and the RSES items divided into positive and negative self-esteem factors.

Subsequently, a CFA analysis was carried out to model ASR as a distinct subdomain of self-esteem by fitting a full bifactor model to the ASRS and RSES items with the ASRS items specified to load on the general self-esteem (g) factor as well as a third (ASR) grouping factor alongside the positive and negative self-esteem method factors. This model was a good fit to the data χ2(102, N = 686) = 384.58, p < .001, RMSEA = .064, TLI = .97, CF I = .98). Factor loadings on the ASR factor were all significant and ranged from .25 to .74, indicating notable amounts of unique variance for the ASR items over the g factor, supporting ASR as a distinctive subdomain of self-esteem (Gomez et al., 2015). See Fig. 1 in Online Resource Appendix F for a pictorial representation of the model with factor loadings. A similar bifactor model but with ASR and positive self-esteem items specified to load on the same grouping factor provided a worse fit: χ2(102, N = 686) = 589.34, p < .001, RMSEA = .083, TLI = .96, CFI = .97.

Lastly, we estimated a correlation of .59 between the ASR factor and the general self-esteem factor in a correlated ASR self-esteem bifactor CFA model in which a bifactor model was fitted just to the RSES items to account for the two positive and negative self-esteem method factors (see Hyland et al., 2014). This also supported the constructs as strongly related but distinct, with 65% of the variance being unshared. The model was a good fit, χ2(107, N = 686) = 484.81, p < .001, RMSEA = .072, TLI = .97, CFI = .97, after allowing the global self-respect item to cross-load on the general self-esteem factor as indicated by a large modification index (see Online Resource Appendix F for factor loadings).

The global self-respect item “I have a lot of respect for myself” also had the smaller loading of .25 on the ASR factor in the full bifactor model while the other ASR items had factor loadings over .50. This item was included to tap perceived respectworthiness of oneself as an honourable person, but as a more global assessment as worthy of respect, may not have been expected to load as highly on the ASR grouping factor. In particular, as a global evaluative item, it is likely to be heavily influenced by global affective feelings of self-worth in addition to grounding in honourable character (Baumeister et al., 2003). Other research has documented face valid global ratings of overall appearance and performance to similarly load more marginally on the appearance and performance domain specific grouping factors in a bifactor analysis (see Clucas, 2019). Supporting the ASRS as measuring a self-evaluation of respectworthiness is the stronger correlation of the total score rating for the six other items with this global item (r = .59) than with global self-esteem (r = .49), William’s t(671) = 4.22, p < .001. In addition, this global item had a high I-ECV value of .94 in the ASRS unidimensionality analysis for this sample, which was not paralleled by the more global RSES item “On the whole I am satisfied by myself” that showed an I-ECV of .46 when added to the ASRS unidimensionality analysis in this sample.

We provide additional evidence for ASR being distinct from global self-esteem in the next two phases by showing different association patterns for ASR and global self-esteem with other constructs. Furthermore, we demonstrate that the ASRS is primarily a measure of “realistic” self-evaluation as being a respectworthy honourable person, as opposed to simply reflecting self-esteem, by showing theoretically expected association patterns with moral self, principledness and prosocial variables predicted to be specifically related to ASR, that were not substantially influenced by adjustment for self-esteem.

Phase 2: Establishing Convergent and Discriminant Validity

Following development and validation of the ASRS, shown to be essentially unidimensional, to be reliable and to measure a construct distinct from global self-esteem, in this phase we aimed to examine the convergent and discriminant validity of the ASRS with existing measures.

Given that a self-perception as moral and as having integrity plays an important role in supporting a self-perception as an honourable and respectworthy person, we selected a range of measures related to the moral self and integrity and certain “moral” personality traits, which we expected to correlate with the ASRS (to establish convergent validity), alongside other measures, which we did not expect to correlate with the ASRS (to establish discriminant validity).

