Original Article

Does the Factor Structure of the

Brief COPE Fit Different Types
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

of Traumatic Events?
A Test of Measurement Invariance
Ashley Wei-Ting Wang,1 Chung-Ping Cheng,2 Cheng-Shyong Chang,3 Dar-Ren Chen,4
Shou-Tung Chen,4 Vincent Shieh,5 Angela Lo,6 and Wen-Yau Hsu1,7
1
Department of Psychology, National Chengchi University, Taipei, Taiwan
2
Department of Psychology, National Cheng Kung University, Tainan, Taiwan
3
Division of Hematology-Oncology, Department of Internal Medicine, Changhua Christian Hospital, Changhua, Taiwan
4
Comprehensive Breast Cancer Center, Changhua Christian Hospital, Changhua, Taiwan
5
Graduate Institute of Gender Education, National Kaohsiung Normal University, Kaohsiung, Taiwan
6
Department of Psychiatry, Kaohsiung Medical University, Kaohsiung, Taiwan
7
Research Center for Mind, Brain, and Learning, National Chengchi University, Taiwan

Abstract: The current study aims to explore the likelihood of there being a higher-order factorial model of the Brief COPE (Carver, 1997) that is
equally appropriate and valid for different traumatic groups (i.e., breast cancer patients and female flood survivors) at 3–4 months after the
traumatic event. Both groups of participants include 180 survivors. Five factorial models for the Brief COPE were identified when reviewing the
literature, and model comparisons were made by conducting confirmatory factor analyses (CFAs). The three dimensions of the Brief COPE
posited by Litman (2006) and Gutiérrez, Peri, Torres, Caseras, and Valdés (2007), namely self-sufficient coping, socially-supported coping, and
avoidant coping, were supported in both traumatic populations. Tests of measurement invariance generally supported invariance (factor form,
factor loadings, item intercepts, error variances, factor variances, and covariances) across groups, implying that groups perceive the latent
factors in the same way and the higher-order coping across different traumatic groups is meaningful. Moreover, the relationship between the
three higher-order coping factors and variables of psychological adjustment was examined. Generally, different relationships between the
subtypes of coping and the outcomes of adaptation existed between female flood survivors and breast cancer patients. Further theoretical and
practical implications for these relationships are discussed.

Keywords: Brief COPE, measurement invariance, breast cancer, Typhoon Morakot, psychological adjustment

The identification of adaptive and maladaptive coping strat- Turecki, 2009). As coping effectiveness depends on the
egies following traumatic life events has gained consider- characteristics of stressful events (Carver & Connor-Smith,
able attention. Researchers generally agree that coping 2010), progress in this area requires a reliable and valid
plays a crucial role in recovery and effective adaptation measurement tool of coping. Accordingly, it is doubtful
(e.g., Norris, Friedman, & Watson, 2002; Skinner, Edge, whether the factor structure of coping assessments is invari-
Altman, & Sherwood, 2003). Generally speaking, the data ant under different traumatic events.
most consistently suggest that avoidance coping is mal- Brief COPE (Carver, 1997) is one of the most widely-
adaptive (Norris, Friedman, Watson, et al., 2002; Penley, used instruments for assessing coping strategies in a variety
Tomaka, & Wiebe, 2002). The relationship between of life-threatening events and medical illnesses, such as
approach forms of coping and psychological distress, how- HIV (Turner-Cobb et al., 2002), strokes (Gillen, 2006),
ever, is less conclusive. Some studies have found approach end-stage renal disease (O’Connor, Jardine, & Millar,
coping strategies are adaptive among breast cancer patients 2008), cancer (Bellizzi & Blank, 2006; Hulbert-Williams,
(Carver et al., 1993; O’Brien & Moorey, 2010), the others Neal, Morrison, Hood, & Wilkinson, 2012), and hurricanes
have found them to be positively related or unrelated to dis- (Glass et al., 2009). The Brief COPE was developed based
tress in natural disaster (Glass, Flory, Hankin, Kloos, & on theoretical models and is the shorter version of the

European Journal of Psychological Assessment (2018), 34(3), 162–173 Ó 2016 Hogrefe Publishing
DOI: 10.1027/1015-5759/a000321
 A. W.-T. Wang et al., Measurement Invariance and Brief Cope 163

COPE (Carver, Scheier, & Weintraub, 1989). Compared to be identified. The 3-factor structure has two substantial
other long-established coping assessment tools, such as the conceptualizations, the first of which conceptualizes coping
Revised Ways of Coping Questionnaire (WOC-R; Folkman strategies as being problem-focused (AC, PL, IS), emotion-
& Lazarus, 1985) and the Coping Strategies Inventory (CSI; focused (ES, PS, HU, AT, RE), or avoidant coping (Gow,
Tobin, Holroyd, Reynolds, & Wigal, 1989), its brevity gives Warren, Anthony, & Hinschen, 2008; Ng & Keung Leung,
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

