Abstract
The objective of this study was to extend the psychometric evaluation of a brief version of the Self Compassion Scale (SCS). A secondary analysis of data from an international sample of 1,967 English-speaking persons living with HIV disease was used to examine the factor structure, and reliability of the 12-item Brief Version Self Compassion Inventory (BVSCI). A Maximum Likelihood factor analysis and Oblimin with Kaiser Normalization confirmed a two-factor solution, accounting for 42.58% of the variance. The BVSCI supported acceptable internal consistencies, with .714 for the total scale and .822 for Factor I and .774 for Factor II. Factor I (lower self compassion) demonstrated strongly positive correlations with measures of anxiety and depression while Factor II (high self compassion) was inversely correlated with the measures. No significant differences were found in the BVSCI scores for gender, age, or having children. Levels of self-compassion were significantly higher in persons with HIV disease and other physical and psychological health conditions. The scale shows promise for the assessment of self-compassion in persons with HIV without taxing participants, and may prove essential in investigating future research aimed at examining correlates of self-compassion, as well as providing data for tailoring self-compassion interventions for persons with HIV.
Keywords: Brief self-compassion inventory, HIV, psychometrics, factor analysis
On the basis of a conceptualization of self-compassion, Neff (2003) developed a psychometrically sound and theoretically valid scale to measure this construct. Since the self-compassion scale (SCS) was first published, the majority of research on self-compassion has used the SCS. Although this measure is increasingly popular in use, several potential issues have been noted. The longer 26-item SCS may be too burdensome or semantically complex for use by some persons who are experiencing a chronic medical illness, including HIV/AIDS (Barnard & Curry, 2011).
A shortened yet structurally equivalent scale, the Brief Version Self Compassion Inventory (BVSCI), was developed to reduce participant burden and also offer researchers an opportunity to gather perspectives from persons who are ill or debilitated in varied clinical settings. The purpose the current study was to extend the psychometric evaluation of the BVSCI using data from an international study of persons with HIV disease.
Method
This secondary analysis explored associations between key self-care concepts (self-compassion, self-efficacy, and self-esteem) and antiretroviral treatment adherence, HIV transmission behaviors, and symptom management (Corless et al., 2012; Nokes et al., 2012; Webel et al., 2012). Study sites included Canada, China, Namibia, Thailand, and the U.S. The study was approved by Institutional Review Boards at each study site and all participants gave written consent. Participants at each site completed a packet of self-report study measures that included the BVSCI. Data on CD4 and length of HIV were also obtained by self-report. This analysis included sites where study measures were completed in English (n = 1,967). Participant characteristics are presented in Table 1.
Table 1.
Demographic Data (n = 1,969)
| Variable | Range | Mean | SD` | Frequency |
|---|---|---|---|---|
| Age | 18–73 yrs. | 45.11 | 9.48 | 1896 |
| CD4 Self Report | 0–10,000 | 482.94 | 431.02 | 1424 |
| Length of HIV Self Report | 1–31 yrs. | 12.93 yrs. | 7.53 | 1886 |
| Self Rating of Physical Condition | 0 – 10 | 6.76 | 2.18 | 1958 |
| Self Rating of Psychological Support | 0 – 10 | 6.79 | 2.38 | 1956 |
| Self Rating of Social Support | 0 – 10 | 6.94 | 2.64 | 1955 |
| Sex | ||||
| Male | 1416 (72.6%) | |||
| Female | 530 (27.2%) | |||
| Decline to State/Other | 4 (0.2%) | |||
| Education | ||||
| 11th Grade or Less | 526 (27.0%) | |||
| High School or GED | 717 (36.8%) | |||
| 2 Yrs. College/AA Degree/Technical School | 425 (21.8%) | |||
| College (BS or BA) | 211 (10.8%) | |||
| Masters Degree | 63 (3.2%) | |||
| Doctorate/Medical Degree/Law Degree | 9 (0.5%) | |||
| Ethnicity | ||||
| Asian/Pacific Islander | 357 (18.4%) | |||
| African American/Black | 750 (38.6%) | |||
| Hispanic | 258 (13.3%) | |||
| Native American Indian | 68 (3.5%) | |||
| White/Anglo (Non Hispanic) | 457 (23.5 %) | |||
| Other | 53 (2.7%) | |||
| Children | ||||
| No | 1000 (51.8%) | |||
| Yes | 931 (48.2%) | |||
| Work for Pay | ||||
| Do Not Work for Pay | 1448 (74.4%) | |||
| Do Work for Pay | 498 (25.6%) | |||
| Adequacy of Income | ||||
| Totally Inadequate | 542 (28.0%) | |||
| Barely Enough | 960 (49.6%) | |||
| Enough | 432 (22.3%) | |||
| Told You Have AIDS | ||||
| No | 1051 (53.0%) | |||
| Yes | 862 (45.0%) | |||
| Don’t Know | 39 (2.0%) | |||
| Other Health conditions | ||||
| Yes | 1182 (64.5%) | |||
| No | 651 (35.5%) |
Measures
Self-Compassion Inventory
The self-report 26-item SCS assesses three major components of self-compassion, including self-kindness, common humanity, and mindfulness. In addition to three major components of self-compassion, the scale assesses six factors that reflect the positive and negative poles of these three components of self-compassion (Neff, 2003).