We expected a moral self-concept to relate to ASR. We also expected characteristics found to motivate moral and principled behaviour to relate to ASR by facilitating a positive self-perception as moral and/or as having personal integrity across situations. These self-characteristics include: 1) moral identity, which motivates moral behaviour through a desire to behave in a manner consistent with one’s sense of self (e.g., Aquino & Reed, 2002), 2) virtue-contingent self-esteem (Crocker et al., 2003), which motivates morally principled behaviour to maintain self-esteem, and 3) commitment to a principled as opposed to expedient ideology, which motivates principled behaviour (Schlenker, 2008). We expected these characteristics to relate more strongly to trait ASR than global self-esteem. Self-control (the ability to self-regulate one’s behaviour) (Tangney et al., 2004) is also likely to facilitate adherence to moral standards and other valued standards of conduct.

In addition, we expected trait ASR to relate to the “moral” personality traits of agreeableness and conscientiousness, because the former represents the tendency to be cooperative, considerate and helpful and the latter incorporates attributes such as self-discipline, diligence and dependability (John & Srivastava, 1999), which facilitate a self-perception as a principled, honourable and respectworthy person (Clucas, 2019; Frei & Shaver, 2002). On the other hand, we did not expect ASR to relate to neuroticism other than through its relationship with global self-esteem. People high in trait ASR could still experience feelings of tension and moodiness.

We also did not expect the humility personality dimensions of modesty and greed-avoidance to relate to ASR in a systematic way; Although modesty is a positive moral trait, modest individuals may not rate themselves highly for the excellence of their character, and motivations for high social status and wealth are not strictly immoral or necessarily dishonourable so long as they are not at the expense of others.

As a specific self-evaluation based on self-perception as being a respectworthy honourable kind of person, we expected trait ASR to be distinct from other self-esteem facets that are not based on moral or honourable character such as academic, social, appearance and physical ability self-evaluations, and to be weakly related to them after adjusting for their common relationship to global self-esteem given the reciprocal relationship between specific self-evaluations attached to different domains and global self-esteem (see Rosenberg et al., 1995).

Subsidiary to testing the relationships described above to provide evidence of convergent and discriminant validity, we aimed to provide additional evidence of convergent validity using a behavioural measure of security in moral self-image; we expected high ASR individuals to possess a secure moral self-image since they have a confidently held self-perception as being a respectworthy honourable person. We also explored the ASRS’s relationship with impression management (IM).

Method

Samples

Demographic sample details on age, gender and country of residence can be found in Table 1. Sample 4a consisted of a subgroup of 456 non-student adults from Sample 4 recruited through Prolific Academic who completed the ASRS, RSES and measures of agreeableness, conscientiousness and neuroticism (John & Srivastava, 1999). Two weeks later (temporal separation introduced to reduce common method bias), 265 were invited to complete measures of moral self-concept, moral self-characteristics facilitating moral behaviour, commitment to a principled ideology (referred to as “principledness” thereafter), self-control, humility personality dimensions and impression management, receiving an additional £3 to the £1.30 initially received. Nine and 17 respondents were respectively excluded from the first and second questionnaire for failing attention checks.

Sample 5 consisted of 407 adults recruited from the UK university student community (students and acquaintances of a student researcher²) through RPS and Facebook (N = 302), or from the American web platform Reddit https://www.reddit.com (N = 105) who completed measures of self-esteem facets not based on moral or honourable character, alongside the ASRS and the RSES. The pattern of results was similar in the university student community and Reddit sample (Mage = 26.60 (SD = 7.76), 64% with degree or above qualification) and adjusting for recruitment source did not change the results, so we combined the samples.

For the subsidiary investigation that aimed to provide additional evidence of convergent validity using a behavioural measure of security in moral self-image, a sample of 109 UK university students (29 males, 79 females; 82 undergraduates, 23 postgraduates) was recruited through RPS and face to face across campus.

Measures

Sample 4a and Sample 5 completed the ASRS (respectively α = .86 and α = .83) and the RSES (α = .92 and α = .91). They also completed the following convergent and/or discriminant validity measures.

Convergent Validity Measures

These were completed by Sample 4a and included the Big Five Personality Inventory (BFI) 5-point Likert 9-item Agreeableness (α = .78) and 9-item Conscientiousness (α = .77) subscales (John & Srivastava, 1999); the 12-item 8-point Likert Honesty/Reliability moral self-concept subscale of the Self-Description Questionnaire III (Marsh & O’Neill, 1984) (α = .81); Aquino and Reed’s (2002) 5-item 5-point Likert measures of moral identity internalisation (α = .81) and symbolisation (α = .83) measuring the extent to which moral traits are central to the self-concept, and are reflected in the person’s actions in life, respectively; 5-item 7-point Likert virtue subscale of the Contingency of Self-Worth scale (Crocker et al., 2003) (α = .86), measuring the extent to which self-esteem is contingent on moral adequacy; the 18-item 5-point Likert Integrity Scale (Schlenker, 2008) (α = .85), measuring the strength of commitment to a principled ideology (i.e., principledness); and 13-item 5-point Likert Brief Self-Control Scale (Tangney et al., 2004) (α = .85), measuring the ability to self-regulate one’s behaviour.