it the advantage of being applied in time-limited situations. 2006; O’Connor et al., 2008). This is based on the theoret-
Therefore, the Brief COPE is an important measurement ical framework of Folkman and Lazarus (1980) and Carver
tool when discovering how coping leads to successful and colleagues. In this categorization, venting was inconsis-
traumatic adaptation. tently conceptualized as either avoidant coping (Model 2a;
The main impediment in coping research has shown to Coolidge, Segal, Hook, & Stewart, 2000) or emotion-
be that there is little agreement among theorists about focused coping (Model 2b; Cartwright, Endean, & Porter,
the way to conceptualize categories of coping strategies 2009; Schnider, Elhai, & Gray, 2007). However, the prob-
(Skinner et al., 2003). Therefore, constructing category sys- lem- and emotion-focused distinction, albeit theory-driven,
tems to conceptualize and measure coping have been cen- has not received empirical support. Previous research has
tral endeavors (Carver & Connor-Smith, 2010; Skinner & consistently found them loaded on the same factor (Litman,
Zimmer-Gembeck, 2007). Skinner and colleagues have 2006; Skinner et al., 2003). Therefore, Litman (2006) and
pointed out the central importance of achieving a consen- Gutiérrez et al. (2007) postulated another 3-factor struc-
sus on categorizing coping into a few pivotal higher-order ture, in which coping strategies were identified as being
dimensions (i.e., families of coping strategies), which would self-sufficient coping (AC, PL, PR, HU, AT), socially-
help to advance coping research and better explain the rela- supported (ES, IS, RE), and avoidant coping. This conceptu-
tionship between coping and adjustment. Thus, for a coping alization was subsequently supported by several studies
measure, the critical issue is whether or not a set of reliable using an exploratory factor analysis (EFA; Feaster &
higher-order coping categories can be identified for trau- Szapocznik, 2002; Moscardino, Scrimin, Capello, Altoè, &
matic events and whether or not they convey the same Axia, 2008; Prado et al., 2004; Schottenbauer et al.,
meaning across different traumatic events. 2006; Snell, Siegert, Hay-Smith, & Surgenor, 2011).
The Brief COPE measures 14 coping strategies (with two Likewise, venting was inconsistently labeled as being either
items each): self-distraction (SD), active coping (AC), denial avoidant coping (Model 3a) or socially-supported coping
(DE), substance use (SU), use of emotional support (ES), (Model 3b; Litman, 2006). Although the 2- and 3-factor
use of instrumental support (IS), behavioral disengagement structures were theory-driven, no study to date has com-
(BD), venting (VE), positive reframing (PR), planning (PL), pared these models on the Brief COPE. To uncover which
humor (HU), acceptance (AT), religion (RE), and self-blame of the aforementioned latent structures fits the coping strat-
(SB). Based on Skinner and colleagues (2003)’s argument, egies of traumatic survivors, this study focuses on two
these lower-order coping strategies can be classified into groups of survivors facing different types of traumatic
higher-order dimensions to represent the core constructs events: breast cancer (personal trauma) and flood (collec-
of coping. tive trauma). The first aim of this study is to examine the
However, a higher-order factorial structure of the Brief factor structure of the Brief COPE by using a confirmatory
COPE in traumatic or stressful events has lacked consen- factor analysis (CFA).
sus. According to the literature, the primary categorizations Psychological traumatic events can be classified as either
of the 14 coping strategies in the Brief COPE are 2- and collective trauma (including natural disasters and man-
3-factor structures. The 2-factor structure classifies coping made disasters) or individual trauma (including life-
strategies as being either adaptive or maladaptive (Bellizzi threatening medical illnesses, such as cancer). The current
& Blank, 2006; Hulbert-Williams et al., 2012; Krägeloh, study focuses on coping at 3–4 months after two types of
2011), based on the theoretical model of Carver and life-threatening events, namely, breast cancer and Typhoon
colleagues (1989). The subscales that form adaptive coping Morakot. Breast cancer is the most frequent cancer in
are AC, PL, AT, RE, HU, RE, ES, and IS. The remaining women worldwide and one of the most overwhelmingly tra-
subscales form maladaptive coping (Model 1). Although gic diagnoses a women can be given (Jemal et al., 2011).
adaptive and maladaptive is a basic coping taxonomy Taiwan was affected by a number of catastrophic natural
(Carver & Connor-Smith, 2010), and several studies disasters within 15 years, with the Chi-Chi earthquake in
adopted this classification for simplicity, some reviews have 1999 and Typhoon Morakot in 2009 being two major nat-
highlighted that this categorization oversimplifies the func- ural disasters in the contemporary era. Typhoon Morakot
tion of coping (Coyne & Racioppo, 2000; Skinner et al., caused the “8/8 Flood Disaster,” which was the costliest
2003). Accordingly, researchers have suggested that two and deadliest flood in 50 years in Taiwan. An entire village
approach-oriented and one avoidance-oriented factors can was completely buried, and 724 people died, 24,775

Ó 2016 Hogrefe Publishing European Journal of Psychological Assessment (2018), 34(3), 162–173
 164 A. W.-T. Wang et al., Measurement Invariance and Brief Cope

civilians were evacuated from their hometowns. After the was conducted in November to December, 2009, 3–4
higher-order structure of the Brief COPE is confirmed, we months after the disaster struck. Considering that gender
then ask if this measurement model is invariant across dif- may be a critical characteristic, we only included female
ferent traumatic survivors. survivors (n = 180) in this sample because the sample of
Therefore, the first goal of this study is to focus on the breast cancer patients only consisted of females. Mean
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

way to conceptualize the Brief COPE into higher-order cat- age was 40.31 years (SD = 14.91, range = 17–80). Among
egories and examine its level of generalizability over differ- this sample, 25% had experienced a friend or neighbor
ent traumatic events, which is usually referred to as being injured, 22% had a family member injured, 21% were
measurement invariance (MI). MI denotes that the factorial injured themselves, 11.3% had a family member or neigh-
model is not perceived differently across populations bor who died in the flood disaster, 33.4% of the sample
(Steenkamp & Baumgartner, 1998) and this is important had their houses destroyed or seriously damaged, and
in accumulating the findings of Brief COPE research 45.8% of the sample had been temporarily relocated to safe
related to diverse traumatic events. In order to expand areas. The vast majority of the participants (94.8%)
the use of the brief COPE and make comparisons across regarded this flood to have been extremely destructive,
different traumatic groups meaningful, we tested levels of and 70% felt that the flood had significantly threatened
invariance to explore to what extent people under different their lives.
traumatic experience perceive the Brief COPE in the same The second sample was recruited from the Breast
way. Several reviews have highlighted a lack of reliable Surgery of a Hospital in Taiwan from 2010 to 2011. Women
structure for coping assessment (Coyne & Racioppo, were eligible if they (1) were diagnosed with breast cancer
2000; Skinner et al., 2003). To explore a structure that at any stage but did not have a history of any other cancer,
can be replicated across different traumatic events, MI test (2) were aged 18–75 years, (3) were aware of the diagnosis
is a critical step (Steenkamp & Baumgartner, 1998). To of breast cancer, (4) had no prior psychiatric history, and (5)
date, no research has applied MI on coping measures. comprehended Mandarin. Of the 201 women who were
Having established the comparability of the measure- contacted, 180 consented to participate. Part of the sample
ment model, the second goal is to test whether these was reported in Wang, Chang, Chen, Chen, and Hsu
higher-order coping factors are meaningfully distinct by (2014). The mean age was 49.33 years (SD = 8.11, range =
examining the relationship between coping and the positive 30–72). They were approached 3–4 months after undergo-
and negative outcomes of adaptation upon controlling for ing cancer surgery. The participants were diagnosed with
demographic variables and trauma severity. Stage 0 (19.4%), Stage I (30.0%), Stage II (32.2%), Stage
III (17.8%), and Stage IV (0.6%)1. About half (55.3%) of
the women had lumpectomies rather than mastectomies
(44.7%). The majority of patients (81.67%) had received
Method chemotherapy, radiation therapy, or both. The demo-
graphic variables of the two samples are provided in Table 1.
Participants and Procedures The breast cancer sample was older and had a higher edu-
cational level and a better income than the flood survivors.
The data used here came from two samples: a sample of In both samples, participants were invited to complete
Typhoon Morakot female survivors and a sample of paper-and-pencil measures including the Brief COPE and
newly-diagnosed breast cancer patients. Written informed assessments of psychological adjustment. The samples used
consent was obtained from all the participants prior to the in the current study came from different studies, in which
beginning of the study, and the National Science Council different outcome measures were administered. However,
of Taiwan (for the first sample) and the Institutional Review both the positive and negative outcomes of the two samples
Board (for the second sample) approved the protocol. were assessed.
For the first sample (260 community residents), this
study uses data from a large-scale post-flood study, which
took place in three major areas (Namaxia, Taoyuan, and
Measures
Maolin in Kaohsiung County) that had been seriously
affected by the typhoon and had the highest rates of The Brief COPE (Carver, 1997) was administered to both
flood-related mortality, morbidity, and damage. The study samples. Participants were asked to respond to the items

1
“Stage” contains the information about the prognosis or severity of the breast cancer diagnosis. According to American Cancer Society (2014),
5-year survival rate for Stage 0 or I is 100% and for Stage II it is 93%. However, if the cancer has spread regionally (Stage III), the survival rate is
72%, and it drops to 22% if the cancer has metastasized (Stage IV). In the current study, our sample consisted of 81.6% early-stage breast
cancer patients.

