Sage Journals: Discover world-class research

Abstract

Lay definitions tend to conceptualize stress as negative, undesirable, and maladaptive. However, contemporary stress models emphasize the differentiation between negative and positive stress responses, known as distress and eustress. Despite prominent theoretical conceptualisations accepting the existence of eustress, the vast majority of stress measures tend to focus exclusively on the distress response. The current study introduces the Adolescent Distress-Eustress Scale (ADES) which holistically captures both aspects of the stress response, bridging the gap between theory and measurement and counteracting the typically negatively focused approach to stress research. The ADES was systematically developed and tested in a socio-educationally diverse sample of 981 adolescents (M_age = 15.19, 50.62% female). The finalized self-report scale consists of two 5-item subscales, individually indexing distress and eustress. Initial psychometric properties of the ADES are promising, and the scale has the potential to meet the needs of researchers, schools, and organizations.

Keywords

measurement and scaling methods research methods social sciences reliability and validity educational psychology applied psychology psychology stress adolescence

Adolescence is a crucially stressful period of the lifespan (Noor & Alwi, 2013; Venning, Eliott, Kettler, & Wilson, 2013). During this time, young people face numerous demands, experiencing numerous psychological, physical, and environmental changes (Moksnes et al., 2016; Noor & Alwi, 2013; Vera et al., 2012). Indeed, in a recent survey of Australian adolescents, “stress” was found to be respondents’ number one personal concern (Bailey et al., 2016). However, while the lay assumption is that “stress” is dysfunctional and detrimental (e.g., Jones & Bright, 2001), theory suggests that stress is not inherently maladaptive.

Defining the Stress Response

In 1974, pioneering researcher Hans Selye defined stress as “the non-specific response of the body to the demands made upon it” (p. 14). Selye argued that the body necessarily produced a response to every demand and therefore considered stress to be ubiquitous and unavoidable (Le Fevre, Matheny, & Kolt, 2003). Crucially Selye’s conceptualization delineated this response into both positive and negative aspects, known as distress and eustress.

Contemporary stress models have retained Selye’s holistic conceptualization, emphasizing the differentiation between positive and negative stress responses. For example, the Transactional Approach (see Lazarus & Folkman, 1984) outlines that an individual’s experience of stress is dependent on their appraisal of their ability to cope with the stressor. When an individual perceives that their coping skills are inadequate, they will experience negative stress. On the other hand, if an individual perceives their coping skills as adequate, they will experience positive stress. Similarly, the Holistic Model (see Nelson & Simmons, 2003) also differentiates positive from negative stress on the basis of individualized appraisal. However, the latter model focusses more on the salient individual differences predicting the stress response. Supporting both models, empirical evidence emphasizes the importance of appraisal in the experience of stress (Lazarus, 1993).

While these models accept the distinction between positive and negative stress responses, they differ in their specific conceptualization of the stress process. This has led to poor comparison across the literature and little replication of empirical findings (Burton & Hinton, 2010). However, while significant variation does exist between models, all incorporate certain key concepts. As such, integrating across Selye’s original work and such contemporary theories, the current study adopts a partial-consensus definition of the stress process, summarized in Figure 1 (see also Branson, Turnbull, Dry, & Palmer, 2018). This definition focuses only on those key elements of the stress process for which there is agreement across the various theoretical models and is thus necessarily broad.

Figure 1.

A visual description of the partial-consensus definition of the stress process.

Here, a stressor is any relevant stimulus that puts a demand on an individual. This stimulus can be physical, psychological, “tangible or mentally evoked” (Meir Drexler & Wolf, 2017, p. 286). Stressors are considered to have no inherent valence, such that the stress response is subjective and dependent upon the individualized appraisal of the demand. The resultant response is delineated into both distress, the negative, undesirable, and harmful response, and eustress, the positive, desirable, and advantageous response. The two responses are considered to be distinct constructs, rather than extremes on a continuum. As such, individuals can simultaneously experience distress and eustress.

Measuring the Stress Response

Despite prominent theoretical conceptualisations accepting eustress, the concept of “positive stress” has received markedly less research interest (e.g., Le Fevre, Kolt, & Matheny, 2006; Le Fevre et al., 2003). Correspondingly, the overwhelming majority of stress measures focus exclusively on what this paper defines as distress. For example, the commonly utilized Perceived Stress Scale (Cohen, Kamarck, & Mermelstein, 1983; Cohen & Williamson, 1988) characterizes stress as a pathological condition. Similarly, another frequently used measure, the Depression Anxiety Stress Scale (DASS; Lovibond & Lovibond, 1995) defines stress as an exclusively negative emotional state. One exception, however, is the Academic Eustress Scale (O’Sullivan, 2011), which focuses on the process of responding positively to academic stressors as well as the positive outcomes of this process. In response to the lack of validated, reliable measures, various authors have used positive and negative emotional states as proxy measures of distress and eustress (e.g., J. R. Edwards & Cooper, 1988; Parker & Ragsdale, 2015).

To the best of our knowledge, only three published scales holistically measure both distress and eustress: the Self-Report Stress Response Questionnaire (Hargrove, Casper, & Quick, 2014), the Valencia Eustress-Distress Appraisal Scale (Rodríguez, Kozusznik, & Peiró, 2013), and the Stress Professionnel Positif et Négatif (De Keyser & Hansez, 1996). However, all three measures have restricted populations of interest, being developed within the context of organizational psychology and specifically focussing on the adult work context. Applying these vocational, adult-focused measures to the adolescent context is inappropriate when considering the unique developmental contexts and idiosyncrasies of young people (e.g., Compas, 1987). There is thus a need for an adolescent-focussed measure that captures the distinction between positive and negative stress.

The Current Investigation

The near-exclusive use of negatively biased measures serves to perpetuate the lack of research on positive eustress. To counteract this negative focus, a more balanced approach is required, which holistically takes into account both the negative and positive aspects of the stress response. The overarching goal of the investigation was therefore to develop a brief, reliable, and valid measure of the adolescent stress response. This approach can be contextualized within the field of Positive Psychology, expanding the exclusively deficit-focussed approach to highlight positive human assets (Seligman & Csikszentmihalyi, 2000; Waters, 2011).

Imposing adult measures on young people discounts the unique developmental context of adolescence (e.g., Compas, 1987). As such, the measure was specifically designed for use in populations aged between 12 and 20 years (as per the South Australian Mental Health Survey definition of adolescence; Venning et al., 2013), with regards to both the content of the scale and the language and format.

The current study introduces the Adolescent Distress-Eustress Scale (ADES). This scale addresses the disjunct between theory and measurement by holistically capturing both aspects of the stress response, with individual subscales indexing distress and eustress (ADES-D and ADES-E, respectively). Specifically, the paper aims to (1) design the ADES by optimizing a preliminary collection of items, (2) evaluate internal and test–retest reliability of the ADES, (3) demonstrate initial construct validity of the measure by assessing convergent and divergent associations, and (4) determine measurement invariance across genders.