Discriminant Validity Measures

These were the BFI 8-item Neuroticism (α = .90) subscale (John & Srivastava, 1999) and the 4-item 5-point modesty and greed-avoidance HEXACO Personality Inventory-Revised subscales (α = .66 for each) (Lee & Ashton, 2004) completed by Sample 4a, as well as the Fleming and Courtney’s (1984) revised Janis-Field Feelings of Inadequacy 5-point Likert subscales measuring physical appearance (5-item, α = .66), academic (7-item, α = .72), physical ability (5-item, α = 82) and social confidence (12-item, α = .88) domain-specific self-evaluations completed by Sample 5.

Impression Management

Impression management (IM) was measured in Sample 4A using the 13-item Marlowe-Crowne Social Desirability Scale (MCSDS) - short form (Reynolds, 1982) with response options yes/no (α = .64).

Subsidiary Investigation: Behavioural Measure of Security in Moral Self-Image

Participants were invited to take part in an adapted Implicit Association test (IAT); they had to rapidly categorise moral trait adjectives and their antonyms into the categories “me or moral” or “not me or immoral” (“non-threatening” critical trials) or into the categories “me or immoral” or “not me or moral” (“threatening” critical trials). It was predicted high ASR participants would experience the “threatening” trials as less threatening to their self-concept and be quicker at accurately categorising the trait adjectives in those trials, showing less difference in mean response latency between the two critical trials for the accurate trials (full details of the method are provided in Online Resource Appendix J).

Statistical Analysis

Pearson (or Spearman) zero-order and partial correlations (obtained through multiple linear regression analyses with ASR and self-esteem as predictors) were used to examine the relationships between ASR and the convergent, discriminant and impression management measures. In both this phase and Phase 3, we examine ASR’s relationship with other variables before and after adjusting for global self-esteem to gain a clearer picture of the influence of grounded or “realistic” ASR over and above the influence of self-enhancing biases and global affective evaluations of self-worth. An EFA using RML extraction and Promin rotation in FACTOR 10.8.04 was also used to support ASR as being distinct from self-esteem facets not based on moral or honourable character.

Results and Discussion

The correlations between ASR and self-esteem were .60 in Sample 4A and .63 in Sample 5.

Moral Self, Principledness, and Personality Variables

In line with our predictions, Pearson (or Spearman) zero-order and partial correlations showed the ASRS (M, SD = 37.54, 6.24) was significantly positively correlated with measures of moral self-concept, moral identity internalisation and symbolisation, virtue-contingent self-esteem, principledness, self-control, agreeableness and conscientiousness (before and after adjusting for self-esteem)⁶⁷ but showed weak and non-significant relationships to constructs theorised to be unrelated to ASR, including neuroticism, modesty (the latter two after adjusting for global self-esteem⁶), and greed-avoidance (see Table 3 and Online Resource Appendix H for correlations between the measures). This provides evidence of convergent and discriminant validity for the ASRS.

Table 3 Zero-order and partial correlations with convergent, discriminant and criterion-related validity measures

In contrast to ASR, self-esteem (M, SD = 29.24, 6.47) showed non-significant correlations with moral identity internalisation and symbolisation, virtue-contingent self-esteem, and principledness, and the correlations were significantly weaker than those with ASR (see Table 3). Moreover, compared to ASR, moral self-concept showed a weaker correlation with self-esteem and was no longer related to self-esteem after adjusting for ASR (see Table 3), which is consistent with our conceptualisation of ASR as encompassing, but broader than, moral self-appraisal, involving a more general evaluation of oneself as being a respectworthy honourable person, and also suggests a critical role of ASR for self-worth relating to positive moral self-appraisal.⁸