The BVSCI possesses psychometric characteristics comparable to Neff’s 26-item SCS. Based on loadings on the SCS's three subscales (four items per subscale), a subset of 12 items that loaded highest on the three factors from Neff's original 26-item SCS was identified, thereby reducing the length of the scale from 26 to 12 items. Scores on these 12 items correlated .92 with total SCS scores (M. Leary, personal communication, December, 2012). The BVSCI was validated with three non-clinical samples of adults including 117 adult community members, 281 participants from a psychology subject pool, and 161 graduate students. The samples completed the 12-item BVSCI and the 26-item SCS, along with various combinations of personality measures. The BVSCI demonstrated patterns of correlations with other psychometric measures that have known relationships with self-compassion. Internal reliability of the 12-item scale was acceptable (α > .82) (M. Leary, personal communication, December, 2012).
Center for Epidemiology Studies Depression Scale (CES-D)
This non-diagnostic 20-item self-report scale measures the current level of depressive symptoms in community populations (Radloff, 1977). Items are rated on a 4-point Likert-type scale, ranging from 0 (rarely or none of the time) to 3 (most of the time). The CES-D yields a total score of 60. The alpha reliability estimate was 0.90 in a sample of 727 AIDS patients (Holzemer, 1999).
Symptom Checklist-90 (SCL-90)
Anxiety symptoms were assessed with the self-report 10-item anxiety subscale from the SCL-90 (Derogatis, Lipman & Covi, 1973; Derogatis & Unger, 2010). This subscale is scored on five-points (0–4) assessing ‘how much’ the respondent was bothered by each anxiety symptom in the past week (1=not at all; 5=extremely). Items are summed to obtain a total score, ranging from 0 to 40. The SCL-90 has been tested across numerous populations 13 years and older and has well-established reliability and validity (Derogatis & Cleary, 1977; Derogatis & Unger, 2010). Internal consistency of the SCL-90 anxiety subscale in this sample was high (α > .95).
Factor Structure of the BVSCI
It was hypothesized that the factor structure of the BVSCI would reflect the single higher order factor in the SCS (Neff, 2003). A Maximum Likelihood (ML) extraction method and Oblimin with Kaiser Normalization was conducted on the BVSCI data. The minimum factor loading was set at .40. The data had a skewness of −.316 (standard error .053) and a kurtosis at 1.219 (standard error .106), indicating normal distribution.
Contrary to expectation, it was found that a one factor model did not fit the data. Two independent factors with drawn eigenvalues greater than 1.0 were found, accounting for 42.58% of the total item variance. Examination of the scree-test showed a distinct break between the steep slope of the first two factors and the gradual trailing off of the remaining factors. The first 6 item factor (low self-compassion items) accounted for 23.79% of the variance while the second factor (high self-compassion items) accounted for 18.59% of the variance. Both subscales were highly correlated with the BVSCI total score (Factor 1, r = .80; Factor 2, r = .60), suggesting a possible higher order factor.