Method

The ADES was established following DeVellis’s (2012) practical guidelines for scale development. This framework, which is based on the tenets of Classical Test Theory, outlines four major steps in the development of a questionnaire: defining the constructs, creating, then reviewing the scale items, then evaluating the psychometric properties of the scale. The initial three stages of this process were informed by a series of preliminary qualitative studies, summarized briefly below and described in more detail elsewhere (Branson et al., 2018). The current paper chiefly focusses, however, on the evaluation stage of scale development, describing the optimisation and testing of the ADES.

Item Generation and Refinement

The creation of the ADES was a collaborative enterprise between the research team and the intended adolescent respondents (Compas, Davis, Forsythe, & Wagner, 1987; Mason & Danby, 2011). To identify potential effect indicators of the stress response, individual interviews were conducted with 20 adolescents (50% female, 13-20 years old), to elicit their personal experience of stress. These interviews focussed on the phenomena that adolescents identified as compellingly and effectively differentiating between distress and eustress. On the basis of these qualitative results, 463 candidate items were generated for consideration in the final scale.

These items were then sent to subject matter experts for feedback regarding content validity, clarity, and developmental appropriateness. Furthermore, cognitive interviews were conducted with 12 adolescents (50% female, 13-19 years old) to identify and amend the elements of the draft questionnaire proving problematic for the intended respondents. Based on this review process, the items were refined, improved, and combined to form a cohesive preliminary scale consisting of 25 candidate items per subscale.

Participants and Procedure

To obtain a broad, generalisable sample, students (over the age of 13) from three different educational institutions of varying socio-educational advantage (independent private school, university, and public government school) were invited to take part in the online survey. Ethical considerations emphasized anonymity, confidentiality, informed consent (participant, and where necessary parental), and safeguarding participants’ emotional wellbeing. All procedures were approved by The University of Adelaide School of Psychology: Human Research Ethics Subcommittee (Code Numbers: 17/10 and 17/65) and The Department of Education and Child Development (Reference CS/17/000,747-1.14).

Split samples procedure

For analysis purposes, the total sample (N = 981) was randomly split into two approximately equivalent subsamples. One half, the Development Subsample, was used for item selection and scale optimization through exploratory factor analysis (EFA). The second, Cross-checking Subsample was used to support these results through confirmatory factor analysis (CFA). The size of each subsample exceeded the commonly recommended minimum of 300 participants required for factor analysis (e.g., DeVellis, 2012).

In addition, all students recruited from the University were asked to complete the preliminary ADES a second time within 1 week of the initial questionnaire. This Follow-Up Subsample was used to evaluate the test–retest reliability of the scale. Internal reliability and validity were evaluated using the total sample.

Description of sample

The socio-demographic characteristics of participants are presented in Table 1.

Table 1.

Sample Sociodemographic Characteristics.

		Subsamples
Characteristic	Main sample (N = 981)	Development subsample (n = 491)	Cross-checking subsample (n = 490)	Follow up subsample (n = 83)
Age, M (SD)	15.19 (1.70)	15.19 (1.96)	15.19 (1.70)	18.73 (0.86)
Gender, n (%)
Male	477 (48.62)	237 (48.27)	240 (48.98)	19 (75.90)
Female	497 (50.62)	251 (51.12)	246 (50.20)	63 (22.89)
Other	7 (0.71)	3 (0.61)	4 (0.82)	1 (1.20)
Language background, n (%)
English	763 (77.78)	383 (78.00)	380 (77.55)	51 (61.45)
Other	218 (22.22)	108 (22.00)	110 (22.45)	32 (38.55)
Educational institution, n (%)
University	93 (9.48)	46 (9.37)	47 (9.59)	83 (100.00)
Private school	563 (57.39)	282 (57.43)	281 (57.35)	0 (0.00)
Public school	325 (33.13)	163 (33.20)	162 (33.06)	0 (0.00)

Materials

In addition to the preliminary ADES described above, the online self-report questionnaire consisted of the six established scales described below.

Short-Form Marlowe-Crowne Social Desirability Scale

Socially desirable responding, or the tendency to deny socially undesirable traits and/or emphasize socially desirable traits (Nederhof, 1985), is a common source of bias affecting the validity of self-report measures. To investigate the influence of socially desirable responding on the ADES, the Reynolds (1982) short-form of the Marlowe-Crowne Social Desirability Scale (MC-SDC-13) was included in the online questionnaire. This reliable and valid short form of the original scale (Crowne & Marlowe, 1960), consists of 13 True-False items and has been used in samples as young as 10-years-old (Wang, Fu, Zhang, & Kou, 2015). The reliability in the current sample was α = .66.

Academic Eustress Scale

The Academic Eustress Scale (AES; O’Sullivan, 2011) defines eustress as both the process of responding positively to stressors as well as the positive outcomes of this process. Specifically, this scale focusses on eustress related to academic stressors in adolescent and young adult populations. Responses to the 10-item scale range from never (0) to always (5), with higher mean scores indicating greater experience of academic eustress. The reliability in the current sample was α = .83.

Perceived Stress Scale

The Perceived Stress Scale (PSS-10; Cohen et al., 1983; Cohen & Williamson, 1988) frames stress as a negative, undesirable, pathological phenomenon, measuring the extent to which one finds their life to be unpredictable, uncontrollable, and overloading. Responses to the 10-item scale range from Never (0) to Very Often (4) with greater sum scores indicating greater experience of negative stress. The PSS-10 has robust psychometric properties in both adult and adolescent samples (Cohen et al., 1983; Cohen & Williamson, 1988; Williams, Turner-Henson, Langhinrichsen-Rohling, & Azuero, 2017). The reliability in the current sample was α = .87.

General Self-Efficacy Scale

Self-Efficacy, defined as “optimistic beliefs about individual ability to deal with tasks at hand” (Luszczynska, Piko, & Januszewicz, 2011, p. 2559) was assessed via the General Self-Efficacy Scale (GSES; Schwarzer & Jerusalem, 1995). Responses to the 10-item scale range from not at all true (1) to exactly true (4), with higher total sum score indicating greater self-efficacy. The measure is considered valid for use in youths, including adolescents (Schwarzer & Jerusalem, 1995). The reliability in the current sample was α = .85.

Orientation to Life Questionnaire

Sense of Coherence (SOC) is defined as

a global orientation that expresses the extent to which one has a pervasive, enduring though dynamic feeling of confidence that one’s internal and external environments are predictable and that there is a high probability that things will work out as well as can be reasonably be expected. (Antonovsky, 1979, p. 123)

This construct was measured using the Orientation to Life Questionnaire (SOC-13; Antonovsky, 1979), which assesses SOC along three dimensions: comprehensibility, meaningfulness, and manageability. Participants respond to 13 items along a semantic differential scale with diametrically labeled continuum ends. Higher sum scores indicate greater SOC. The scale has shown acceptable reliability in studies with adolescents (e.g., Margalit & Eysenck, 1990; Moksnes, Espnes, & Haugan, 2014). The reliability in the current sample was α = .79.