Self-Evaluations Not Based on Honourable Character Traits

The ASRS (M, SD = 37.40, 6.18) was significantly related to physical appearance, academic, physical ability and social confidence domain specific self-evaluations, but correlations ranged between .22 and .35 (see Table 3), indicating ASR was distinct from these, only sharing between 4.8% and 12.3% of variance with each. An EFA also supported five distinct factors, with correlations between ASR and the other four self-evaluations ranging between .17 and .32; the ASR rotated factor explained 19% of the variance (see Online Resource Appendix I for a fuller description of results). Correlations were close to zero and non-significant after adjusting for global self-esteem⁶, further supporting the scale’s discriminant validity. On the other hand, relationships between non-honourable domain specific self-evaluations and self-esteem (M, SD = 27.06, 6.17) were moderate to large, stronger than those with ASR, and remained substantial after adjusting for ASR (see Table 3). This provides additional evidence of ASR being distinct from self-esteem.

Relationship with Impression Management

The ASRS was significantly correlated with the MCSDS (respectively r(245) = .42, p < .001 and r(243) = .31, p < .001, before and after adjusting for self-esteem). This could indicate a response bias with respondents claiming to have high ASR to present themselves favourably or could be explained by the finding that high IM scorers tend to have an agreeable and interpersonally adjusted personality with high levels of self-control (Uziel, 2010). However, acknowledging the caveat that social desirability responding (SDR) scale scores are inherently ambiguous in that it is difficult to separate self-report bias from true personality, a similar pattern of results was found when controlling for SDR scores, with relationships being slightly weaker but still significant.

Subsidiary Investigation: Behavioural Measure of Security in Moral Self-Image

In support of high ASR participants having a more secure moral self-image, higher ASR participants showed less of a difference in response latency between the critical trials (r = −.22, p = .036). Self-esteem was not related to this measure. Interestingly, high ASR participants also recalled more of the moral trait adjectives presented during the adapted IAT (r = .27, p = .009), consistent with having a more highly articulated self-schema as moral, as expected if deriving worth from evaluating themselves positively in the moral domain (see Online Resource Appendix J for full details of the analyses and results).

Phase 3: Examining Associations with Behavioural and Psychological Outcomes

In this section, we examine the ASRS’s association with behavioural and psychological outcomes, including prosocial behaviour and well-being.

We expected trait ASR to relate to prosocial behaviour since acting in ways that benefit others (for instance, donating blood or money, or volunteering) demonstrates noble and admirable character, which attracts honour and respect (Frei & Shaver, 2002; Uskul et al., 2012). The desire to maintain a self-perception as being a respectworthy honourable person is therefore likely to motivate prosocial behaviour, regardless of the perceived moral obligation to perform the prosocial behaviour (see De Groot & Steg, 2009).

However, ASR also necessitates behaving in ways that do not compromise honour by asserting one’s worth when wronged (Telfer, 1968). An individual high in ASR may, therefore, be reluctant to adopt the prosocial response, for instance, forgive a friend when they have betrayed them, despite forgiveness being the superior moral course of action, whilst restraining from engaging in revenge tendencies of returning harm for harm since this is a dishonourable response.

We also expected trait ASR to relate to well-being through increased self-esteem, but also over-and-above self-esteem level because of secure feelings of self-worth from a confident well-founded self-perception as having honourable character traits that call for respect from oneself and others (Dillon, 2010; Paradise & Kernis, 2002). Moreover, we expected ASR to be associated with well-being as a specific self-evaluation associated with enhanced self-regulation and value-congruent behaviour (Hofmann et al., 2014), and greater connection to community through prosocial acts and reputation as trustworthy, and possibly a greater sense of purpose or meaning in life.

Method

Samples

Demographic sample details on age, gender and country of residence can be found in Table 1. Sample 6 was used to investigate the relationship between ASR and prosocial behaviour and consisted of 251 university students and acquaintances of a student researcher² recruited through RPS, social networking sites and face to face contacts on and off campus. In addition, Sample 4A of non-student adults recruited through Prolific Academic completed a quasi-behavioural measure of prosocial behaviour two-weeks after completing the ASRS and RSES.