The 6 reverse scored items in factor one formed a separate factor from the 6 straightforward worded items in factor two, raising the question of whether such factors should be considered meaningful or discarded as an expected artifact of method effects or response bias (Brown, 2003; Weeks et al., 2005). The weak negative correlation found between the two subscales (r = −.011) offered support for a two-factor solution.
Means, standard deviations, factor loadings, and communalities for each item of the BVSCI are reported in Table 2.
Table 2.
Means, Standard Deviations, Factor Loadings, and Communalities for the BVSCI (n = 1,967)
|
Scale Items |
Factor Loadings | Communalities | ||
|---|---|---|---|---|
| Mean (SD) | 1 | 2 | ||
| *When I see aspects of myself that I don’t like, I get down on myself. | 3.21 (1.2) | .662 | −.244 | .517 |
| *When I’m feeling down, I tend to feel like most other people are probably happier than I am. | 3.21 (1.2) | .652 | −.168 | .452 |
| *I can be a bit cold-hearted towards myself when I’m unhappy. | 3.43 (1.2) | .650 | −.118 | .441 |
| *When I’m feeling down I tend to obsess and fixate on everything that’s wrong. | 3.34 (1.2) | .625 | −.246 | .451 |
| *When I’m really struggling, I tend to feel like other people must be having an easier time of it. | 3.25 (1.2) | .614 | −.281 | .465 |
| *When something painful happens, I tend to blow the incident out of proportion. | 3.47 (1.2) | .565 | −.142 | .388 |
| When something painful happens, I try to take a balanced view of the situation. | 3.30 (1.2) | .302 | .642 | .558 |
| When I fail at something important to me I try to keep things in perspective. | 3.45 (1.2) | .320 | .617 | .505 |
| When I’m going through a very hard time, I give myself the caring and tenderness I need. | 3.22 (1.3) | −.055 | .578 | .474 |
| I’m kind to myself when I’m unhappy or experiencing suffering. | 3.09 (1.2) | .432 | .549 | .384 |
| When I feel inadequate in some way, I try to remind myself that most people feel inadequate from time to time. | 2.81 (1.1) | −.103 | .533 | .455 |
| When I’m down, I remind myself that there are lots of other people in the world feeling like I am. | 2.89 (1.2) | .333 | .506 | .875 |
| Eigenvalue | 2.879 | 2.231 | ||
| Percent Variance | 23.99% | 18.59% | ||
Note: Extraction Method: Maximum Likelihood; Rotation method: Oblimin and Promax with Kaiser Normalization Loadings in bold are values greater than .40 and are retained for that factor.
Reverse coded items
Findings show that the total BVSCI was sufficiently relaible (α = .714). Cronbach’s alpha for Factor 1 subscale was .822, while the alpha for the Factor 2 subscale was .774.
It was hypothesized that the BVSCI would demonstrate similar patterns of correlations between the SCS and measures of anxiety and depression. The mean score for the CES-D was 21.5 (SD 11.7), while the mean score for the SCL-90-anxiety subscale was 17.8 (SD 9.1). The 6-item subscale based on Factor I (lower self compassion) demonstrated strongly positive correlations with the CES-D (r = .435; p = .000) and the SCL-90 anxiety subscale (r = .571; p = .000). The 6-item subscale based on Factor II (high self compassion) was inversely correlated with the CES-D (r = −.152; p = .000) and the SCL-90 anxiety subscale (r = −.289; p = .000).
The mean score for the BVSCI was 29.97 (SD 6.23), with a range of 10–50. Differences in mean BVSCI scores were found across race/ethnicity (p<.026), length of HIV disease (p<.017), and for having other chronic conditions (p<.000). Significant correlations were observed between Factor I scores and education (p<.010) and whether or not participants worked for pay (p<.006). Significant correlations were also observed between Factor II and race/ethnicity (p<.000), education (p<.026), having health insurance (p<.001), working for pay (p<.000), and having other health conditions (p<.000).
Discussion
This study provided an investigation of the psychometric properties of the BVSCI with a large international sample of people living with HIV disease. In contrast to a single factor solution, a two-factor structure was found. Findings suggest that this shortened instrument has acceptable reliability and validity, and they support the utility of using the BVSCI in future research.