Big Five Inventory

The Five Factor Model (e.g., McCrae & Costa, 1997) was used to conceptualize personality in the current study. This model describes personality along five dimensions: Openness to experience, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. Utilizing the Five Factor theoretical framework, personality was measuring using the Big Five Inventory (BFI; John, Naumann, & Soto, 2008). The scale consists of 44 items with responses ranging from disagree strongly (1) to agree strongly (5). Subscale scores are computed as the average of the corresponding items, with higher scores indicating greater endorsement of the respective personality trait. The BFI is accessible to children as young as 10 (Soto, John, Gosling, & Jeff, 2011). Reliability in the current sample ranged from α = .70 to .84.

Data Analysis

Prior to data analysis, data were screened for obviously frivolous responses (Fan et al., 2006). Outliers for each variable were identified and trimmed using the Hoaglin and Iglewicz (1987) modification of the Tukey (1977) Outlier Labeling rule.¹ Missing data were managed via Listwise deletion (Allison, 2001; Schreiber, 2008). CFA was conducted using the software Amos Graphics (Arbuckle, 2017). All remaining analyses were conducted in SPSS Version 24 (SPSS Inc., 2017).

Measure optimisation

In the first stage of analysis, the scale was optimized to be suitably parsimonious, with an ideal aim of 5 items per subscale (Costello & Osborne, 2005). For balance, equal numbers of items were included in the distress and eustress subscale.

Following the recommendations of DeVellis (2012), the preliminary item pool was screened for deficient psychometric properties prior to entry into factor analysis, according to three elimination criteria. First, inter-item correlations were considered. Considering the distress and eustress items separately, items would be discarded if they shared inconsistent correlation patterns (i.e., positive correlation with some items and negative relationship with others). Furthermore, in the case of multicollinearity (r ≥ .8), only one item would be retained. Second, to ensure appropriate distribution and variance of item responses, highly skewed (skew > |2|) and/or leptokurtic (kurtosis > 2) items would be discarded. Finally, items sharing more than a moderate correlation with the MC-SDC-13, thereby being influenced by social desirability bias, would be discarded.

The remaining items of the preliminary ADES were subjected to a series of iterative EFAs (maximum likelihood extraction method) using the Development Subsample. As distress and eustress are expected to correlate, the solution was rotated using an oblique direct oblim rotation (Δ = 0). To produce a suitably parsimonious scale, items were deleted iteratively according to factor loadings, until no cross loadings exceeded ≥ 0.3 and either all items loaded on one factor ≥ 0.7 or a minimum of 5 items loaded on each factor ≥ 0.5 (Costello & Osborne, 2005). Furthermore, beyond exclusively statistical criteria, the appropriate theoretical alignment of items and the interpretability of the retained pool as a cohesive questionnaire were also considered (DeVellis, 2012).

Following EFA, CFA (maximum likelihood estimation method) was conducted using the Cross-checking Subsample to confirm the structure of the ADES. Model fit was evaluated primarily using the root mean square error of approximation (root mean square error approximation [RMSEA]), the comparative fit index (CFI), and the Tucker–Lewis Index (Tucker–Lewis index [TLI]). A RMSEA less than 0.08 combined with a CFI and TLI greater than 0.95 was considered to indicate good model fit (Hu & Bentler, 1999; Schreiber, 2008). Following the recommendations of Schreiber (2008), the χ² statistics was also reported; however, this value was not used to judge model fit as it is extremely sensitive to sample size (e.g., Cheung & Rensvold, 2002). For comparison, a one-factor model (all items loading on a single “stress response” factor) and a second-order hierarchical model (items loading on two subscales, which load on a single higher-order “stress response” factor) were also estimated. Model comparison was assessed using the χ² difference test. However, as this test is also sensitive to sample size, differences between models were only considered practically meaningful if ΔCFI ≥ 0.01 (Cheung & Rensvold, 2002).

Measure testing

Next, the psychometric properties of the optimized 10-item measure were tested.

Reliability

To estimate internal consistency, Cronbach’s alpha was computed for the finalized subscales using the re-combined total sample. DeVellis (2012) and Rattray and Jones (2005) suggest an alpha value of 0.7 as a minimum for novel scales.

To assess the temporal stability of the ADES, test–retest coefficients (Pearson’s r) were calculated in a convenience subsample consisting only of the university student cohort (n = 83; see Table 1 for specific cohort demographics). A correlation coefficient exceeding 0.8 indicates good temporal stability, though all values exceeding 0.7 are considered useful (De Vriendt et al., 2011).

Validation

Evidence for construct validity was provided by demonstrating that the ADES (1) was associated with other measures designed to measure the same thing (convergent validity) and (2) related as expected with other measures of non-stress constructs (discriminant validity; Churchill, 1979).

As there is no existing measure of distress and eustress in adolescents, no scale reflects identical constructs to the ADES. The AES and the PSS-10 were consequently selected as convergent validity constructs, as their theoretical conceptualization of “stress” is the closest analogue to each of the ADES subscales. Validity coefficients (Pearson’s r) were calculated between the ADES-E and the AES and the ADES-D and the PSS-10, with convergent validity determined by relatively strong positive correlations. However, according to Classical Test Theory, these correlations should not exceed $\sqrt{α}$ (DeVellis, 2006).

To examine discriminant validity, validity coefficients (Pearson’s r) were calculated between the ADES and three related, non-stress constructs: personality, sense of coherence, and self-efficacy. These constructs were chosen as validation items as they (1) share significant conceptual and theoretical overlap and empirically demonstrated relationships with the stress response and (2) were considered key concepts associated with stress by the adolescent participants of the preliminary qualitative studies. Extant literature cites strong theoretical and empirical links between the stress response and both self-efficacy (e.g., Luszczynska et al., 2011; Nelson & Simmons, 2003) and sense of coherence (e.g., Antonovsky, 1979; Moksnes et al., 2014), such that the constructs are negatively related to distress and positively related to eustress. Furthermore, each of the Big 5 personality traits were expected to relate to the stress response. Conscientiousness and Neuroticism share the greatest theoretical and conceptual overlap with stress, with the former being positively related with eustress (e.g., Rice, 1999; Saksvik & Hetland, 2011) and the latter with distress (e.g., B. D. Edwards, Franco-Watkins, Cullen, Howell, & Acuff, 2014; Saksvik & Hetland, 2011). Openness, Extraversion, and Agreeableness share little conceptual overlap with the stress response. In the sense that they broadly represent positive interactions with the world, these traits may relate positively to eustress and negatively to distress; however, these relationships were expected to be weak-to-negligible (Saksvik & Hetland, 2011). Appropriate discrimination was considered according to the traditional Campbell and Fiske (1959) cut-off, whereby a correlation less than 0.8 demonstrates evidence of discriminant validity.