European Journal of Psychological Assessment (2018), 34(3), 162–173 Ó 2016 Hogrefe Publishing
 A. W.-T. Wang et al., Measurement Invariance and Brief Cope 165

Table 1. Sample characteristics of flood survivors (N = 180) and breast 0 to 100, with higher scores indicating better health. We
cancer patients (N = 180)
use the Chinese version of the SF-36 (Lu, Tseng, & Tsai,
Characteristic Flood survivors (%) Breast cancer patients (%) 2003). Coefficient alpha reliability was .83 and .88 for
M = 40.31 M = 49.33 PCS and MCS, respectively.
(SD = 14.91) (SD = 8.11)
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

Age (years) Range 17–80 Range 30–72

Psychological Adjustment Measures for the Flood
Education
Elementary school 31.6 17.2 Survivor Sample
Junior high school 21.6 16.7 The 17-item Posttraumatic Diagnostic Scale (PDS; Foa,
Senior high school 40.3 41.7 Cashman, Jaycox, & Perry, 1997) was used to assess PTSD
College and above 13.4 48.0
symptoms. Using a 4-point scale (scored 0–3), the respon-
Marital status 73.3 87.8
Employed 50.5 49.71 dents were asked to rate 17 items (e.g., “Have you had
Family income (dollars/month) recurrent or intrusive distressing thoughts and recollections
< 1,000 67.5 30.0 about Typhoon Morakot?”) representing the DSM-IV PTSD
1,000–2,000 14.5 31.7
symptoms experienced in the past 30 days: Reexperiencing,
2,000–3,000 10.8 17.2
> 3,000 7.2 13.4 Avoidance, and Arousal. A higher score indicated more
severe PTSD symptomatology. The test-retest (3-week
interval) coefficient was r = .83 (Su & Chen, 2008). Coeffi-
to reflect the way in which they dealt with their cancer diag- cient alpha reliability was .93.
nosis and the 8/8 Flood on a 4-point Likert scale (score 1–4). Depression was assessed using the 10-item short form of
Having obtained permission to use the Brief COPE from the the Center for Epidemiological Studies Depression Scale
author, following the guidelines of translation and back- (10-item CES-D). Each of the 10 items was rated on a
translation, we translated it into Mandarin Chinese. 4-point scale (scored 0–3), indicating the frequency of
experiencing each symptom in the past week. The 10-item
Psychological Adjustment Measures for the Breast CES-D statements include: Not interested in eating and
Cancer Sample have a poor appetite. The responses were reversely scored
The Hospital Anxiety and Depression Scale (HADS; when necessary so that higher scores represented greater
Zigmond & Snaith, 1983) consists of 14 items, divided into levels of symptom frequency. The 10-item CES-D was
two 7-item subscales: Anxiety (e.g., I feel tense or “wound revised from Radloff (1977)’s 20-item CES-D and translated
up”) and Depression (e.g., I have lost interest in my appear- into Chinese by the Bureau of Health Promotion of the
ance). The respondent rates each item on a 4-point scale Taiwan Department of Health in a nationally representative
ranging from 0 (= absence) to 3 (= extreme presence). It is a survey of the Taiwan Longitudinal Study on Aging (TLSA)
tool created for the detection of anxiety and depression in from 1989 to 2007, and has been validated to show good
people with physical health problems, including cancer. internal consistency reliability (Cronbach’s α = 0.79–.97;
The items do not assess somatic complaints. The Chinese Lee, Ou, Chen, & Weng, 2009). Coefficient alpha reliability
version has shown adequate reliability and validity (Wang, was .90 for the current study.
Tu, Liu, Yeh, & Hsu, 2013). Coefficient alpha reliability The Life Satisfaction Index (LSI; Neugarten, Havighurst,
was .81 and .76 for anxiety and depression, respectively. & Tobin, 1961) contains 10 items to assess life satisfaction.
The Affects Balance Scale (ABS; Bradburn, 1969) mea- These items were dichotomously rated yes or no and higher
sures two general dimensions of emotional experiences, LSI scores corresponded with better life satisfaction. The
PA (positive affect) and NA (negative affect), each of which LSI statements include: “Has your life been better than
comprises five specific mood-related adjectives, rated yes- most people’s lives?” “Are you satisfied with your life?”,
no, in the past week. The PA statements include: Did you and “Do you find what you do interesting?” The LSI was
feel particularly excited or interested in something? And also translated and adopted in the survey of the TLSA,
the NA statements include: Did you feel so restless that and its validity and reliability were demonstrated (Lee
you couldn’t sit long in a chair? Cronbach’s alpha was .73 et al., 2009). Coefficient alpha reliability was .56.
and .63 for PA and NA, respectively.
The SF-36 Health Survey (SF-36; Ware, Snow, Kosinski, Exposure Level of the Flood Survivor Sample
& Gandek, 1993) measures both the physical (PCS) and The Earthquake Exposure Index (EEI; S. H. Chen, Hung,
psychosocial (MCS) aspects of health-related quality of life. Lin, & Tseng, 2002) is a 6-item assessment measuring
The SF-36 contains eight subscales, four of which cover the trauma severity in the flood survivor sample. The EEI con-
physical quality of life (PCS) and four the mental and social sists of index for death and injury (including physical injury
quality of life (MCS). Since norm-based scoring for SF-36 is level in self, injury, and/or death of family member(s),
used, scores on the total PCS and MCS scores range from relatives, and close friends) and index for property loss

Ó 2016 Hogrefe Publishing European Journal of Psychological Assessment (2018), 34(3), 162–173
 166 A. W.-T. Wang et al., Measurement Invariance and Brief Cope

(i.e., degree of house damage). Its validity and reliability conceptualized in the same way across all groups
were demonstrated in studies of Chi-Chi earthquake (Steenkamp & Baumgartner, 1998). Scalar (item intercepts)
(S. H. Chen et al., 2002). Coefficient alpha reliability in and strict (residual errors) invariance suggests that the
the current study was .52. internal anchors for the subscales are equal across groups,
which is generally a requirement for comparing group
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

