Background: The World Health Organization has asserted the importance of enhancing participation of people with disabilities within the International Classification of Functioning, Disability and Health framework. Participation is regarded as a vital outcome in community-based rehabilitation. The actualization of the right to participate is limited by social stigma and discrimination. To date, there is no validated instrument for use in Chinese communities to measure participation restriction or self-perceived stigma. This study aimed to translate and validate the Participation Scale and the Explanatory Model Interview Catalogue (EMIC) Stigma Scale for use in Chinese communities with people with physical disabilities. Methods: The Chinese versions of the Participation Scale and the EMIC stigma scale were administered to 264 adults with physical disabilities. The two scales were examined separately. The reliability analysis was studied in conjunction with the construct validity. Reliability analysis was conducted to assess the internal consistency and item-total correlation. Exploratory factor analysis was conducted to investigate the latent patterns of relationships among variables. A Rasch model analysis was conducted to test the dimensionality, internal validity, item hierarchy, and scoring category structure of the two scales. Results: Both the Participation Scale and the EMIC stigma scale were confirmed to have good internal consistency and high item-total correlation. Exploratory factor analysis revealed the factor structure of the two scales, which demonstrated the fitting of a pattern of variables within the studied construct. The Participation Scale was found to be multidimensional, whereas the EMIC stigma scale was confirmed to be unidimensional. The item hierarchies of the Participation Scale and the EMIC stigma scale were discussed and were regarded as compatible with the cultural characteristics of Chinese communities. Conclusion: The Chinese versions of the Participation Scale and the EMIC stigma scale were thoroughly tested in this study to demonstrate their robustness and feasibility in measuring the participation restriction and perceived stigma of people with physical disabilities in Chinese communities. This is crucial as it provides valid measurements to enable comprehensive understanding and assessment of the participation and stigma among people with physical disabilities in Chinese communities. Keywords: Participation, Disability, Stigma, Community-based rehabilitation * Correspondence: firstname.lastname@example.org Department of Special Education and Counseling, The Education University of Hong Kong, 10 Lo Ping Road, Tai Po, New Territories, Hong Kong School of Medical and Health Sciences, Tung Wah College, 31 Wylie Road, Homantin, Kowloon, Hong Kong Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 2 of 12 Background immediately noticeable by an observer . The effect Community-based rehabilitation (CBR) aims to promote of stigma on participation as experienced by people with the rights and opportunities for people with disabilities physical disabilities is adverse, and it affects their mental . Through CBR programs, people with disabilities health, physical health, and overall quality of life. In are enabled to participate in their community and soci- Chinese communities, lower self-concept and fewer ety. Within a human rights framework, CBR is promoted quality social relationships are evident among people to remove the obstacles, barriers, and discrimination with physical disabilities as a result of stigmatization [4, that hinder the participation of people with disabilities. 33]. For a comprehensive understanding and to assess It is also advocated to promote the active participation participation and stigma among affected people, it is es- of people with disabilities and their caregivers through sential to have a validated instrument that can be effect- appropriate measures to attain their maximum inde- ively used by communities. pendence and full participation in all aspects of life . This study aimed to translate and validate two in- Participation refers to involvement in life situations . struments, namely the Participation Scale and the Ex- Problems an individual may experience in involvement planatory Model Interview Catalogue (EMIC) stigma in life situations are classified as participation restric- scale, for use in Chinese communities. Both the Par- tions . Activity limitations and restrictions on partici- ticipation Scale and the EMIC focus on health-related pation are more critical to the affected person than the stigma; the EMIC assesses perceived stigma and the underlying health condition. Evidence on the social par- Participation