Measurement invariance

Measurement invariance refers to the extent to which a scale performs equivalently across different groups of respondents. If measurement invariance is not established, one cannot decisively ascertain if score differences across groups reflect true construct difference between those groups of differences in the scale’s performance across the groups (Cheung & Rensvold, 2002; DeVellis, 2006). As extant literature suggests that gender differences should be expected on ADES scores (e.g., Almeida & Kessler, 1998; Flook, 2011), the current study considered the measurement invariance of the ADES across gender groups via multi-group confirmatory factor analysis (MCFA). MCFA examines the changes in fit indices as increasingly restrictive cross-group constraints are progressively imposed on the measurement model (Brown, 2015; Cheung & Rensvold, 2002).

According to the recommendations of Vandenberg and Lance (2000), three increasingly restrictive models were iteratively examined to determine the degree of model invariance across genders. In the first model, only the measurement model pattern is constrained to be equal across groups (known as configural invariance), then the factor loadings (metric invariance), and finally the factor variances and covariance (variance-covariance invariance). As with regular CFA, meaningful model differences were considered at ΔCFI ≥ 0.01 (Cheung & Rensvold, 2002).

Results

Measure Optimisation

Prior to performing EFA, two items were eliminated from the preliminary item pool for being strongly negatively skewed. No items were found to display inconsistent correlation patterns, share strong multicollinearity, or be overly influenced by social desirability bias.

Suitability of the remaining 48 items for EFA was established, with the Kaiser-Meyer-Olkin value (Kaiser, 1974) exceeding 0.6 (KMO = 0.95) and the Bartlett’s Test of Sphericity (Bartlett, 1954) reaching statistical significance. Data extraction revealed the presence of 7 factors with Eigenvalues greater than 1. The first two factors contained 44.9% of the total variance in the analysis (factor one and two accounted for 32.29% and 12.63% of variance respectively). The third factor accounted for 4.46% of the variance and each subsequent factor less than 3%. Inspection of the scree plot was inconclusive, suggesting either a two or three factor solution. Parallel analysis supported a three-factor solution, with three components with Eigenvalues exceeding the corresponding criterion values for a randomly generated data matrix of the same size (50 variables, 491 respondents). To determine optimal factor structure, both the two- and three-factor solutions were examined. Comparing the factor loading tables, the two-factor solution resulted in stronger factor loadings and less cross-loadings. Given this comparison, the large differences between variance accounted for by factors one and two compared to factor three, and the increased interpretability and theoretical-alignment of a two-factor solution, subsequent EFA fixed the number of factors to two.

The items loading on the first factor were predominantly intended to measure distress, while the items loading on the second factor were predominantly intended to measure eustress. This indicated Factor 1 represents Distress, while Factor 2 represents Eustress. All items loaded on one factor ≥ 0.32, establishing that they share more than 10% overlapping variance with other items in the factor.

Next, to produce a suitably parsimonious scale, item deletion occurred iteratively. Items with the lowest factor loadings were dropped in sequence until no item showed cross loading ≥ 0.3 and 5 items loaded on each factor ≥ 0.5 (Costello & Osborne, 2005). In addition, attention was paid to the theoretical alignment of items and the interpretability of the remaining items as a cohesive questionnaire.

The final factor solution after oblique rotation (see Table 2) accounted for 64.70% of the variance. The correlation between the factors was weak (r = -.32), suggesting that the subscales serve as suitably independent dimensions.

Table 2.

Pattern Matrix for EFA With Direct Oblim Rotation of the Final Two Factor Solution of the Retained Preliminary ADES Items.

	Factor
Item	1	2
I felt anxious.	.86	.03
I felt overwhelmed.	.79	.02
I felt panicked.	.78	.05
I was frustrated with myself.	.68	−.02
My mind was racing out of control.	.68	−.12
I felt the outcome was worth the effort.	.04	.78
I felt determined.	.03	.73
I felt proud for dealing with the pressure.	−.13	.72
I felt motivated.	.09	.70
I was satisfied with how I dealt with the pressure.	−.12	.66

Note. Major loadings for each item are bolded. Factor correlation r = −.32. EFA = exploratory factor analysis; ADES = Adolescent Distress-Eustress Scale.

To confirm the 10-item, two-factor oblique structure found in EFA (see Figure 2), CFA was conducted using the cross-checking subsample. The two-factor model demonstrated acceptable model fit; Table 3 summarizes the latent factor loadings and fit indices. Furthermore, neither the one-factor nor the hierarchical model meaningfully improved data fit (Table 4). Together, these results support the two-factor oblique model found via EFA as the most appropriate design of the ADES.

Figure 2.

Two factor oblique model with 10 indicator items.

Table 3.

Latent Factor Loadings and Fit Indices in CFA for the Final 10-Item Measure (See Figure 2 for the Estimated Model).

Factor/question		Estimates
Eustress
I felt motivated	λ_E1	.77
I felt the outcome was worth the effort	λ_E2	.66
I was satisfied with how I dealt with the pressure	λ_E3	.62
I felt determined	λ_E4	.75
I felt proud for dealing with the pressure	λ_E5	.60
Distress
My mind was racing out of control	λ_D1	.74
I felt panicked	λ_D2	.80
I felt overwhelmed	λ_D3	.79
I felt anxious	λ_D4	.76
I was frustrated with myself	λ_D5	.69
Latent factor covariances
Distress~Eustress	Ф_{D, E}	−.34
Model Fit
RMSEA [90CI]		.07 [.06, –.09]
CFI		.95
TLI		.94
χ² (df)		123.41** (34)

Note. All estimates are standardized. CFA = confirmatory factor analysis; RMSEA = root mean square error of approximation; CI = confidence interval; CFI = comparative fit index; TLI = Tucker–Lewis Index.

p < .01.

Table 4.

Model Fit Statistics for a Two-Factor Model, One-Factor Model, and Second-Order Hierarchical Model of ADES Items.

Model	Comparison	χ² (df)	RMSEA [90CI]	CFI	TLI	Δχ² (df)	ΔCFI
1. Two-factor model		123.41** (34)	.07 [.06, –.09]	.95	.94
2. One-factor model	2-1	721.17** (35)	.20 [.19, –.21]	.65	.55	597.76** (1)	−.30
3. Second-order hierarchical model	3-1	123.41** (34)	.07 [.06, –.09]	.95	.94	N/A^a	.00

Note. ADES = Adolescent Distress-Eustress Scale; RMSEA = root mean square error of approximation; CI = confidence interval; CFI = comparative fit index; TLI = Tucker–Lewis Index.

Models 1 and 3 are equivalent and cannot be distinguished on statistical grounds; comparison must therefore be based on theory and interpretability.

p < .01.

Final measure and instructions

At the conclusion of the measure optimisation process, the ADES was finalized to consist of two correlated subscales each consisting of 5 items (see Table 5). The scale is evaluative rather than prescriptive, exclusively describing the adolescent stress response rather than offering any clinical or diagnostic criterion.

Table 5.

Final 10-Item ADES Measure.

Item	Question
E1	I felt motivated.
D1	My mind was racing out of control.
E2	I felt the outcome was worth the effort.
E3	I was satisfied with how I dealt with the pressure.
D2	I felt panicked.
D3	I felt overwhelmed.
D4	I felt anxious.
E4	I felt determined.
E5	I felt proud for dealing with the pressure.
D5	I was frustrated with myself.