Sample 7 was used to investigate the relationship between ASR and forgiveness in response to a betrayal and consisted of 386 adults: 188 UK university students recruited through RPS and 105 Reddit users from Sample 5, with another 93 UK university students and acquaintances of a student researcher³ recruited through RPS and face-to-face on and off campus. As for Sample 5, the pattern of results was similar in the university student community and Reddit sample so the samples were combined.

Sample 4B (described on p.9) was used to investigate the relationship between ASR and well-being.

Measures

The 20-item 5-point Likert Self-report Altruism Scale (Rushton et al., 1981) was completed by Sample 6, alongside the ASRS (α = .85), the RSES (α = .91). It measures the frequency of engaging in altruistic acts towards strangers or acquaintances (α = .85).

A quasi-behavioural measure of prosocial behaviour was completed by Sample 4a, which asked participants to report on whether they had engaged in six typical altruistic activities within the past two years to encourage concrete reflection on behaviours performed, including donating goods/money to the needy, donating blood, buying charity holiday cards, volunteering at a charitable organisation, helping push a stranger’s car that was broken or out of gas (or called for help) and visiting friends/family at a nursing home (adapted from Aquino & Reed, 2002 and Rushton et al., 1981). They were considered to have engaged in prosocial behaviour if they had performed at least one of the activities, following Aquino and Reed (2002).

Forgiveness in response to a betrayal was measured in Sample 7 by asking participants to imagine themselves in a situation in which a close friend had betrayed them, which was adapted from Luchies et al. (2010) (secondary to the investigation, the study also manipulated amends made in the scenario, see Online Resource Appendix J for the scenario verbatim and fuller details of the method) and then indicate their level of forgiveness towards the friend who hurt them by completing the 12-item 5-point Likert Transgression-Related Interpersonal Motivation Scale (McCullough et al., 1998), measuring post-transgression avoidance (α = .87) and revenge motivations (α = .89) underpinning forgiveness. They first completed the ASRS (α = .83) and single 7-point Likert scale global self-esteem item validated by Robins et al. (2001).

Mental well-being (or positive mental health) was completed by Sample 4B using the 14-item 5-point Likert Warwick-Edinburgh Mental Well-being scale (Tennant et al., 2007) (α = .94), which encapsulates affective-emotional as well as cognitive-evaluative dimensions of well-being and psychological functioning (Tennant et al., 2007), alongside the ASRS (α = .86) and the RSES (α = .92).

Statistical Analysis

Pearson zero-order and partial correlations (obtained through multiple linear regression analyses with ASR and self-esteem as predictors) were used to examine the relationship of ASR with altruism scale scores and well-being. A multiple regression model that also included the amends manipulation, and the interaction between amends and ASR was used to examine the relationship between ASR and forgiveness scores. Logistic regression models were used to examine the relationship between ASR and the quasi-behavioural measure of prosocial behaviour, and examine the influence of moral self-variables, principledness and moral personality variables.

Results and Discussion

The correlation between ASR and self-esteem was .60 in Sample 4A, .59 in Sample 4B, .56 in Sample 6 and .55 in Sample 7.

Prosocial Behaviour

As predicted, altruism scale scores were significantly related to ASR (M, SD = 36.79, 6.07) over-and-above self-esteem (M, SD = 28.33, 5.81) (see Table 3)⁷.

Percentages of participants having engaged in each altruistic activity ranged from 16.9% to 50%, except donating goods or money to the needy, which 90.7% of respondents reported having done in the past two years. Because this behaviour was performed by most respondents, it was excluded from the overall measure of altruistic activities within the past two years, and analysed separately. This resulted in 186 respondents (75%) having engaged in at least one of the 5 remaining activities and 62 respondents (25%) not having engaged in any of the 5 activities. The behaviour of donating goods or money to the needy might have been more common because it entailed a greater perceived personal moral obligation due to a lack of engagement being less easily excusable than for the other activities (De Groot & Steg, 2009). Therefore, we might have expected moral self-variables and principledness should relate to this behaviour, but not necessarily the overall measure of altruism, whilst ASR should independently relate to both measures.

A one standard deviation increase in ASR was associated with a 57% increase in the odds of having engaged in altruistic activities within the past two years. Self-esteem was also significantly associated with this outcome but ASR remained significant after controlling for self-esteem (see Table 4)⁷.

Table 4 Results of single and multiple binary logistic regressions for prosocial behaviour measures

None of the moral self-variables, principledness or moral personality variables (including modesty) significantly related to this outcome.