means (Cheung & Rensvold, 2002). Finally, the factor var-

Statistical Analysis iance and covariance across groups signify that the concep-
Five CFA models were estimated and compared to assess tualization and prioritization are the same for the factor
the factor validity of the Brief COPE using Lisrel 8.7 meaning at a higher level.
(Jöreskog & Sörbom, 2004). CFAs were conducted on the Invariance testing is a sequential test, which involves
samples of flood survivors and breast cancer survivors sep- comparing a more restricted model and a baseline model.
arately. All models (Models 1, 2a, 2b, 3a, and 3b) were That is, restrictions are added while holding the previous
tested independently in both groups. If the best-fitting restrictions in successive models. Differences of several
model of the two groups was found to be different, this goodness-of-fit indices for nested models are proposed to
would be interpreted as a lack of configural invariance test measurement invariance, with less than .01 change in
and MI will not be tested. CFI (ΔCFI) (Cheung & Rensvold, 2002) and less than
Since the scale for “Substance Use” showed a very low .015 change in RMSEA (ΔRMSEA) (F. F. Chen, 2007). In
variance in the breast cancer sample (SD = 0.25), it was testing measurement invariance, the chi-square difference
not included in all the analyses. Thirteen subscales of the statistic (Δw2) was used in previous studies. However, it
Brief COPE were designated as observed variables. The has been suggested that ΔCFI and ΔRMSEA are better indi-
robust maximum likelihood method was employed to esti- cators than Δw2, and are sensitive to lack of invariance (F. F.
mate all models because initial assumption checks revealed Chen, 2007; Cheung & Rensvold, 2002). In this study, the
that none of the Brief COPE subscales were normally dis- invariance at each step was evaluated by the ΔCFI and
tributed (Shapiro-Wilks p < .0005). Missing values were ΔRMSEA, as well as Δw2. If the full invariance hypothesis
imputed by mean (average of the observed values). was rejected, a partial measurement invariance model
Several goodness-of-fit measures were used to evaluate was then estimated. Modification index (MI) was used to
the models. These included acceptable cutoffs for the determine which items were invariant.
Bentler-Bonett Normed Fit Index (NFI), the Non-Normed To address the second research question, the relation-
Fit Index (NNFI), and the Comparative Fit Index (CFI) ships between coping strategies and psychological adjust-
which are .95 or higher, and .08 or lower for the Standard- ment were examined by hierarchical multiple regression.
ized Root-Mean-Square Residual (SRMR) and the Root- We statistically controlled demographic variables (age, edu-
Mean-Square Error of Approximation (RMSEA; Hu & cation, marital status, and income) and trauma severity
Bentler, 1999). Difference of CFI (ΔCFI) more than .01 (disaster exposure level or cancer stage). The entering steps
would be interpreted as significantly different. The model of the independent variables were as follows: demographic
with higher CFI value is better than that with lower CFI variables and trauma severity were assessed first, and then
value (Cheung & Rensvold, 2002). Furthermore, the index all coping strategies were entered together in Step 2. To
for model comparison, the Akaike information criterion avoid overestimation of what can be predicted in the
(AIC; Akaike, 1987), was used; this model gives the lowest regression models, insignificant predictors will be elimi-
AIC value and is considered to be the best fit. nated from the regression equation before conducting the
Having achieved the best-fitting model, a multigroup final analyses.
CFA was then used to test for measurement invariance in
the two samples. When the factor structure held for the
two groups, the next step was to test if the factor loadings,
intercepts, residual variances, and factor variance and Results
covariance were also equivalent. Equality restrictions on
the parameters of the hypothesized factor model can be Confirmatory Factor Analyses and Tests
applied in a series of nested CFA models, in order to test
of Measurement Invariance
the different levels of equality (Vandenberg & Lance,
2000). All these levels of invariance have a significant As shown in Table 2, Model 3b significantly outperformed
implication for the use and the psychometric properties of the other four models in both samples and produced an
the Brief COPE. When configural (factor structure) acceptable fit to the data of both (Akaike, 1987; Cheung
invariance and metric (factor loadings) invariance are & Rensvold, 2002; Hu & Bentler, 1999). All the fitted path
met, it means that the higher-order coping factors can be estimates of Model 3b were significant (see Figure 1).

European Journal of Psychological Assessment (2018), 34(3), 162–173 Ó 2016 Hogrefe Publishing
 A. W.-T. Wang et al., Measurement Invariance and Brief Cope 167

Table 2. Summary of test statistics for confirmatory factor analyses

w2 df w2/df NFI NNFI CFI SRMR RMSEA AIC
Flood survivors
Model 1 215.82 64 3.37 0.87 0.89 0.91 0.093 0.115 269.82
Model 2a 212.68 62 3.48 0.87 0.88 0.91 0.092 0.117 270.68
Model 2b 180.29 62 2.91 0.89 0.91 0.93 0.086 0.103 238.29
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

Model 3a 155.63 62 2.51 0.91 0.93 0.94 0.085 0.092 213.63

Model 3b 111.76 62 1.80 0.93 0.96 0.97 0.073 0.067 169.76
Breast cancer patients
Model 1 211.00 64 3.30 0.90 0.91 0.93 0.097 0.113 265.00
Model 2a 210.13 62 3.39 0.90 0.91 0.93 0.099 0.116 268.13
Model 2b 246.11 62 3.97 0.88 0.89 0.91 0.095 0.129 304.11
Model 3a 94.95 62 1.53 0.96 0.98 0.98 0.067 0.054 152.95
Model 3b 91.80 62 1.48 0.96 0.98 0.99 0.067 0.052 149.80
Notes. NFI = Bentler-Bonett Normed Fit Index; NNFI = Non-Norm Fit Index; CFI = Comparative Fit Index; SRMR = Standardized Root-Mean-Square Residual;
RMSEA = Root-Mean-Square Error of Approximation; AIC = Akaike Information Criterion. These data were obtained by 13 subscales of the Brief COPE
because “Substance Use” was excluded.

Figure 1. The final model (Model 3b) of

.40/.31 Active Coping the Brief COPE. The results of confirma-
.77 tory factor analyses. The first value
/.8
.42/.25 Positive Reframing 3 (bold) represents the flood-survivor
.76
/.8 sample and the values below slash
6 (italic) represents the breast-cancer
.38/.30 Planning .79/.8
4 sample.
Self-
.52/.48 .69/.72 sufficient
Humor
9
.77/.7
Coping
.41/.37 Acceptance

.76/.66
.45/.28 Emotional Support .74/
.85
.37/.35 Instrumental Support .79/.81 Socially-
Supported .26/.54
.70/.64 Coping
.50/.59 Venting
/.57
.41
.83/.68 Religion
.56/.75

.69/.52 Self-Distraction .56/.69

Avoidant
.41/.57 Coping
.84/.68 Denial
.26
.61/
2
.63/.93 Behavior Disengage /.4
.54

.71/.82 Self-Blame

The sequential testing procedure of MI and statistical significant. In accordance with Δw2 criterion, we tested
results are presented in Table 3. The model for configural the partial invariance to examine whether invariance may
invariance fits the data very well, as indicated by a RMSEA hold for a subset of items. Our examination of MIs revealed
value of .059 and CFI = .98. Examining differences in the that only the loading of Behavioral Disengagement varied
fit statistics between the metric invariance and configural between samples. We then constrained all factor loadings
invariance models, the metric invariance model was sup- except that of Behavioral Disengagement (partial metric
ported by the ΔCFI being smaller than .01 and the ΔRMSEA invariance model). The fit indices showed that there was
being smaller than .015. However, the Δw2 value was a nonsignificant Δw2 between the partial metric invariance

Ó 2016 Hogrefe Publishing European Journal of Psychological Assessment (2018), 34(3), 162–173
 168 A. W.-T. Wang et al., Measurement Invariance and Brief Cope

Table 3. Tests for measurement invariance of the Brief COPE

Comparison
Invariance model w2 df Δw2 Δdf NFI CFI ΔCFI RMSEA Δ RMSEA AIC models
M1: Configural (form) 202.137 124 0.947 0.979 0.0593 318.137
M2: Metric (loadings) 226.080 134 23.943* 10 0.941 0.975 0.004 0.0620 0.0027 322.080 M1
M2-1: Metric (loadings) with all loadings 213.713 133 11.576 9 0.944 0.978 0.001 0.0582 0.0011 311.713 M1
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

constrained but behavior disengagement

M3: Scalar (intercepts) 216.757 146 3.044 13 0.943 0.981 0.003 0.0520 0.0062 340.757 M2-1
M4: Strict (residuals) 242.709 159 25.952* 13 0.936 0.977 0.004 0.0542 0.0022 340.709 M3
M4-1: Strict (residuals) with all residuals 230.614 158 13.857 12 0.939 0.980 0.002 0.0507 0.0035 330.614 M3
constrained but venting
M5: Factor variance and covariance 251.467 164 20.853* 6 0.934 0.976 0.004 0.0546 0.0039 339.467 M4-1
Notes. NFI = Bentler-Bonett Normed Fit Index; CFI = Comparative Fit Index; RMSEA = Root-Mean-Square Error of Approximation; AIC = Akaike Information
Criterion; M = Model. These data were obtained by 13 subscales of the Brief COPE because “Substance Use” was excluded. *p < .05.