Note. Participants were given the following instructions: “These questions are about how you respond to pressure. Everybody responds to pressure differently at different times. Pressure can be good for you, bad for you, or a bit of both. For each item below, please choose the answer that best describes how you responded to pressure in the last 7 days.” Each item is scored on a 5-point Likert-type scale; only the two extremes and the midpoint are labeled: not like me (0), somewhat like me (2), and very much like me (4). Scores are computed as the sum of the 5 corresponding items, and results presented separately across subscales: ADES-Distress = Sum(D1, D2,D3,D4,D5); ADES-Eustress = Sum(E1, E2,E3,E4,E5). ADES = Adolescent Distress-Eustress Scale.

Table 6 displays the descriptive statistics for the ADES in the current sample.

Table 6.

Descriptive Statistics of the ADES in the Current Sample (N = 981).

	M	SD	Minimum	Maximum	Skewness	Kurtosis	Interquartile range
ADES-E	10.43	4.60	0.00	20.00	−0.11	−0.53	6.50
ADES-D	9.08	5.39	0.00	20.00	0.18	−0.88	8.00

Note. ADES subscale scores could theoretically range from 0 to 20, with higher scores indicating greater experience of that aspect of the stress response. ADES = Adolescent Distress-Eustress Scale.

Measure Testing

Using the final 10-item scale, reliability, validity, and measurement invariance were evaluated.

Reliability

Estimates of internal consistency were computed for the finalized subscales using the re-combined total sample. According to DeVellis’s (2012) conventions, both subscales had very good reliability (ADES-D: α = .87; ADES-E: α = .83).

The Follow up subsample completed the ADES a second time within 1 week of the initial questionnaire (mean number of days between Time 1 and Time 2 was 3.31, SD = 1.17). Test–retest reliability was strong for both the distress subscale, r(81) = .86, p < .01, and the eustress subscale, r(81) = .81, p < .01, indicating good temporal stability of the ADES scores.

Validity

The ADES was appropriately correlated with the convergent validity scales. As expected, there were strong positive relationships between the ADES-E and the AES, r(874) = .60, p < .001, and between the ADES-D and the PSS-10, r(874) = .68, p < .001. Neither of these correlations exceeded $\sqrt{α}$ for their respective subscales (ADES-E = .91, ADES-D = .93). These results provide evidence for the convergent validity of the ADES.

Table 7 summarizes the expected relationships between the ADES and the three individual difference variables (self-efficacy, sense of coherence, and personality), based on the direction and relative strength of the correlation. All correlations were below .80, providing evidence for discriminant validity and indicating that the ADES is sufficiently distinct from these related, non-stress constructs (Campbell & Fiske, 1959). Encouragingly, the ADES showed comparable or superior discriminant validity when compared to the existing stress measures (see Supplemental Table S-1 for correlations between the validation constructs and the PSS and AES).

Table 7.

Evidence for Discriminant Validity: Predicted and Observed ADES Correlations With Individual Difference Variables.

	Expected pattern of results		Observed correlation
	Distress	Eustress	Distress	Eustress
Self-efficacy	– – –	+ + +	−.39**	.46**
Sense of coherence	– – –	+ + +	−.53**	.40**
Personality
Openness to experience	–	+	.02	.23**
Conscientiousness	– –	+ +	−.19**	.48**
Extraversion	–	+	−.15**	.26**
Agreeableness	–	+	−.06	.25**
Neuroticism	+ +	– –	.66**	−.31**

Note. Predictions are based on the expected pattern.—indicates a negative correlation is expected; + indicates a positive correlation is expected. Relative strength is indicated by the number of symbols. Listwise n = 876. ADES = Adolescent Distress-Eustress Scale.

p < .01.

In addition, the results generally adhered to the expected pattern of correlations, with some exceptions. As expected, the ADES subscales shared relatively stronger correlations with the more similar constructs of Self-Efficacy, SOC, Conscientiousness, and Neuroticism. While these relationships were in the direction predicted, the strength of the relationships between Eustress~Conscientiousness and Distress~Neuroticism were stronger than expected. However, as expected, the weakest and non-significant correlations are with the least similar variables: Openness, Agreeableness, and Extraversion.

Measurement invariance

While participants had the option to indicate “Other” when reporting gender, this group was too small in size (n = 7) to include in the analysis. As such, only male and female participants were considered when evaluating the measurement invariance of the ADES across genders.

MCFA results (Table 8) indicated that the measurement invariance constraints resulted in no substantial decrement in model fit, indicating that the ADES had appropriate equivalency across genders.

Table 8.

Fit Indices and Difference Statistics for Measurement Invariance Models by Gender.

	Model description	Comparison	χ² (df)	Δχ² (df)	CFI	ΔCFI
1	Configural invariance^a		237.82** (68)		.96
2	Metric invariance^b	2-1	245.63** (76)	7.81 (8)	.96	.00
3	Variance-Covariance invariance^c	3-2	250.20** (79)	4.57 (3)	.96	.00

Note. CFI = comparative fit index.

Measurement model pattern constrained across gender group.

Model 1 + Factor loadings constrained across gender group.

Model 2 + Variances and covariance between factors constrained to be equal across gender group.

p < .01.

As measurement invariance was established, Hotelling’s T ² was run to determine the effect of gender on the ADES (see Table 9 for descriptive statistics according to gender).

Table 9.

Descriptive Statistics for the ADES According to Gender.

	ADES-E		ADES-D
	M	SD	M	SD
Female^a	10.19	4.54	10.67	5.21
Male^b	10.68	4.65	7.38	5.02

Note. ADES subscale scores could theoretically range from 0 to 20, with higher scores indicating greater experience of that aspect of the stress response. ADES = Adolescent Distress-Eustress Scale.

n = 497.

n = 477.

The differences between genders on the combined dependent variables was statistically significant, F(2, 971) = 51.14, p < .001, Wilks’ λ = .19, partial η² = .10. Using Bonferroni adjusted α level of .025, post hoc testing showed females scored higher on the ADES-D (M_difference = 3.29, 95% CI [2.56, 4.03], p < .001), but no statistically significant difference was found between genders for ADES-E scores (M_difference = .49, 95% CI [-1.15, 0.17], p = .10).

Discussion

The ADES was systematically developed and tested in a socio-educationally diverse sample of 981 adolescents. This scale was specifically designed for adolescent participants, with input from young people at every stage of item generation and scale refinement.

The first aim of the current study was to design the ADES from a collection of preliminary items. The scale was optimized using a pre-defined, iterative procedure incorporating item performance statistics and EFA. These results were then cross-checked in a separate subsample, with a two-factor oblique model supported as the most appropriate design of the ADES. The finalized scale consists of two 5-item subscales, which individually index distress and eustress. The two subscales were only weakly negatively correlated, suggesting that the scales are related, but suitably independent dimensions.