The sample in this analysis consisted of adults with increased levels of depressive and anxiety symptoms, common co-morbidities in HIV disease (Kemppainen et al., 2006, 2012; Rabkin, 2008). In this analysis, the measures of anxiety and depression were independently associated with the two factors in the BVSCI. Items representing lower self compassion were positively associated with anxiety and depression, while items representing higher self compassion demonstrated inverse associations. Researchers have found that the longer, single factored SCS is negatively correlated with measures of both anxiety and depression (Barnard & Curry, 2011). It is noteworthy that the participants in previous studies studies were undergraduates with subclinical anxiety and depression. Although a recent meta-analysis by MacBeth & Gumley (2012) provides robust data on the association between self-compassion and lower psychopathology, there is a continued need for research focused on populations with increased levels of anxiety and depression.
In contrast with previous studies of undergraduates that found lower levels of self-compassion in women than men (Neff, 2003; Neff, Hseih, & Dejitthirat, 2005; Neff & McGeehee, 2010), results in this study were fairly consistent across gender. Study findings may reflect positive psychological changes that frequently result in both males and females following a negative life event such as being diagnosed with HIV (Milam, 2004). Despite struggles with a life threatening diagnosis, many people living with HIV disease experience positive changes, including a new outlook on life.
In a noteworthy finding, levels of self-compassion were significantly higher in persons with HIV disease who had other health conditions. A growing body of literature reports the associations between quality of life in HIV disease and mindfulness (Robinson, Matthews, & Witek-Janusek, 2003), self-forgiveness (Martin, Vosick, & Riggs, 2012), and self esteem (Anderson, 2000; Simoni, Huang, Goodry & Montaoya, 2006). An interesting future direction for research would be to examine correlates of self-compassion in persons who deal with other physical or psychological conditions.
Participants with a length of HIV at 10 years or less reported significantly higher scores on the BVSCI than participants with HIV greater than 10 years. This finding may relate to asymptomatic illness stage, fewer HIV-related symptoms, or levels of social support (Parkenham & Rinaldi, 2001). Disclosure rates have also been associated with the use of more adaptive coping strategies (Simoni, Demas, Mason, Drossman & Davis, 2000). Little work has been published, however, that examines the role of self-compassion as a psychological adaptation strategy across the stages of HIV disease.
Several important imitations must be considered. Although our sample was international, it primarily represented persons with HIV disease from North America. The smaller number of participants from China, Thailand, and Namibia limit generalizeability. The instruments used in the original study further limited our analysis. An additional limitation relates to the sample in this study. While self-compassion studies tend to include more females than males, the sample is reversed in this analysis. The primarily male sample reflects the demographics of the HIV epidemic. Future research should examine levels of self-compassion related to age and race or ethnicity. While the BVSCI would add to an understanding of self-compassion in this population, further testing of the 12-item scale is recommended.
Conclusions
Research instruments that effectively measure levels of self-compassion in persons with HIV disease can facilitate the development of interventions aimed at increasing levels of self-compassion. The BVSCI shows promise for examining the construct of self-compassion without taxing study participants.
Table 3.