A one standard deviation increase in ASR was associated with a 49% decrease in the odds of not donating goods or money to the needy within the past two years. The relationship remained significant after adjusting for self-esteem, which was not significantly related to the outcome (see Table 4)⁷. After adjusting for internalisation and modesty, which emerged as unique predictors amongst all moral self-variables, principledness and moral personality variables, ASR was still significantly related to donating goods or money (OR (CI_95%) = .91 [.845, .983], Wald = 5.75, p = .016, standardised OR = .55). Importantly, ASR was independently associated with prosocial behaviour over-and-above the influence of moral self and principledness constructs. ASR was not significantly related to gender, and controlling for gender did not change the pattern of results.

With regard to ASR’s relationship to forgiveness, ASR (M, SD = 37.50, 5.91) was significantly negatively related to revenge motivations - including after adjusting for self-esteem⁸ (M, SD = 3.88, 1.57) - but was not significantly related to avoidance (see Table 3). Self-esteem was not related to revenge. There was a main effect of amends but no interactions between ASR (or self-esteem) and amends (see Online Resource Appendix J for a fuller account of the results). These findings further support ASR as being distinct from self-esteem and its unique implications for pro-relationship behaviour (see Kumashiro et al., 2002).

Well-Being

ASR (M, SD = 37.37, 6.55) was indeed found to be significantly related to mental well-being, independently of self-esteem (M, SD = 26.71, 6.40) (see Table 3)^7,9 highlighting its value for psychological outcomes.

General Discussion

The term ‘self-respect’ is widely used in everyday discourse. However, rather surprisingly, the concept it reflects has been the subject of little psychological research, but rather discussion of a more philosophical nature. We have developed a short unidimensional measure of trait appraisal self-respect (ASR), conceptualised as a disposition to perceive oneself as a respectworthy honourable person. We developed items to capture key respectworthy honourable character traits and have presented evidence to support the scale’s internal and test-retest reliability, unidimensional factor structure as well as convergent, discriminant and criterion validity, in line with theoretical expectations.

In this empirical work, trait ASR was found, as predicted, to be related to, yet distinct from related constructs, including measures of moral self, principledness and global self-esteem, supporting its construct validity. ASR was shown to be strongly related but distinct from self-esteem in EFA, bifactor CFA and SEM analyses and correlations with other constructs. We provided support for ASR being a distinctive subdomain of self-esteem, and also showed it to be distinct from academic, social, appearance and physical ability self-esteem facets. As a subjective self-evaluation as being a respectworthy honourable person, ASR is likely influenced by global self-esteem to some extent in addition to having an honourable character. Despite this, ASR uniquely predicted prosocial behaviour and well-being, and showed positive association patterns with moral self-variables and principledness, not shown by self-esteem. Also, in line with our conceptualisation of ASR as a narrower self-evaluation or subcomponent of self-esteem grounded in respectworthy honourable character (see Marsh & Craven, 2006), we found trait ASR to be more consistently related to prosocial behaviour than self-esteem was.

Moreover, trait ASR was related to prosocial behaviour independently of moral self-constructs levels, further supporting the distinction between the moral self and ASR. The desire to maintain a positive self-evaluation as honourable and respectworthy (rather than purely moral) is likely to motivate prosocial behaviour. Given that not all prosocial behaviours entail a strong perceived moral obligation (De Groot & Steg, 2009), trait ASR is likely to prove a useful variable in predicting such behaviour. Trait ASR is a broader construct than moral self-appraisal. Moral courage and the impetus to defend one’s honour and worth are also important aspects of trait ASR that do not necessarily characterise favourable moral self-appraisal. For instance, kind and self-effacing individuals who appraise themselves as moral may not have high ASR if they fail to assert their worth when wronged. Consistent with this argument, we show that ASR related to a lower likelihood of revenge motivations in a betrayal situation, but not of avoidance motivations. A valuing, as opposed to unduly humble, or critical attitude, towards one’s honourable character traits is also likely to be important in ASR, but more research is needed on this topic.