Initial psychometric properties for the ADES are promising. Addressing Aim 2, the internal reliability and temporal stability of both subscales was very good and exceeded the minimum requirements for a novel scale (DeVellis, 2012; De Vriendt et al., 2011; Rattray & Jones, 2005). Furthermore, results provided promising initial evidence for construct validity. Addressing Aim 3, the ADES was strongly correlated with established stress measures and related as expected with other non-stress constructs. Finally, in investigating Aim 4, the scale demonstrated measurement invariance across genders. This indicates the score differences found between males and females using the ADES may be interpreted to indicate true differences in the stress response, rather than as artifacts of the scale’s performance across groups. This is pertinent given the current female participants were found to have significantly higher ADES-D scores.

Implications

The ADES is, to the best of our knowledge, the first measure that holistically takes into account both the positive and negative aspects of the adolescent stress response. As such, this measure serves to bridge the gap between theory and measurement, more appropriately reflecting the two-factor approach of prominent conceptualisations of stress (e.g., Lazarus & Folkman, 1984; Nelson & Simmons, 2003; Selye, 1974). Furthermore, by highlighting the positive aspects of stress, the ADES serves to counteract the negative-focus and provide a more balanced approach to stress research.

Limitations and Future Research Directions

While the current results are promising, it is recognized that demonstrating the psychometric properties of a novel scale is an ongoing, cumulative effort (DeVellis, 2012). Several important considerations should be taken when interpreting the results of the present study.

Restrictive sampling

Attempts were made to avoid restricted sampling by considering both the size and the composition of the development sample (Cohen et al., 1983; DeVellis, 2012). However, the present sample was relatively homogeneous with regard to several demographic factors, most pertinently cultural and language diversity. In the current sample, 77.8% of participants exclusively spoke English at home, exceeding the national rate of 72.7% (Australian Bureau of Statistics, 2017). Furthermore, by sampling from exclusively educational contexts, adolescents in the workforce, vocational training, and those unengaged in any formal system were overlooked. In addition, all participants were volunteers and the majority required parental consent, likely leading to selection bias.

These issues of restrictive sampling were compounded in the examination of test–retest reliability. Given the pragmatic restrictions around collecting data in schools, the analysis was performed on a convenience subsample of only university students, leading it open to several limitations such as non-generalization and bias (De Vriendt et al., 2011). Furthermore, participant drop-out between initial and follow-up assessment was potentially selective. For example, participants may have dropped out due to higher levels of stress (Laferton, Stenzel, & Fischer, 2018).

Together these sampling limitations constrain the generalizability of the current results. Researchers utilizing the ADES should thus consider how their specific research situation differs from the current setting, how these differences may affect the validity of the scale, and the implications of this on the research conclusions (DeVellis, 2012). Future work, should look to reproduce the current findings in a broader, diverse, more generalisable sample. A further priority is to examine the psychometric properties of the ADES in specific populations, such as cross-cultural and Indigenous groups or in adolescents not engaged in the education system.

Further validation work

Validation of a scale is a long-term process (Peacock & Wong, 1990); the current study provides only initial support for construct validity and future research must examine a wider range of constructs. Furthermore, by only including one type of measurement method (self-report), the current study cannot account for common-method biases (Churchill, 1979), defined as “variance that is attributable to the measurement method rather than to the constructs the measures represent” (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003, p. 879). Further work should therefore look to determine the associations of the ADES with non-self-report measures of the same constructs, such as parent- or teacher-report scales.

Influence of contextual factors

As part of the development study, the current participants completed the scale together with all preliminary, subsequently discarded, items. This unique condition likely exerted an influence on pertinent contextual factors, such as respondent fatigue, question order, and motivation, thereby effecting responses to the scale items (DeVellis, 2012). Replication of results utilizing only the finalized ADES is therefore necessary.

Clinical cut-offs and norms

As the ADES was developed as an exclusively descriptive tool, no specific clinical cut-offs or diagnostic criteria were established. Given then that the units of the ADES are arbitrary, individual scores viewed in isolation may not provide a researcher and/or clinician with adequate meaning. Future research could look to develop population norms, which would impute more meaning into individual scores (Churchill, 1979). Furthermore, researchers may look to develop threshold levels for intervention purposes. While not diagnostic criteria, such thresholds would identify individuals likely to benefit from intervention (Kern, Benson, Steinberg, & Steinberg, 2016).

Conclusion

Limitations notwithstanding, the initial results presented here suggest the ADES as a brief, reliable, and psychometrically sound scale. Given the clarity and simplicity of both delivery and scoring, this self-report scale has the potential to meet the needs of researchers, schools, and other adolescent-focused organizations in the fields of both education and psychology. In conclusion, with replication in broader samples and further validation the ADES provides a promising tool for both theory and practice.

Supplemental Material

Supplemental_Material – Supplemental material for The Adolescent Distress-Eustress Scale: Development and Validation

Supplemental material, Supplemental_Material for The Adolescent Distress-Eustress Scale: Development and Validation by Victoria Branson, Matthew J. Dry, Edward Palmer and Deborah Turnbull in SAGE Open

Footnotes

This paper is based on the first author’s thesis for the degree of Doctor of Philosophy/Master of Psychology (Clinical) at The University of Adelaide. We are grateful to the adolescents who participated in the study and to the school and university staff who facilitated and supported the data collection.

Authors’ Note

Prior Dissemination—some of the data from this paper was presented at 38 th Stress and Anxiety Research Society Conference in Hong Kong (July 2017).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This work was supported through the provision of an Australian Government Research Training Program Scholarship and the Healthy Development Adelaide and Channel 7 Children’s Research Foundation (PhD top-up scholarship).

ORCID iD

Victoria Branson

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Victoria Branson is a PhD/master of Psychology (Clinical) candidate at the University of Adelaide. Her primary areas of research and clinical interest are in the promotion of healthy youth development and wellbeing and the application of Positive Psychology to educational settings.

Matthew J. Dry completed his PhD in Psychology in 2007 at the University of Adelaide. His research has focused on the mathematical modelling of cognitive and perceptual processes,psychopharmacology,and pedagogical research into tertiary education teaching and learning.

Edward Palmer,an associate professor publishes regularly in the field of technology in education in high quality journals. He has 20 years of experience in the design and evaluation of teaching initiatives and has played a key role in developing surveys and evaluation methodologies for national projects.

Deborah Turnbull,professor was awarded the chair in Psychology at the University of Adelaide in 2005 and has been researching in the area of health and clinical psychology for over 20 years. Turnbull’s work has been published in high quality outlets including The Lancet,International Journal of Behavioural Medicine,and Medical screening. Her research group regularly presents at the peak international meeting relevant to her field,including the International Congress of Behavioural Medicine.

References

Allison

(2001). Missing data (Vol. 136). Thousand Oaks, CA: Sage.

Almeida

D. M.

Kessler

R. C.

(1998). Everyday stressors and gender differences in daily distress. Journal of Personality and Social Psychology, 75, 670-680.

Antonovsky

(1979). Health, stress and coping: New perspectives on mental and physical well-being. San Francisco, CA: Jossey-Bass.

Arbuckle

J. L.

(2017). Amos (Version 24.0.0). Chicago, IL: IBM SPSS.