Means and Standard Deviations for the Total BVSCI, Factor I, and Factor 2
| Total BVSCI* | Factor I | Factor 2 | ||||
|---|---|---|---|---|---|---|
| Variable | M | SD | M | SD | M | SD |
| Total Sample: | 29.97 | 6.17 | 18.81 | 4.94 | 10.98 | 3.75 |
| Gender: | ||||||
| Male | 29.91 | 6.14 | 18.83 | 4.90 | 11.09 | 3.72 |
| Female | 29.57 | 6.16 | 18.81 | 4.99 | 10.78 | 3.82 |
| Age: | ||||||
| Age 50 and older | 30.37 | 5.85 | 19.68 | 5.01 | 10.72 | 3.68 |
| Under age 50 | 29.75 | 6.28 | 18.57 | 4.90 | 11.18 | 3.78 |
| Children: | ||||||
| Having children | 29.84 | 6.34 | 18.92 | 4.95 | 10.93 | 3.84 |
| Without children | 29.79 | 5.97 | 18.75 | 4.93 | 11.05 | 3.66 |
| Ethnicity: | ||||||
| White | 30.38* | 4.80 | 18.60 | 4.91 | 11.79 | 3.73 |
| Non-White | 29.64* | 6.49 | 18.89 | 4.94 | 10.77 | 3.73 |
| Other Health Conditions: | ||||||
| Yes | 30.23*** | 5.82 | 18.58 | 4.90 | 11.47 | 3.77 |
| No | 28.98*** | 6.68 | 18.89 | 5.08 | 10.13 | 3.61 |
| AIDS | ||||||
| Received an AIDS diagnosis | 29.58 | 6.27 | 18.66 | 4.88 | 10.94 | 3.71 |
| Not received an AIDS diagnosis | 30.03 | 6.01 | 18.93 | 5.00 | 11.09 | 3.79 |
| Not sure | 28.28 | 7.60 | 18.81 | 5.47 | 9.37 | 3.66 |
| Length of HIV Disease | ||||||
| 10 years or less | 30.17* | 5.96 | 19.24 | 4.96 | 10.93 | 3.60 |
| Greater than 10 years | 29.48* | 6.26 | 18.43 | 4.87 | 11.07 | 3.91 |
BVSCI total scores were calculated by reverse coding of 6 items and then summing both subscale means.
Acknowledgements
This project was supported in part by: NIH UL1 285 RR024131; NIHT32NR007081; NIH KL2RR024990; NIH R15NR011130; International Pilot Award, University of Washington Center for AIDS Research; University of Washington, School of Nursing; University of British Columbia School of Nursing Helen Shore Fund; Duke 290 University School of Nursing Office of Research Affairs; MGH Institute of Health Professions; Rutgers College of Nursing; City University of New York; Irwin Belk Distinguished Professorship Fund-University of North Carolina Wilmington . The contents of this manuscript 295 are solely the responsibility of the authors and do not necessarily represent the official views of the NIH or any other funders.
References
- Anderson E. Self-esteem and optimism in men and women infected with HIV. Nursing Research. 2000;49(5):262–271. doi: 10.1097/00006199-200009000-00005. [DOI] [PubMed] [Google Scholar]
- Barnard L, Curry J. Self-compassion: Conceptualizations, correlates, & interventions. Review of General Psychology. 2011;15(4):289–303. [Google Scholar]
- Brown TA. Confirmatory factor analysis of the Penn State Worry Questionnaire: Multiple factors or method effects? Behavior Research and Therapy. 2003;41:1411–1426. doi: 10.1016/s0005-7967(03)00059-7. [DOI] [PubMed] [Google Scholar]
- Cargill V, Stone V. HIV/AIDS: A minority health issue. Medical Clinics of North America. 2005;89(4):895–912. doi: 10.1016/j.mcna.2005.03.005. [DOI] [PubMed] [Google Scholar]
- Corless IB, Guarino AJ, Nicholas PK, Tyer-Viola L, Kirksey K, Brion J, Holzemer W. Mediators of Antiretroviral Adherence: A Multisite International Study. AIDS Care. 2012 doi: 10.1080/09540121.2012.701723. [ePub ahead of print]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Derogatis L, Lipman R, Covi L. SCL-90: An outpatient psychiatric rating scale-preliminary report. Psychopharmacology Bulletin. 1973;9(1):13–27. [PubMed] [Google Scholar]
- Kemppainen J, Eller L, Bunch E, Hamilton M, Dole P, Holzemer W, Tsai Y-F. Strategies for self-management of HIV-related anxiety. AIDS Care. 2006;18(6):597–607. doi: 10.1080/09540120500275726. [DOI] [PubMed] [Google Scholar]