model and the configural model. Item intercepts were Table 4. Means, Standard Deviations, and Cronbach’s alpha coeffi-
cients for coping factors and psychological adjustment
then constrained; the scalar invariance model was sup-
variables
ported by nonsignificant Δw2, the ΔCFI being smaller
Breast cancer
than .01 and the ΔRMSEA being smaller than .015,
Flood survivors patients
between the scalar invariance model and the partial met-
M SD α M SD α
ric invariance model. We further constrained error resid-
uals. According to the ΔCFI and ΔRMSEA, the strict Self-sufficient Coping 31.31 5.46 0.89 30.21 6.85 0.91
Socially-supported Coping 23.87 4.62 0.81 21.35 5.60 0.86
invariance model was acceptable. But the Δw2 value was Avoidant Coping 17.68 3.79 0.69 16.09 4.09 0.69
significant. MIs showed that there were significant differ- PTSD 10.23 9.58 0.93
ences across the two samples on Venting error variances. Depression (10-item CED-S) 9.06 6.41 0.90
We constrained all error variances except that of Venting Life Satisfaction 6.44 1.86 0.56
PCS 50.00 8.10 0.83
(partial strict invariance model). The fit indices showed MCS 50.00 9.18 0.88
that there was a nonsignificant Δw2 between the partial PA 2.29 1.64 0.73
strict invariance model and the scalar model. Last, factor NA 0.99 1.26 0.63
variance and covariance invariance model was tested. Anxiety (HADS) 5.90 3.75 0.81
Depression (HADS) 5.07 4.10 0.76
These constraints did not significantly degrade the model
fit according to the ΔCFI and ΔRMSEA, whereas the Δw2 Notes. PTSD = symptoms of posttraumatic stress disorder; PCS = physical
quality of life; MCS = mental and social quality of life; PA = positive affect;
value was significant. NA = negative affect; HADS = The Hospital Anxiety and Depression Scale;
In summary, when factor loadings, item intercepts, error 10-item CES-D = the 10-item short form of the Center for Epidemiological
Studies Depression Scale.
residuals, and factor variance and covariance were con-
strained in consecutive tests, these constraints did not sig-
nificantly degrade the model fit according to the ΔCFI Hierarchical Regression Analyses
and ΔRMSEA. The goodness-of-fit of the final model was Predicting Positive and Negative
satisfactory, providing reasonable support that the measure-
Psychological Adjustment
ment model of the Brief COPE was invariant for the two
samples. According to the Δw2 value, there were slight dif- The second research questions concerned whether three
ferences in the final model. Behavioral Disengagement coping strategies showed differential relationships to psy-
had a larger factor loading on its corresponding subscale chological adjustment and whether different associations
“Avoidant Coping” in the flood sample than in the breast existed in different traumatic samples. Table 5 presents
cancer sample. In addition, Venting had a larger error resid- the results of hierarchical regression analyses. Only signifi-
ual in the breast cancer sample than in the flood sample. cant predictors were retained in the final models.
Table 4 presents the means, standard deviations, and In the flood survivor sample, results indicated that avoid-
Cronbach’s alpha coefficients for the higher-order coping ant coping was positively related to PTSD symptoms and
factors and the indices for psychological adjustment. In depression and negatively associated with life satisfaction.
terms of the coping strategies used, the flood survivors Self-sufficient coping was associated with less PTSD and
endorsed more socially-supported and avoidant coping than less depression. Socially-supported coping was positively
the breast cancer patients (t = 4.66, 3.81; p < .001). associated with life satisfaction.
Cronbach’s alpha coefficients were all acceptable for three In the breast cancer patient sample, self-sufficient coping
coping strategies in these two samples. was positively related to positive effect (PA) and physical

European Journal of Psychological Assessment (2018), 34(3), 162–173 Ó 2016 Hogrefe Publishing
 A. W.-T. Wang et al., Measurement Invariance and Brief Cope 169

Table 5. The results of the hierarchical regression analyses

Final β ΔR2 Final β ΔR2 Final β ΔR2

Flood survivors Predicting PTSD Predicting depression Predicting life satisfaction

Step 1
Age .32*** .22** .14*** .02*
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

Education .26***
Marital status
Income .14*
Trauma severity .44*** .24***
Step 2
Self_coping .20** .08*** .19** .14*** .08***
Social_coping .24**
Avoidant_coping .23*** .34*** .24**

Breast cancer patients Predicting anxiety Predicting depression Predicting PA

Step 1
Age .10*** .12***
Education .14* .21**
Marital status
Income
Cancer stage
Step 2
Self_coping .28*** .27*** .55*** .26*** .36*** .11***
Social_coping
Avoidant_coping .54*** .34***

Breast cancer patients Predicting NA Predicting PCS Predicting MCS

Step 1
Age .06*** .10*** .01
Education .20**
Marital status
Income .17*
Cancer stage .30***
Step 2
Self_coping .23** .14*** .18* .04* .38*** .24***
Social_coping
Avoidant_coping .38*** .17* .47***
Notes. Self_coping = self-sufficient coping; Social_coping = socially-supported coping; PCS = physical quality of life; MCS = mental and social quality of life;
PA = positive affect; NA = negative affect. *p < .05. **p < .01. ***p < .001.

and mental quality of life (PCS and MCS), and negatively This taxonomy can be conceptualized under theoretical
associated with negative affect (NA), anxiety, and depres- and cultural framework. Evolutionary perspective, intro-
sion. Avoidant coping was associated with increased NA, duced by Gutiérrez and colleagues as a way of explaining
anxiety, and depression, and low PCS and MCS. Socially- this factor structure, conceives coping strategies as evolved
supported coping was not associated with any indicator of mechanisms and consequences of natural selection to pro-
psychological outcomes in the breast cancer patient sample. tect the organism against outside threats and adaptive to
the circumstances. The first dimension of the Brief COPE,
self-sufficient coping, embraces problem- and emotion-
focused strategies, echoing the sentiment that these two
Discussion types of coping are greatly covaried (Lazarus, 1996; Skinner
et al., 2003). According to the evolutionary view of coping,
The current study used MI to explore whether or not there problem-focused coping involves maneuvers associated
is a higher-order factorial model of the Brief COPE that is with executive functions and thus demands emotion-
equally appropriate and valid for different traumatic groups focused strategies to be jointly mobilized to reduce the feel-
at 3–4 months after the traumatic events. Our results sup- ings of threat and physical arousal that hamper effective
ported three dimensions of the Brief COPE posited by problem-solving. Therefore, these two coping strategies
Litman (2006) and Gutiérrez et al. (2007), namely, self- work simultaneously and form the factor “self-sufficient
sufficient coping, socially-supported coping, and avoidant coping.” The second dimension, socially-supported coping,
coping. is associated with strategies oriented to the social