Australian Bureau of Statistics. (2017). 2016 census quickstats. Retrieved from https://quickstats.censusdata.abs.gov.au/census_services/getproduct/census/2016/quickstat/036

Bailey

A.-M. B.

Cave

Fildes

Perrens

Plummer

Wearring

(2016). Mission Australia’s 2016 youth survey report. Sydney, New South Wales: Mission Australia.

Bartlett

M. S.

(1954). A note on the multiplying factors for various chi square approximations. Journal of the Royal Statistical Society: Series B, 16, 296-298.

Branson

Turnbull

Dry

M. J.

Palmer

. (2018). How do young people experience stress? A qualitative examination of the indicators of distress and eustress in adolescence. International Journal of Stress Management. Advance online publication. doi: 10.1037/str0000102

Brown

T. A.

(2015). CFA with equality constraints, multiple groups, and mean structures. In Brown

T. A.

(Ed.), Confirmatory factor analysis for applied research (pp. 236-319). New York, NY: Guilford Press.

10.

Burton

R. F.

Hinton

J. W.

(2010). Up the tower of psychobabel: Does lexical anarchy impede research into “psychological stress?” Medical Hypotheses, 74, 644-648. doi:10.1016/j.mehy.2009.10.043

11.

Campbell

D. T.

Fiske

D. W.

(1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81-105. doi:10.1037/h0046016

12.

Cheung

G. W.

Rensvold

R. B.

(2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 9, 233-255. doi:10.1207/S15328007SEM0902_5

13.

Churchill

G. A.

Jr. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16, 64-73. doi:10.2307/3150876

14.

Cohen

Kamarck

Mermelstein

(1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 385-396. doi:10.2307/2136404

15.

Cohen

Williamson

(1988). Perceived stress in a probability sample of the United States. In Spacapam

Oskamp

(Eds.), The social psychology of health: Claremont symposium on applied social psychology (pp. 31-67). Newbury Park, CA: Sage.

16.

Compas

B. E.

(1987). Stress and life events during childhood and adolescence. Clinical Psychology Review, 7, 275-302.

17.

Compas

B. E.

Davis

G. E.

Forsythe

C. J.

Wagner

B. M.

(1987). Assessment of major and daily stressful events during adolescence: The Adolescent Perceived Events Scale. Journal of Consulting and Clinical Psychology, 55, 534-541. doi:10.1037/0022-006x.55.4.534

18.

Costello

A. B.

Osborne

J. W.

(2005). Best practices in exploratory factor analysis: Four recommendations for getting the most from your analysis. Practical Assessment, Research & Evaluation, 10(7). Retrieved from http://pareonline.net/getvn.asp?v=10&n=7

19.

Crowne

D. P.

Marlowe

(1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349-354. doi:10.1037/h0047358

20.

De Keyser

Hansez

. (1996). Vers une perspective transactionnelle du stress au travail: pistes d’évaluations méthodologiques [Towards a transactional perspective on workplace stress: Pathways to methodological evaluation]. Cahiers de Médecine du Travail, 33, 133-144.

21.

DeVellis

R. F.

(2006). Classical test theory. Medical Care, 44, 50-59. doi:10.1097/01.mlr.0000245426.10853.30

22.

DeVellis

R. F.

(2012). Scale development: Theory and applications (3rd ed.). Thousand Oaks, CA: Sage.

23.

De Vriendt

Clays

Moreno

L. A.

Bergman

Vicente-Rodrigues

Nagy

. . . De Henauw

. (2011). Reliability and validity of the Adolescent Stress Questionnaire in a sample of European adolescents: The HELENA study. BMC Public Health, 11, Article 717. doi:10.1186/1471-2458-11-717

24.

Edwards

B. D.

Franco-Watkins

A. M.

Cullen

K. L.

Howell

J. W.

Acuff

R. E.

Jr. (2014). Unifying the challenge-hindrance and sociocognitive models of stress. International Journal of Stress Management, 21, 162-185. doi:10.1037/a0034730

25.

Edwards

J. R.

Cooper

C. L.

(1988). The impacts of positive psychological states on physical health: A review and theoretical framework. Social Science & Medicine, 27, 1447-1459. doi:10.1016/0277-9536(88)90212-2

26.

Fan

Miller

B. C.

Park

K.-E.

Winward

B. W.

Christensen

Grotevant

H. D.

Tai

R. H.

(2006). An exploratory study about inaccuracy and invalidity in adolescent self-report surveys. Field Methods, 18, 223-244. doi:10.1177/152822X06289161

27.

Flook

(2011). Gender differences in adolescents’ daily interpersonal events and well-being. Child Development, 82, 454-461. doi:10.1111/j.1467-8624.2010.01521.x

28.

Hargrove

M. B.

Casper

W. J.

Quick

J. C.

(2014, May). Scale validation for the Self-Report Stress Response Questionnaire (SRSRQ). Paper presented at the Annual Conference of the Society of Industrial and Organizational Psychologists, Honolulu, HI.

29.

Hoaglin

D. C.

Iglewicz

(1987). Fine tuning some resistant rules for outlier labeling. Journal of the American Statistical Association, 82, 1147-1149. doi:10.2307/2289392

30.

L.-t.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55. doi:10.1080/10705519909540118

31.

John

O. P.

Naumann

L. P.

Soto

C. J.

(2008). Paradigm shift to the integrative big-five trait taxonomy: History, measurement, and conceptual issues. In John

O. P.

Robins

R. W.

Pervin

L. A.

(Eds.), Handbook of personality: Theory and research (3rd ed., pp. 114-158). New York, NY: Guilford Press.

32.

Jones

Bright

(2001). Stress: The concept. In Jones

Bright

(Eds.), Stress: Myth, theory and research (pp. 3-15). Harlow, UK: Pearson Education.

33.

Kaiser

H. F.

(1974). An index of factorial simplicity. Psychometrika, 39, 31-36. doi:10.1007/bf02291575

34.

Kern

M. L.

Benson

Steinberg

E. A.

Steinberg

(2016). The EPOCH measure of adolescent well-being. Psychological Assessment, 28, 586-597. doi:10.1037/pas0000201

35.

Laferton

J. A. C.

Stenzel

N. M.

Fischer

(2018). The Beliefs About Stress Scale (BASS): Development, reliability, and validity. International Journal of Stress Management, 25, 72-83. doi:10.1037/str0000047

36.

Lazarus

R. S.

(1993). From psychological stress to the emotions: A history of changing outlooks. Annual Review of Psychology, 441, 1-22. doi:10.1146/annurev.ps.44.020193.000245

37.

Lazarus

R. S.

Folkman

(1984). Stress, appraisal, and coping. New York, NY: Springer.

38.

Le Fevre

Kolt

G. S.

Matheny

. (2006). Eustress, distress and their interpretation in primary and secondary occupational stress management interventions: Which way first? Journal of Managerial Psychology, 21, 547-565. doi:10.1108/02683940610684391

39.

Le Fevre

Matheny

Kolt

G. S

. (2003). Eustress, distress, and interpretation in occupational stress. Journal of Managerial Psychology, 18, 726-744. doi:10.1108/02683940310502412

40.