Scale assesses the impact of stigma on ticipation of people with disabilities is essential in pro- social participation . These two scales are fre- gram planning, monitoring, and assessing the effect of quently used and put in the disability toolkit for use interventions aimed at reducing participation restric- in community-based inclusive development programs tions. Knowledge regarding the degree of participation . The Participation Scale is an interview-based in- restriction of a person is useful in informing the progress strument for measuring the level of participation re- of the person as a result of an intervention. However, striction of people with disabilities . The there is no universal accepted definition of participation instrument has good content validity because it covers , participation restrictions are a very widespread most of the domains of participation in the Inter- phenomenon, and scientific evidence and data on par- national Classification of Functioning, Disability and ticipation restrictions are limited . Health . Validation studies have demonstrated its Social stigma and discrimination constitute a critical high internal consistency (Cronbach’s alpha = 0.92), environmental factor that limits participation and con- high interrater reliability (r = 0.80), and high discrim- tributes to disabilities . Stigma is regarded as a set of inant validity for use with different target groups, prejudices, stereotypes, discriminatory beliefs, and biases such as people with leprosy and AIDS, in Nepal, linked to the characteristics that differentiate a person India, and Brazil . The EMIC stigma scale is an from others . Social stigma is defined as the attitudes interview-based instrument for assessing perceived of others toward people with disabilities; enacted stigma stigma. The EMIC stigma scale has been adopted in a refers to the actual episodes of discrimination against non-Chinese context for people with HIV/AIDS and people with disabilities; felt stigma is the stigmatization leprosy with acceptable discriminant and convergent as experienced by the person; and self-perceived stigma validity, interitem reliability, and test–retest reliability is the stigma perceived when having a painful inner [26, 30]. struggle about a disability, even without any encounter This study provides data for answering two research with actual stigmatization . Perception of stigma and questions regarding the validity of the Participation Scale experience of discrimination cause people to feel and EMIC stigma scale. First, the Participation Scale and ashamed and may cause anxiety, depression, and isola- the EMIC stigma scale are rarely employed to study tion . Measuring stigma is crucial because the evi- people with physical disabilities. However, they are dence obtained from such assessment constitutes a widely adopted in the fields of mental illness  and valuable part of a situational analysis in the planning, chronic disease . It is unclear whether they can be monitoring, and evaluation of CBR service. Evidence ob- equally valid when they are applied to people with phys- tained regarding intensity of stigma is helpful in advocat- ical disabilities. The second question is whether the val- ing the participation rights of people with disabilities in idity of the Participation Scale and EMIC stigma scale in society. a Chinese cultural context is as clear as that in a Evidence regarding measurement of stigma and par- non-Chinese context, where the stigmatization of dis- ticipation is essential in building a strong evidence base abilities in Chinese society is distinctively influenced by for CBR in Chinese communities. Physical disabilities its traditional cultural values. The specific objectives of are regarded as visible disabilities and thus are this study are: Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 3 of 12 1. To translate the Participation Scale and the EMIC levels of participation by score: (1) no significant restriction stigma scale into a traditional Chinese version. (0–12), (2) mild restriction (13–22), (3) moderate restric- 2. To examine the reliability and construct validity of tion (23–32), (4) severe restriction (33–52), and (5) extreme the Participation Scale and the EMIC stigma scale. restriction (53–90). The instrument, in its original language, has good content validity as it covers nine domains of par- Methods ticipation: learning and applying knowledge, general tasks First, both the Participation Scale and the EMIC stigma and demands, communication, mobility, self-care, domestic scale were translated from English to Chinese according life, interpersonal interactions and relationships, major life to the guidelines stated by the authors . A back areas, and