Ó 2016 Hogrefe Publishing European Journal of Psychological Assessment (2018), 34(3), 162–173
 170 A. W.-T. Wang et al., Measurement Invariance and Brief Cope

environment. The help-seeking strategies are evolved social residual error of Venting, and the model fell to meet factor
abilities eliciting care, connectedness, and protection from variance and covariance invariance. The contribution of
group. Achieving support is potentially adaptive because Behavioral Disengagement to the “avoidant” factor was
social support provides sympathy or tangible help. In this higher in the flood survivor sample as compared to the
dimension, religion and venting both involve an orientation breast cancer patient sample. We may infer that Behavioral
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

toward others (reciprocity) and the surroundings (turning to Disengagement would be a less important aspect of avoid-
others), at least in the short-term aftermath of trauma. ant coping for the breast cancer patients than for the flood
Therefore, they can be classified as socially-supported cop- survivors. Fell to meet residual invariance in Venting indi-
ing. The third dimension, avoidant coping, is among the cated that is attributable to differential error variances
most consistent factors that have achieved consensus in (unique variances) of Venting. Sometimes it represents
most studies (e.g., Cartwright et al., 2009; Schnider et al., unmodeled sources of systematic variance that operate dif-
2007). This dimension encompasses disengagement-based ferently in the two groups. We may infer that while 3–4
strategies that withdraw efforts from the threat. Thus, it has months after the flood is the period of recovery, 3–4 months
been regarded as maladaptive, especially if these strategies after cancer surgery is a fighting period, during which
are used frequently and intensely (Gutiérrez et al., 2007). women are fighting with adjunctive treatments and dis-
From cultural perspective, the independent-interdependent charging of negative emotions to other people could signify
self-construal framework describes that Chinese people are helplessness.
more relational and interdependent oriented, compared In the hierarchical regression analyses, avoidant coping
with people in Western culture (Markus & Kitayama, was not only consistently associated with negative psycho-
1991). Since maintaining coalitions and soliciting aid is pre- logical outcomes but related to decrease in life satisfaction
valent and accessible in Chinese culture, the independent- among flood survivors, signifying that avoidant coping is a
interdependent personal construal system may affect many maladaptive coping strategy. Past research has shown that
aspects of individual experiences, including coping. Thus, avoidant coping had the greatest effect on the relationship
the approach-oriented coping can be broadly separated into with psychological distress in the aftermath of natural disas-
two dimensions, an individual-intrapersonal dimension ter (Norris, Friedman, Watson, et al., 2002; Penley et al.,
(self-sufficient coping) and a relational-interpersonal 2002). On the other hand, self-sufficient coping only asso-
dimension (socially-supported coping). ciated with low PTSD and depression, and socially-
In addition to cross-validate the factor structure in two supported coping only associated with higher life
samples, this study tested measurement invariance to pro- satisfaction.
vide further psychometric information on the Brief COPE. The uncontrollability of natural disaster may have con-
The results of the increasingly stringent invariance tests tributed to the effectiveness of coping strategies. The flood
indicated that the factorial model of the Brief COPE was causes widespread destruction and dramatic environmental
invariant across traumatic groups. According to two com- change, especially in the initial stage of reconstruction, the
monly used fit indices for invariance models, ΔCFI and damage can hardly be avoided, no matter what one does.
ΔRMSEA (F. F. Chen, 2007; Cheung & Rensvold, 2002), Therefore, approach-based coping can hardly perform a
the results of this study showed that there was no signifi- beneficial function at this stage. Furthermore, it may also
cant departure from measurement invariance in terms of be possible that approach-based coping involves a substan-
factor form, factor loadings, item intercepts, and error vari- tial investment in energy, which may offset the adaptive
ances, and provided a prerequisite for a comparison of function at this acute stage (Glass et al., 2009; Wu, Hung,
means and correlations. Although we did not attempt to & Chen, 2002). Thus, our results revealed that avoidant
further explore these differences in this study, this is an coping contributed more variance to psychological adjust-
interesting avenue for future research, extending it to ment following an uncontrollable natural disaster like flood.
including the relationship to other constructs (Steenkamp It is worth noting that, even in the overwhelming circum-
& Baumgartner, 1998). The results also showed that vari- stances, self-sufficient coping was related to fewer PTSD
ances and covariances of three factors are invariant across symptoms in the three months following a flood. It implies
two groups in the study, implicating that the female flood that self-sufficient coping is a protective strategy at least in
survivors and breast cancer patients shared a common the short term after the flood disaster.
understanding of higher-order coping categories as mea- In terms of the breast cancer patients, self-sufficient cop-
sured by the Brief COPE. Using a stringent criterion – ing was consistently related to low psychological distress
chi-square difference test, results supported that the and to high life quality and PA, indicating it is an adaptive
3-factor structure was relatively invariant across these dif- and engaged-based coping style (Litman, 2006). Avoidant
ferent samples with a slight invariance of the factor loading coping was associated with more distress and low physical
of Behavioral Disengagement, a slight invariance of the and psychosocial aspects of quality of life, but was

European Journal of Psychological Assessment (2018), 34(3), 162–173 Ó 2016 Hogrefe Publishing
 A. W.-T. Wang et al., Measurement Invariance and Brief Cope 171

unrelated to PA. These results support that self-sufficient and treatment plans. Therefore, self-sufficient coping has
coping is adaptive when tackling cancer stress (e.g., been more critical at this phase. Penley and colleagues’
Stanton, Danoff-burg, & Huggins, 2002), but disengage- (2002) meta-analysis has supported that the effectiveness
ment coping relates to increased psychological symptoms. of socially-supported coping depends on stressor type and
Our findings also suggest that the magnitude and direction duration.
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