Lovibond

S. H.

Lovibond

P. F.

(1995). Manual for the Depression Anxiety Stress Scales (2nd ed.). Sydney, New South Wales, Australia: Psychology Foundation.

41.

Luszczynska

Piko

Januszewicz

(2011). Self-efficacy. In Levesque

R. J. R.

(Ed.), Encyclopedia of adolescence. doi:10.1007/978-1-4419-1695-2

42.

Margalit

Eysenck

(1990). Prediction of coherence in adolescence: Gender differences in social skills, personality, and family climate. Journal of Research in Personality, 24, 510-521. doi:10.1016/0092-6566(90)90036-6

43.

Mason

Danby

(2011). Children as experts in their lives: Child inclusive research. Child Indicators Research, 4, 185-189. doi:10.1007/s12187-011-9108-4

44.

McCrae

R. R.

Costa

P. R.

Jr. (1997). Personality trait structure as a human universal. American Psychologist, 52, 509-516. doi:10.1037/0003-066x.52.5.509

45.

Meir Drexler

Wolf

O. T

. (2017). Stress and memory consolidation. In Axmacher

Rasch

(Eds.), Cognitive neuroscience of memory consolidation (pp. 285-300). Cham, Switzerland: Springer.

46.

Moksnes

U. K.

Espnes

G. A.

Haugan

(2014). Stress, sense of coherence and emotional symptoms in adolescents. Psychology & Health, 29, 32-49. doi:10.1080/08870446.2013.822868

47.

Moksnes

U. K.

Løhre

Lillefjell

Byrne

D. G.

Haugan

(2016). The association between school stress, life satisfaction and depressive symptoms in adolescents: Life satisfaction as a potential mediator. Social Indicators Research, 125, 339-357. doi:10.1007/s11205-014-0842-0

48.

Nederhof

A. J.

(1985). Methods of coping with social desirability bias: A review. European Journal of Social Psychology, 15, 263-280.

49.

Nelson

D. L.

Simmons

B. L.

(2003). Health psychology and work stress: A more positive approach. In Quick

J. C.

Tetrick

L. E.

(Eds.), Handbook of occupational health psychology (pp. 97-119). Washington, DC: American Psychological Association.

50.

Noor

N. M.

Alwi

(2013). Stressors and well-being in low socio-economic status Malaysian adolescents: The role of resilience resources. Asian Journal of Social Psychology, 16, 292-306. doi:10.1111/ajsp.12035

51.

O’Sullivan

(2011). The relationship between hope, eustress, self-efficacy, and life satisfaction among undergraduates. Social Indicators Research, 101, 155-172. doi:10.1007/s11205-010-9662-z

52.

Parker

K. N.

Ragsdale

J. M.

(2015). Effects of distress and eustress on changes in fatigue from waking to working. Applied Psychology: Health and Well-Being, 7, 293-315. doi:10.1111/aphw.12049

53.

Peacock

E. J.

Wong

P. T. P.

(1990). The Stress Appraisal Measure (SAM): A multidimensional approach to cognitive appraisal. Stress Medicine, 6, 227-236. doi:10.1002/smi.2460060308

54.

Podsakoff

P. M.

MacKenzie

S. B.

Lee

J.-Y.

Podsakoff

N. P.

(2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879-903. doi:10.1037/0021-9010.88.5.879

55.

Rattray

Jones

M. C.

(2005). Essential elements of questionnaire design. Issues in Clinical Nursing, 16, 234-243. doi:10.1111/j.1365-2702.2006.01573.x

56.

Reynolds

W. M.

(1982). Development of reliable and valid short forms of the Marlowe-Crowne Social Desirability Scale. Journal of Clinical Psychology, 38, 119-125. doi:10.1002/1097-4679(198201)38:1<119::AID-JCLP2270380118>3.0.CO;2-I

57.

Rice

P. L.

(1999). Stress and health (3rd ed.). Pacific Grove, CA: Brooks/Cole.

58.

Rodríguez

Kozusznik

M. W.

Peiró

J. M.

(2013). Development and validation of the Valencia Eustress-Distress Appraisal Scale. International Journal of Stress Management, 20, 279-308. doi:10.1037/a0034330

59.

Saksvik

I. B.

Hetland

(2011). The role of personality in stress perception across different vocational types. Journal of Employment Counseling, 48, 3-16. doi:10.1002/j.2161-1920.2011.tb00106.x

60.

Schreiber

J. B.

(2008). Core reporting practices in structural equation modeling. Research in Social and Administrative Pharmacy, 4, 83-97. doi:10.1016/j.sapharm.2007.04.003

61.

Schwarzer

Jerusalem

(1995). Generalized self-efficacy scale. In Weinman

Wright

Johnston

(Eds.), Measures in health psychology: A user’s portfolio (pp. 35-37). Windsor, UK: NFER-NELSON.

62.

Seligman

M. E. P.

Csikszentmihalyi

(2000). Positive psychology: An introduction. American Psychologist, 55, 5-14. doi:10.1037/0003-066X.55.1.5

63.

Selye

(1974). Stress without distress. London, England: Hodder and Stoughton.

64.

Soto

C. J.

John

O. P.

Gosling

S. D.

Jeff

(2011). Age differences in personality traits from 10 to 65: Big Five domains and facets in a large cross-sectional sample. Journal of Personality and Social Psychology, 100, 330-348. doi:10.1037/a0021717

65.

SPSS Inc. (2017). IBM SPSS statistics (Version 24). Armonk, NY: IBM Corp.

66.

Tukey

J. W.

(1977). Exploratory data analysis. Reading, MA: Addison-Wesley.

67.

Vandenberg

R. J.

Lance

C. E.

(2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3, 4-70. doi:10.1177/109442810031002

68.

Venning

Eliott

Kettler

Wilson

(2013). Complete mental health in South Australian youth: Prevalence, measurement, and promotion. In Keyes

C. L. M.

(Ed.), Mental well-being (pp. 29-50). New York, NY: Springer.

69.

Vera

E. M.

Vacek

Blackmon

Coyle

L. D.

Gomez

Jorgenson

. . . Steele

J. C.

(2012). Subjective well-being in urban, ethnically diverse adolescents: The role of stress and coping. Youth & Society, 44, 331-347. doi:10.1177/0044118X11401432

70.

Wang

Zhang

Kou

(2015). The impacts of moral evaluations and descriptive norms on children’s and adolescents’ tolerance of transgression. Journal of Pacific Rim Psychology, 9, 86-96. doi:10.1017/prp.2015.11

71.

Waters

(2011). A review of school-based positive psychology interventions. The Australian Educational and Developmental Psychologist, 28, 75-90. doi:10.1375/aedp.28.2.75

72.

Williams

S. G.

Turner-Henson

Langhinrichsen-Rohling

Azuero

(2017). Depressive symptoms in 9th graders: Stress and physiological contributors. Applied Nursing Research, 34, 24-28. doi:10.1016/j.apnr.2017.01.001

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.08 MB