community, social, and civil life. Van Brakel and translation to English was performed by another bilin- colleagues  validated the instrument scores against ex- gual translator. A panel of academic and clinical experts, pert scores and supported the external validity of the Par- including an occupational therapist, a clinical psycholo- ticipation Scale. gist, and a sociologist, was formed to review the content The EMIC stigma scale is a 15-item instrument, origin- validity of the Chinese version. Minor amendments to ally designed to measure stigma among leprosy-affected some of the wording were made to ensure readability. people. Because this study employed the EMIC stigma The psychometric properties and construct validity of scale to measure stigma among people with physical dis- the revised scales were examined. abilities, “leprosy” was replaced with “physical disability” in each question. Each question was measured with four Participants options, which were “yes,”“possibly,”“uncertain,” and A total of 264 adults with physical disabilities were re- “no.” Scores were generated by assigning 3 points to “yes,” cruited for this study. People affiliated with the local or- 2to “possibly,” 1to “uncertain,” and 0 to “no” for all ques- ganizations for persons with physical disabilities (DPOs) tions except question 2, in which a reverse scoring method were targeted. Physical disabilities are operationally de- was employed. A composite score was obtained for each fined as a chronic physical impairment affecting one or respondent by adding the scores of the 15 questions. A more areas of the body, including the central nervous higher score implied a higher level of perceived stigma system, spinal cord, peripheral nervous system, and per- faced by the respondent. The internal consistency of the ipheral structures . The inclusion criteria were (1) an original scale (as applied in non-Chinese communities) is age of 18 to 65 years; (2) not being in an acute phase of good, with a Cronbach’s alpha coefficient of 0.79. an illness or condition; (3) being mentally clear; and (4) having sufficient cognitive ability to comply with the in- Data collection structions to complete the test. The participants were re- Upon consent of the participants, the Participation Scale cruited from six types of DPO: ankylosing spondylitis, and the EMIC stigma scale were administered in a spinal cord injuries, developmental conditions with face-to-face interview. The interviewers were trained ac- physical disabilities, brain damage, rheumatoid arthritis, cording to the guidelines and protocol of the IELP . and work-related orthopedic injuries. DPOs were contacted and liaised by the principal in- Data analysis vestigator. Upon consent of the DPOs to participate in Reliability analysis and convergent validity of the two in- this study, the research team sent invitation letters and struments was performed using SPSS 21.0. Internal information sheets to all members. Ethical approval from consistency and item-total correlation were examined. the Committee on the Use of Human and Animal Sub- Reliability means that a measure consistently reflects the jects in Teaching and Research of Tung Wah College construct that it measures. Cronbach’s alpha was calcu- was obtained (HASC1415H04). All participants con- lated to examine the internal consistency of the two sented to participating in this study. scales. If a scale is reliable, the overall reliability is not expected to be greatly affected by any one item. It is Instruments therefore essential to also investigate the value of Cron- The Participation Scale is an 18-item interview-based in- bach’s alpha if an item is deleted. All values of alpha are strument for measuring the level of participation among approximately 0.8 or higher in a reliable scale. The people with disabilities. When respondents reported restric- values of the corrected item-total correlation should be tion in a specific area (“no” or “sometimes”), they were above 0.3 to confirm that all items are correlated with asked to indicate the level of restriction. The choices were the total score . Convergent validity indicates that (1) no problem, (2) a small problem, (3) a moderate prob- two measures that are considered to reflect the same lem, and (4) a large problem. The sum of scores was calcu- underlying phenomenon will correlate significantly . lated, with a higher total score representing a lower level of Convergent validity was tested by analyzing the correl- general participation. The respondents were ranked in five ation coefficient of the two measures. The Pearson Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 4 of 12 product-moment correlation coefficient was planned if entail the use of a Likert