of the association of coping strategies with psychological To our knowledge, this is the first study that addresses
outcomes may depend on the type of traumatic event. the invariance of the Brief COPE across traumatic events.
Self-sufficient coping is more important in explaining While the invariance of the 3-factor higher-order structure
adjustment outcomes for breast cancer survivors than for has an important implication for theoretical and methodo-
flood survivors. One of the reasons for this may be that logical perspectives, there are some limitations that need
breast cancer is an internal source of stress, and, according to be addressed. First of all, the drug use subscale was
to cognitive adaptation theory, regaining mastery and con- excluded from the current study because the sample of
trollability over the event counteracts the negative impact breast cancer patients reported very little usage. There-
of cancer (Taylor & Brown, 1988). Roesch and Weiner fore, whether or not the factor structure remains the same
(2001) have found that active coping is related to internal, when including drug use needs further exploration.
unstable, and controllable explanation of personal illness. Secondly, to control for traumatic stage and gender, the
Moreover, in early-stage cancer, fighting spirit, comprising current study limits the attention to the short-term after-
a belief that the disease is controllable and a determination math of trauma and female survivors. Since timing and
to cope with the situation using various active coping meth- gender may affect the coping structure, future studies
ods, is associated with better adjustment (Wang et al., should address this issue and retest the stability of the fac-
2013). In general, women actively attempt to alter or elim- tor structure of the Brief COPE on male and long-term
inate the stressors adjust well in cancer. survivors. Thirdly, because the psychological adjustment
Although we assumed that socially-supported coping is measures were different for the two groups, it was impos-
adaptive, it is unrelated to any psychological outcomes sible to fully compare the relationship between coping and
measured in this study, except for the positive association adjustment. Therefore, the data could only provide preli-
with life satisfaction in the flood survivor sample. To under- minary validity for the classification of the three coping
stand the nature of socially-supported coping, we examined strategies. It is hoped that future research will provide
the correlations between socially-supported coping and psy- more details of this ongoing issue. Finally, an inherent
chological outcomes in both sample. In the flood survivor confounding for the relationship between coping and psy-
sample, socially-supported coping was positively correlated chological adjustment, especially in cross-sectional design,
with life satisfaction (r = .173, .161, p < .05). In the breast is that the reciprocal effect is difficult to capture (whether
cancer patient sample, socially-supported coping had a neg- coping leads to reduced distress or distress leads to
ative correlation with depression (r = .190, p < .01), and a increased coping). Future research with prospective design
positive correlation with PA (r = .281, p < .001). Accord- is obviously required.
ingly, while socially-supported coping was correlated with
some psychological outcomes in both samples, the links dis-
appeared when other coping strategies were controlled for. References
The findings suggest that socially-supported coping, while
Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52,
important, may not be as prominent as self-sufficient cop-
317–332.
ing and avoidant coping at 3–4 months after the traumatic American Cancer Society. (2014). Breast cancer survival rates by
events. There are few explanations. First, due to the types stage. Retrieved from http://www.cancer.org/cancer/breast-
of stressors studied, socially-supported coping was less cancer/detailedguide/breast-cancer-survival-by-stage
Bellizzi, K. M., & Blank, T. O. (2006). Predicting posttraumatic
available and useful in the current study. For example,
growth in breast cancer survivors. Health Psychology: Official
because 8/8 Flood involved the vast displacement of so Journal of the Division of Health Psychology, American Psycho-
many people, as well as our sample, many survivors may logical Association, 25, 47–56.
not have had adequate access to their neighborhoods, com- Bradburn, N. M. (1969). The structure of psychological well-being.
Chicago, IL: Aldine.
munities, and sometimes even families. In addition, other
Cartwright, T., Endean, N., & Porter, A. (2009). Illness perceptions,
survivors may have been too overwhelmed to serve as coping and quality of life in patients with alopecia. The British
effective supports. Secondly, regarding trauma duration, Journal of Dermatology, 160, 1034–1039.
in the short-term aftermath of disaster, there is not enough Carver, C. S. (1997). You want to measure coping but your
protocol’s too long: Consider the brief cope. International
time to mobilize social network. For example, within 3–4
Journal of Behavioral Medicine, 4, 92–100.
months after cancer diagnosis, survivors have to deal with Carver, C. S., & Connor-Smith, J. (2010). Personality and coping.
not only the fact of being affected by cancer but surgery Annual Review of Psychology, 61, 679–704.

Ó 2016 Hogrefe Publishing European Journal of Psychological Assessment (2018), 34(3), 162–173
 172 A. W.-T. Wang et al., Measurement Invariance and Brief Cope

Carver, C. S., Pozo, C., Harris, S. D., Noriega, V., Scheier, M. F., Krägeloh, C. U. (2011). A systematic review of studies using the
Robinson, D. S., . . . Clark, K. C. (1993). How coping mediates Brief COPE: Religious coping in factor analyses. Religions, 2,
the effect of optimism on distress: a study of women with early 216–246.
stage breast cancer. Journal of Personality and Social Psychol- Lazarus, R. S. (1996). The role of coping in the emotions and how
ogy, 65, 375–390. coping changes over the life course. In C. Magai & S. H.
Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing McFadden (Eds.), Handbook of emotion, adult development, and
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

coping strategies: A theoretically based approach. Journal of aging (pp. 289–306). San Diego, CA: Academic Press.
Personality and Social Psychology, 56, 267–283. Lee, K. L., Ou, Y. L., Chen, S. H., & Weng, L. J. (2009). The
Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of psychometric properties of a short form of the CES-D used in
measurement invariance. Structural Equation Modeling, 14, the Taiwan longitudinal study on aging. Formosa Journal of
464–504. Mental Health, 22, 383–410.
Chen, S. H., Hung, F. C., Lin, Y. S., & Tseng, H. M. (2002). Trauma Litman, J. A. (2006). The COPE inventory: Dimensionality and
and psychosocial aftermath among high- and low-exposure relationships with approach-and avoidance-motives and posi-
adults three months post the 921 Chi-Chi earthquake in tive and negative traits. Personality and Individual Differences,
Taiwan. Chinese Journal of Psychology, 44, 167–188. 41, 273–284.
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of- Lu, J., Tseng, H., & Tsai, Y. (2003). Assessment of health-related
fit indexes for testing measurement invariance. Structural quality of life in Taiwan (I): Development and psychometric
Equation Modeling, 9, 233–255. testing of SF-36 Taiwan version. Taiwan Journal of Public
Coolidge, F. L., Segal, D. L., Hook, J. N., & Stewart, S. (2000). Health, 22, 501–511.
Personality disorders and coping among anxious older adults. Markus, H. R., & Kitayama, S. (1991). Culture and the self:
Journal of Anxiety Disorders, 14, 157–172. Implications for cognition, emotion, and motivation. Psycholog-
Coyne, J. C., & Racioppo, M. W. (2000). Never the Twain shall ical Review, 98(2), 224.
meet? Closing the gap between coping research and clinical Moscardino, U., Scrimin, S., Capello, F., Altoè, G., & Axia, G. (2008).
intervention research. The American Psychologist, 55, 655–664. Psychological adjustment of adolescents 18 months after the
Feaster, D., & Szapocznik, J. (2002). Interdependence of stress terrorist attack in Beslan, Russia: A cross-sectional study.
processes among African American family members: Influence Journal of Clinical Psychiatry, 69, 854–859.
of HIV serostatus and a new infant. Psychology and Health, 17, Neugarten, B. L., Havighurst, R. J., & Tobin, S. S. (1961). The
339–363. measurement of life satisfaction. Journal of Gerontology, 16,
Foa, E. B., Cashman, L., Jaycox, L., & Perry, K. (1997). The 134–143.
validation of a self-report measure of posttraumatic stress Ng, S. K., & Keung Leung, W. (2006). A community study on the
disorder: The Posttraumatic Diagnostic Scale. Psychological relationship between stress, coping, affective dispositions and
Assessment, 9, 445–451. periodontal attachment loss. Community Dentistry and Oral
Folkman, S., & Lazarus, R. S. (1980). An analysis of coping in a Epidemiology, 34, 252–266.
middle-aged community sample. Journal of Health and Social Norris, F. H., Friedman, M. J., & Watson, P. J. (2002). 60,000
Behavior, 21, 219–239. disaster victims speak: Part II. Summary and implications of
Folkman, S., & Lazarus, R. S. (1985). If it changes it must be a the disaster mental health research. Psychiatry: Interpersonal
process: Study of emotion and coping during three stages of a and Biological Processes, 65, 240–260.
college examination. Journal of Personality and Social Psychol- Norris, F. H., Friedman, M. J., Watson, P. J., Byrne, C. M., Diaz, E.,
ogy, 48, 150–170. & Kaniasty, K. (2002). 60,000 disaster victims speak: Part I. An
Gillen, G. (2006). Coping during inpatient stroke rehabilitation: An empirical review of the empirical literature, 1981–2001. Psy-
exploratory study. The American Journal of Occupational Ther- chiatry: Interpersonal and Biological Processes, 65, 207–239.
apy, 60, 136–145. O’Brien, C. W., & Moorey, S. (2010). Outlook and adaptation in
Glass, K., Flory, K., Hankin, B. L., Kloos, B., & Turecki, G. (2009). advanced cancer: a systematic review. Psycho-Oncology, 19,
Are coping strategies, social support, and hope associated with 1239–1249.
psychological distress among Hurricane Katrina survivors? O’Connor, S. M., Jardine, A. G., & Millar, K. (2008). The prediction
Journal of Social and Clinical Psychology, 28, 779–795. of self-care behaviors in end-stage renal disease patients
Gow, K., Warren, C., Anthony, D., & Hinschen, C. (2008). Retention using Leventhal’s Self-Regulatory Model. Journal of Psychoso-
and intentions to quit among Australian male apprentices. matic Research, 65, 191–200.
Education and Training, 50, 216–230. Penley, J. A., Tomaka, J., & Wiebe, J. S. (2002). The association of
Gutiérrez, F., Peri, J. M., Torres, X., Caseras, X., & Valdés, M. coping to physical and psychological health outcomes: A
(2007). Three dimensions of coping and a look at their meta-analytic review. Journal of Behavioral Medicine, 25,
evolutionary origin. Journal of Research in Personality, 41, 551–603.
1032–1053. Prado, G., Feaster, D. J., Schwartz, S. J., Pratt, I. A., Smith, L., &
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in Szapocznik, J. (2004). Religious involvement, coping, social
covariance structure analysis: Conventional criteria versus new support, and psychological distress in HIV-seropositive African
alternatives. Structural Equation Modeling: A Multidisciplinary American mothers. AIDS and Behavior, 8, 221–235.
Journal, 6, 1–55. Radloff, L. S. (1977). The CES-D scale: A self-report depression
Hulbert-Williams, N., Neal, R., Morrison, V., Hood, K., & Wilkinson, C. scale for research in the general population. Applied Psycho-
(2012). Anxiety, depression and quality of life after cancer logical Measurement, 1, 385–401.
diagnosis: What psychosocial variables best predict how patients Roesch, S. C., & Weiner, B. (2001). A meta-analytic review of
adjust? Psycho-Oncology, 21, 857–867. coping with illness: Do causal attributions matter? Journal of
Jöreskog, K., & Sörbom, D. (2004). LISREL 8.7 Computer Software. Psychosomatic Research, 50, 205–219.
Lincolnwood, IL: Scientific Software International. Schnider, K. R., Elhai, J. D., & Gray, M. J. (2007). Coping style use
Jemal, A., Bray, F., Center, M. M., Ferlay, J., Ward, E., & Forman, D. predicts posttraumatic stress and complicated grief symptom
(2011). Global cancer statistics. CA: A Cancer Journal for severity among college students reporting a traumatic loss.
Clinicians, 61, 69–90. Journal of Counseling Psychology, 54, 344–350.