- Kemppainen J, Wantland D, Voss J, Kirksey K, Corless I, Willard S, Portillo C. Self-care behaviors and activities for managing HIV-related anxiety. Journal of the Association of Nurses in AIDS Care. 2012;23(2):111–123. doi: 10.1016/j.jana.2011.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- MacBeth A, Gumley A. Exploring compassion: A meta-analysis of the association between self-compassion and psychopathology. Clinical Psychology Review. 2012;32:545–552. doi: 10.1016/j.cpr.2012.06.003. [DOI] [PubMed] [Google Scholar]
- Martin L, Vosvick M, Riggs S. Attchment, forgiveness, and physical health quality of life in HIV+ adults. AIDS Care. 2012 doi: 10.1080/09540121.2011.648598. [DOI] [PubMed] [Google Scholar]
- Milam J. Posttraumatic growth among HIV/AIDS patients. Journal of Applied Social Psychology. 2004;34:2353–2376. [Google Scholar]
- Neff K. Development and validation of a scale to measure self-compassion. Self and Identity. 2003;2:223–250. [Google Scholar]
- Neff K, Hseih Y, Dejitthirat K. Self-compassion, achievement goals, and coping with academic failure. Self and Identity. 2005;4:263–287. [Google Scholar]
- Neff K, Pisitsungkagarn K, Hsieh Y-P. Self-compassion and self-construal in the United States, Thailand, and Taiwan. Journal of Cross cultural Psychology. 2008;39:267–285. [Google Scholar]
- Neff K, McGeehee P. Self-compassion and psychological resilience among adolescents and young adults. Self and Identity. 2010;9:225–240. [Google Scholar]
- Neff K. Self-compassion, self-esteem, and well-being. Social and personality Psychology Compass. 2011;5(1):1–12. [Google Scholar]
- Nokes K, Johnson MO, Webel A, Dawson Rose C, Phillips JC, Sullivan K, Holzemer WL. Focusing on treatment self-efficacy to improve HIV treatment adherence. Journal of Nursing Scholarship. 2012 Nov; doi: 10.1111/j.1547-5069.2012.01476.x. [Epub ahead of print] [DOI] [PMC free article] [PubMed] [Google Scholar]
- Parkenham K, Rinaldi S. The role of illness, resources, appraisal and coping strategies in adjustment to HIV/AIDS: The direct and buffering effects. Journal of Behavioral Medicine. 2001;24(3):259–279. doi: 10.1023/a:1010718823753. [DOI] [PubMed] [Google Scholar]
- Rabkin J. HIV and depression: 2008 review and update. Current HIV/AIDS Reports. 2008;5:163–171. doi: 10.1007/s11904-008-0025-1. Accessed http://faculty.washington.edu/jsimoni/Rabkin2008.pdf. [DOI] [PubMed] [Google Scholar]
- Radloff L. The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement. 1977;1:385–401. [Google Scholar]
- Robinson F, Matthews H, Witek-Janusek L. Psycho-endocrine-immune response to mindfulness-based stress reduction in individuals infected with the Human Immunodeficiency Virus: A quasiexperimental study. The Journal of Alternative and Complementary Medicine. 2003;9(5):683–694. doi: 10.1089/107555303322524535. [DOI] [PubMed] [Google Scholar]
- Simoni J, Demas P, Mason H, Drossman J, Davis M. HIV disclosure rates among women with African American descent: Association with coping, social support, and psychological adaptation. AIDS and Behavior. 2000;4(2):147–154. [Google Scholar]
- Simoni J, Huang B, Goodry E, Montoya H. Social support and depressive symptomatology among HIV-positive women: The mediating role of self-esteem and mastery. Women and Health. 2006;42(4):1–15. doi: 10.1300/j013v42n04_01. [DOI] [PubMed] [Google Scholar]
- Weeks J, Heimberg R, Fresco D, Hart T, Turk C, Schneier F, Liebowitz M. Empirical validation and psychometric evaluation of the Brief Fear of Negative Evaluation Scale in patients with social anxiety disorder. Psychological Assessment. 2005;17(2):179–190. doi: 10.1037/1040-3590.17.2.179. [DOI] [PubMed] [Google Scholar]
- Webel A, Phillips JC, Dawson Rose C, Holzemer WL, Chen W-T, Tyer-Viola L, Salata RA. A cross-sectional description of social capital in an international sample of persons living with HIV/AIDS (PLWH) BMC Public Health. 2012;12:188. doi: 10.1186/1471-2458-12-188. [DOI] [PMC free article] [PubMed] [Google Scholar]