scale to collect data and both the data were found to be normally distributed; if not, had more than two response options. In this analysis, the Spearman rank correlation coefficient was planned. the data were first cleaned based on misfit person diag- This study used a score of r < 0.25 to indicate a weak nosis. A person was excluded if the point measure cor- correlation; r = 0.25 to 0.5 a moderate correlation; and relation was negative, the outfit mean square value r > 0.50 a strong correlation . (MNSQ) was greater than 2, or the Z-standard value Exploratory factor analyses for the two scales were was greater than 2 . Rasch model analysis was per- conducted separately using SPSS. Factor analysis entails formed using Winsteps 3 software to examine the sum- examining the structure within numerous variables. mary statistics, category structure, dimensionality, and Constructs must be defined by relevant measurable vari- model fit. ables that can be collated to form a conceptual package Dimensionality is a key part of the assessment of con- called a factor. Using an exploratory approach to factor struct validity; it shows whether the items are measuring analysis allows the researcher to sort through numerous a single underlying dimension or several separate dimen- variables to reveal latent patterns of relationships among sions . In Rasch model analysis, the principal com- variables. When used to test the construct validity of an ponent analysis of the residuals allows for a test of the instrument, it simulates the process of theory testing, local independence of items. The absence of any mean- which means that the factors emerging from the process ingful pattern in the residuals supports the assumption of analysis should match a hypothesized variable group- of unidimensionality. To confirm that the scale is unidi- ing . The correlation of an individual item with a mensional, the unexplained variance in the first contrast factor is called a factor loading. A correlation above +.30 should not be greater than 2, and it should be smaller or below −.30 indicates that an item contributes mean- than the raw variance explained by the items . Items ingfully to a factor . In this study, exploratory factor were considered misfit if the point measure correlation analysis was performed using principal axis factoring was negative, the value of the ZSTD exceeded 2, and with eigenvalues greater than 1. Principal axis factoring both the infit and outfit MNSQs exceeded 1.64 . Not- is preferable to principal component analysis because ably, both exploratory factor analysis and Rasch model principal component analysis is only a data deduction analysis were employed to test the dimensionality of the method rather than factor analysis [7, 13]. The purpose two scales. This study conducted exploratory factor ana- of exploratory factor analysis is to derive a more parsi- lysis first and then Rasch model analysis. Exploratory monious conceptual understanding of a set of variables factor analysis was conducted to explore the underlying by determining the number and nature of common fac- concepts that the items are measuring, with the poten- tors required to account for the pattern of correlations tial to explore the meaning of subscale scores. Rasch among the measured constructs . Exploratory factor model analysis was performed to test a unidimensional analysis is based on the common-factor model. Principal score (measurement) scale. In this case, exploratory fac- axis factoring analyzes shared variance among the items. tor analysis revealed the patterns of relationships among It is a factor analysis method that entails extracting fac- items to form latent constructs . It helped to define tors on the basis of a reduced correlation matrix by the underlying construct of participation (participation using a priori communality estimates. Oblique rotation scale) and self-stigma (EMIC). Rasch model analysis was used in this study because the factors might be cor- entailed using data for measurement, and the objective related with each other. An oblique rotation theoretically of this analysis was to test and confirm a unidimensional renders a more accurate and reproducible solution, be- interval scale . In other words, exploratory factor ana- cause it is generally expected in the social sciences that lysis was used in this study as an exploratory device to some correlation exists among factors . The absolute make sense of the data. Once the factors had been iden- values were suppressed in the coefficient display when tified, Rasch model analysis was used to further confirm the factor loading was less than 0.30. that the measurement was unidimensional . Rasch model analyses for the EMIC stigma scale and the