In addition, other findings supported ASR as a self-evaluation that is broader than a moral self-concept. Indeed, in our data, moral self-concept was no longer related to self-esteem after adjusting for ASR. Trait ASR may therefore help strengthen self-esteem in individuals with a moral self-concept through a broader self-evaluation as being a respectworthy honourable person. This is an important finding as moral functioning and self-esteem research have typically proceeded along separate paths (Power, 2004). Trait ASR’s close association with global self-esteem also differentiates it from other self-constructs predicting moral and prosocial behaviour (e.g., moral identity, principledness and virtue-contingent self-esteem). Our research highlights the need to consider individuals’ ASR in interventions seeking to strengthen moral self-characteristics to promote moral behaviour, in order to promote feelings of self-worth.

Promoting ASR, therefore, appears important in the enhancement of self-worth and can benefit personal and societal functioning as it is associated with moral and prosocial behaviour. Also of theoretical significance, trait ASR had relevance for personal well-being beyond self-esteem, which we expected in view of secure feelings of self-worth from a relatively stable and well-founded self-perception as being a respectworthy honourable person who has value and warrants the respect of oneself and others independently of personal affinities or needs, supporting positive psychological functioning (Paradise & Kernis, 2002). This was also expected in view of the benefits of enhanced self-regulation and value-congruent behaviour for positive affect and increased satisfaction with life choices, and a greater connection to community through prosocial acts and reputation as trustworthy. Indeed, we found trait ASR to be related to trait self-control and conscientiousness, again, over-and-above self-esteem. Moreover, better balancing of “vice-virtue” conflicts by favouring the virtuous course of action, has been shown to explain trait self-control’s relationship to affective well-being and life satisfaction (Hofmann et al., 2014).

Our newly developed ASR scale has some potentially important practical implications. For example, it could be used to investigate the implications of trait ASR for enhanced performance, engagement in citizenship behaviours and well-being within the workplace. The ASR scale might also predict risk behaviours that may be perceived as disreputable, such as drug use. ASR could also play an important role in preventing criminal behaviour or other unhealthy psychological or social behaviour in individuals who feel rejected by others. However, trait ASR could also have a darker side in that it is dependent on one’s moral code and one’s perception of what constitutes honourable and respectworthy behaviour. One can imagine that some social groups that do not conform with mainstream society might endorse non-traditional notions of honourable behaviour, for instance certain gangs (e.g., as seen in the Omerta code of the Cosa Nostra, better known as the Italian-American Mafia), and trait ASR could paradoxically be related to criminal behaviour in such groups. Indeed, future research could be extended to such outlying subgroups to understand the relationship between the concept of self and criminal behaviour.

The majority of the research involved the use of self-report scales, which could have inflated correlations due to common-method variance. However, in partial mitigation, many of the correlations presented entailed a two-week temporal separation between the ASR and self-esteem measures on the one hand, and the convergent and criterion measures on the other hand. Overall, the findings indicated major differences between ASR and self-esteem, which implies high discriminatory power. Most research was also cross-sectional, although not all, with the longitudinal data collected showing acceptable test-retest reliability for the ASRS. In addition, we provided evidence that the ASRS is associated with a more secure moral self-image and a more elaborate moral self-schema using objective reaction time and recall measures. A quasi-behavioural measure of prosocial behaviour was also used, where common method variance is less of a problem. Findings may also not generalise to non-Western cultures where being honourable and respectworthy may hold a different meaning, or to groups less well represented in our samples. In particular, some samples were all or predominantly students, but findings were consistent with those obtained in general adult samples. Additional longitudinal research is also needed to further investigate the antecedents and consequences of ASR.

Conclusions

Despite extensive study by philosophers and its popularity in everyday discourse, ASR has not been systematically treated as a distinct construct in its own right. In this article, we showed that trait ASR - conceptualised as a disposition to perceive oneself as a respectworthy honourable person - was related to, yet distinct from, measures of self-esteem, moral self and principledness. Trait ASR was also found to be important for self-worth in moral individuals, as well as for mental well-being and prosocial behaviour beyond self-esteem. The broader focus on having and maintaining a self-perception as a respectworthy honourable person, and its close relationship with self-esteem, differentiates it from other moral self-constructs. Our findings emphasise the need for further research into this psychologically relevant concept. Our newly developed measure of trait ASR shows good psychometric properties and could be used alongside other multidimensional self-esteem measures to improve understanding of global self-esteem, psychological and behavioural outcomes.