European Journal of Psychological Assessment (2018), 34(3), 162–173 Ó 2016 Hogrefe Publishing
 A. W.-T. Wang et al., Measurement Invariance and Brief Cope 173

Schottenbauer, M. A., Klimes-Dougan, B., Rodriguez, B. F., and recommendations for organizational research. Organiza-
Arnkoff, D. B., Glass, C. R., & LaSalle, V. H. (2006). Attachment tional Research Methods, 3, 4–70.
and affective resolution following a stressful event: General Wang, W. T., Tu, P. C., Liu, T. J., Yeh, D. C., & Hsu, W. Y. (2013).
and religious coping as possible mediators. Mental Health, Mental adjustment at different phases in breast cancer
Religion & Culture, 9, 448–471. trajectory: Re-examination of factor structure of the Mini-
Skinner, E. A., Edge, K., Altman, J., & Sherwood, H. (2003). MAC and its correlation with distress. Psycho-Oncology, 22,
https://econtent.hogrefe.com/doi/pdf/10.1027/1015-5759/a000321 - Friday, January 26, 2024 9:35:37 PM - University of Wolverhampton IP Address:134.220.1.139

Searching for the structure of coping: A review and critique of 768–774.

category systems for classifying ways of coping. Psychological Wang, A. W. T., Chang, C. S., Chen, S. T., Chen, D. R., & Hsu, W. Y.
Bulletin, 129, 216–269. (2014). Identification of posttraumatic growth trajectories in
Skinner, E. A., & Zimmer-Gembeck, M. J. (2007). The development the first year after breast cancer surgery. Psycho-Oncology, 23
of coping. Annual Review of Psychology, 58, 119–144. (12), 1399–1405.
Snell, D. L., Siegert, R. J., Hay-Smith, E. J. C., & Surgenor, L. J. Ware, J., Snow, K., Kosinski, M., & Gandek, B. (1993). SF-36 health
(2011). Factor structure of the Brief COPE in people with mild survey manual and interpretation guide. Boston, MA: The Health
traumatic brain injury. The Journal of Head Trauma Rehabili- Institute, New England Medical Center.
tation, 26, 468–477. Wu, Y. C., Hung, F. C., & Chen, S. H. (2002). “Changes or not” is the
Stanton, A. L., Danoff-burg, S., & Huggins, M. E. (2002). The first question: The meaning of posttraumatic stress reactions
year after breast cancer diagnosis: Hope and coping strategies one year after the Taiwan chi-chi earthquake. Journal of the
as predictors of adjustment. Psycho-Oncology, 11, 93–102. Chinese Institute of Engineers, 25, 609–618.
Steenkamp, J.-B. E., & Baumgartner, H. (1998). Assessing mea- Zigmond, A. S., & Snaith, R. P. (1983). The hospital anxiety and
surement invariance in cross-national consumer research. depression scale. Acta Psychiatrica Scandinavica, 67, 361–370.
Journal of Consumer Research, 25, 78–107.
Su, Y.-J., & Chen, S.-H. (2008). A three-month prospective
Received November 12, 2013
investigation of negative cognitions in predicting posttraumatic
Revision received June 11, 2015
stress symptoms: The mediating role of traumatic memory
Accepted August 10, 2015
quality. Chinese Journal of Psychology, 50, 167–186.
Published online April 29, 2016
Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: a social
psychological perspective on mental health. Psychological
Wen-Yau Hsu
Bulletin, 103(2), 193–210.
National Chengchi University
Tobin, D. L., Holroyd, K. A., Reynolds, R. V., & Wigal, J. K. (1989).
Department of Psychology
The hierarchical factor structure of the Coping Strategies
64, Sec.2, Zhinan Rd. Wenshan District
Inventory. Cognitive therapy and Research, 13, 343–361.
Taipei City 11605
Turner-Cobb, J. M., Gore-Felton, C., Marouf, F., Koopman, C., Kim, P.,
Taiwan (R.O.C)
Israelski, D., & Spiegel, D. (2002). Coping, social support, and
Tel. +866 2 2938-7379
attachment style as psychosocial correlates of adjustment in men
Fax +866 2 2936-6725
and women with HIV/AIDS. Journal of Behavioral Medicine, 25,
E-mail [email protected]
337–353.
Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of
the measurement invariance literature: Suggestions, practices,

Ó 2016 Hogrefe Publishing European Journal of Psychological Assessment (2018), 34(3), 162–173