Participation Scale were conducted separately. The Results Rasch model is based on the concept that useful meas- Basic demographics urement involves examining only one human attribute The Participation Scale and the EMIC stigma scale were at a time (unidimensionality) on a hierarchical line of administered to 264 adults with physical disabilities. All inquiry . If an instrument is valid, each of the items participants were aged 18 to 65; 50.8% were married; should contribute meaningfully to the construct being and 38.7% were of low socio-economic status. Types of investigated, and the recorded performance is a reflec- condition were rheumatoid arthritis, acquired brain tion of a single underlying construct. A polytomous damage, spinal cord injury, ankylosing spondylitis, model was chosen in this study because both scales orthopedic injuries, and congenital physical disabilities. Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 5 of 12 The basic demographics of the participants are shown in stigma scale were less than 0.3, meaning that they Table 1. weakly correlated with the total score. Reliability analysis Convergent validity The Cronbach’s alpha values representing internal The total scores of the two scales were correlated to test consistency were 0.93 and 0.897 for the Participation the convergent validity and assess the relationship of the Scale and the EMIC stigma scale, respectively. If an item two scales. Because the total scores of the two scales was deleted, the value of Cronbach’s alpha for all items were not normally distributed, the Spearman’s rank in each scale was higher than 0.8 (Tables 2 and 3). The order correlation was used in the analysis. The Spear- reliability of both the Participation Scale and the EMIC man’s rank order correlation showed a moderate to stigma scale was confirmed. However, the corrected strong correlation (r = 0.48, p = 0.001) among the find- item-total correlations for items 1 and 2 of the EMIC ings of the two scales . The convergent validity of the two scales was therefore confirmed. Table 1 Demographic and clinical characteristics of participants Exploratory factor analysis (n = 264) Data were cleaned to exclude all cases with missing data Variables Frequency Percent from the analysis. For the exploratory factor analysis of Gender the Participation Scale, a total of 256 valid cases were in- Male 116 43.9 cluded. The Kaiser–Meyer–Olkin measure of sampling Female 148 56.1 adequacy was 0.924, which meant that the data were ad- Age equate for exploratory factor analysis. Using principal axis factoring and oblique rotation (promax), three fac- 18–25 8 3.0 tors were extracted with eigenvalues greater than 1, and 26–35 21 8.0 the absolute values of factor loadings less than 0.30 were 36–45 44 16.7 suppressed (Table 4). Factor 1 comprised items 4 (travel 46–55 85 32.2 outside your neighborhood), 5 (take part in festivals), 6 56–65 106 40.1 (take part in social activities), 7 (being as socially active Education as peers), 12 (move around the house and village), 13 (visit public places), and 14 (do household work). Factor Uneducated 2 0.8 2 consisted of items 8 (have respect in the community), Primary school 54 20.5 9 (have opportunity to take care of oneself and others), Secondary school 163 61.7 10 (have opportunity to enter into and maintain College 45 17.0 long-term relationships), 11 (visit other people in the Marital status community), 15 (opinion count in family discussion), 16 Single 80 30.3 (help other people), 17 (comfortable meeting new people), and 18 (confident to learn and try new things). Married 134 50.8 Factor 3 comprised items 1 (find job), 2 (work as hard as Divorced 32 12.1 others), and 3 (contribute to household economically). Widowed 18 6.8 For the EMIC stigma scale, a total of 245 cases were Household Income (HKD) included in the analysis. The value of the Kaiser– < $5000 40 15.2 Meyer–Olkin measure of sampling adequacy was 0.905. $5001 - $10,000 62 23.5 Using principal axis factoring and oblique rotation (pro- max), two factors were extracted with eigenvalues of $10,001 - $ 20,000 79 30.0 greater than 1, and the absolute values of factor loadings $20,001 - $40,000 56 21.2 less than 0.30 were suppressed (Table 5). Factor 1 com- > $40,001 27 10.2 prised items 3 (reduced pride or self-respect), 4 (feel Condition ashamed or embarrassed), 5 (neighbors, colleagues, or Ankylosing spondylitis 31 11.7 others have less respect), 6 (contact might have bad ef- Spinal cord injury 46 17.4 fects on others), 7 (others avoid you), 8 (some people re- fuse to visit you), 9 (colleagues and neighbors think less Congenital physical disabilities 18 6.8 of your family), 10 (cause problems for the children), 11 Acquired brain damage 70 26.5 (problem in getting married and marriage), 12 (the dis- Rheumatoid arthritis 103 39.0 ease makes it difficult for family members to marry), 13 Orthopaedic injuries 27 10.2 (have been asked to stay away from social groups), 14 Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 6 of 12 Table 2 Item-total Statistics of the Participation Scale Corrected Item-Total Correlation Cronbach’s Alpha if Item Deleted P-Scale 1 Find job .451 .930 2 Work as hard as others .577 .927 3 Contribute to household economically .568 .927 4 Travel outside your neighborhood .722 .923 5 Take part in festivals .729 .923 6 Take part in social activities .715 .923 7 Socially active as peers .739 .923 8 Have respect in the community .602 .926 9 Have opportunity to take care of oneself and others .582 .926 10 Have opportunity to enter into /maintain long term relationship .570 .927 11 Visit other people in the community .636 .925 12 Move around the house and village .655 .925 13 Visit public places .670 .924 14 Do household work .589 .927 15 Opinion count in family discussion .651 .925 16 Help other people .687 .924 17 Comfortable meeting new people .634 .925 18 Confident to learn and try new things .586 .926 (decided to stay away from work or social groups), and Rasch model analysis 15 (people think that you also have other health prob- Participation scale lems). Factor 2 consisted of items 1 (keep people from knowing about the disability) and 2 (discussing the prob- Internal validity Using data cleaning (as described in lem with others), as shown in Table 4. Items 1 and 2 the previous section), six persons were removed from were evidently clustered in one factor of self-disclosure. the data file of the Participation Scale. Determined from Table 3 Item-total Statistics of the EMIC stigma scale Corrected Item-Total Correlation Cronbach’s Alpha if Item Deleted EMIC 1 Keep people from knowing about the disability .188 .904 2 Discuss the problem with others .222 .902 3 Reduce the pride or self-respect .658 .887 4 Feel ashamed or embarrassed .699 .885 5 Have less respect for you because of your problems .693 .885 6 The contact might have bad effects on others .657 .887 7 The others avoid you .712 .885 8 Some people refuse to visit you .675 .886 9 The colleagues and neighbors think less of your family .664 .887 10 Cause problems for the children .632 .888 11 Problem in getting married and marriage .525 .893 12 makes it difficult for family members to get married .632 .888 13 asked to stay away from social groups .537 .892 14 Stay away from work or social groups .551 .891 15 People think that you also have other health problems .558 .891 Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 7 of 12 Table 4 Exploratory factor analysis of the Participation Scale (n = 256) Factor 12 3 4 Travel outside your neighborhood .879 13 Visit public places .874 12 Move around the house and village .735 7 Socially active as peers .652 6 Take part in social activities .630 5 Take part in festivals .578 14 Do household work .444 18 Confident to learn and try new things .889 17 Comfortable meeting new people .887 8 Have same respect in the community .691 16 Help other people .454 10 Have opportunity to enter into /maintain long term relationship .428 15 Opinion count in family discussion .420 .331 9 Have opportunity to take care of oneself and others .387 11 Visit other people in the community .336 .340 1 Find job .800 2 Work as hard as others .751 3 Contribute to household economically .555 Note. Absolute values were blanked in the coefficient display when the factor loading was less than 0.30 the summary statistics of the Rasch analysis, the item re- another parallel set of items measuring the same construct liability was 0.72 and the person reliability was 0.81. . The commonly accepted range for the mean-square These statistics demonstrate the good reliability of the (MNSQ) is 0.6 to 1.4 and −2to +2for thestandardized scale and high replicability of person ordering, indicating value (ZSTD) . The results of the analysis showed that that the results would not vary if this sample were given the person fit was good, with the infit and outfit MNSQs being 1.05 and 1.03 and the ZSTDs being − 0.4 and − 0.5, Table 5 Exploratory factor analysis of the EMIC (n = 245) respectively. The item fit was confirmed as good, with the Factor infit and outfit MNSQs being 1.01 and 1.03 and the ZSTDs being − 0.1 and 0.0. Cronbach’s alpha was 0.93, 5 Have less respect for you because of your problems .762 and the value of item separation was 1.59, which indicated 7 The others avoid you .760 good reliability of the Participation Scale. 9 The colleagues and neighbors think less of your family .740 Dimensionality The Rasch-residual-based principal 8 Some people refuse to visit you .731 component analysis (PCAR) showed that the unex- 4 Feel ashamed or embarrassed .684 plained variance explained by the first contrast was 2.2 12 makes it difficult for family members to get married .672 (6.9%), and the raw variance explained by the items was 6 The contact might have bad effects on others .672 29.7%. However, examining the misfit order of all items 15 People think that you also have other health problems .654 revealed that items 1 and 3 were misfitted because the value of the ZSTD exceeded 2 and both the infit and 10 Cause problems for the children .648 outfit MNSQs exceeded 1.64. With an objective to test 14 Stay away from work or social groups .626 and confirm a unidimensional interval measurement 3 Reduce the pride or self-respect .608 scale using Rasch model analysis, items 1 and 3 were re- 11 Problem in getting married and marriage .588 moved from the scale and dimensionality of the 16-item 13 asked to stay away from social groups .576 Chinese version of the Participation Scale was evaluated. 1 Keep people from knowing about the disability .681 This 16-item scale was then found to be unidimensional because the unexplained variance in the first contrast 2 Discuss the problem with others .566 was reduced to 2.0 (6.8%) and the raw variance ex- Notes. Absolute values were blanked in the coefficient display when the factor loading was less than 0.30 plained by items was 27.8%. Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 8 of 12 Dimensionality across the three extracted factors was version Participation Scale. Rasch item maps traditionally analyzed using Rasch model analysis to test and confirm show the distribution of item difficulties with the easiest a unidimensional interval measurement scale. When the items at the bottom and the most difficult items at the items were grouped according to the three factors as ex- top. For the Participation Scale, the higher the score, the tracted from the exploratory factor analysis, they were higher the participation restriction is. A successful out- found to be unidimensional. For Factor 1, the eigenvalue come of CBR means a high level of participation. For for unexplained variance in the first contrast was 1.8 judging enhanced participation, the interpretation of the (12.4%), and the raw variance explained by items was results of the item–person map is reversed. In this case, 27.2%. For Factor 2, the unexplained variance in the first the easiest items were at the top and the most difficult contrast was 1.7 (10.8%), and the raw variance explained items were at the bottom. The item hierarchy was re- by items was 26.7%. For Factor 3, the unexplained vari- vealed clearly in the construct key map (Fig. 1). The least ance in the first contrast was 1.7. difficult items were items 17 (comfortable meeting new Integrating the findings from testing of dimensionality, people), 8 (have respect in the community), and 10 (have it is concluded that the Participation Scale is a opportunity to enter into and maintain long-term rela- multi-dimensional scale. If a unidimensional interval tionships). The most difficult items were items 14 (do scale is required for measurement purposes, it is sug- household work), 7 (being as socially active as peers), and gested to use this 16-item scale. This 16-item scale was 11 (visit other people in the community). then used in this study for determining item difficulty and person hierarchy. Scoring category structure The scoring structure of the Participation Scale was satisfactory because the average Item hierarchy The measurement scale must be unidi- measures increased monotonically, which indicated that mensional to determine the item difficulty and the abil- on average those with higher ability endorsed the higher ity of people. Accordingly, the item misfit order showed category . Table 6 shows that the average measures that items 1 and 3 were misfit. After these two items increased monotonically across the rating scales, which were removed from the scale, the 16-item scale was means that they functioned as expected. Moreover, fit found to be unidimensional, and the examination of statistics show that the outfit mean squares of every cat- item difficulty and ability of people through Rasch model egory was less than 2, meaning that no particular cat- analysis was then legitimate. An item–person map was egory introduced noise into the measurement process therefore computed based on the 16-item Chinese (Table 6). Fig. 1 Item hierarchy of the 16-item Participation Scale as shown in the construct key map Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 9 of 12 EMIC stigma scale than 2 in the PCAR. A previous study on the validation of the Participation Scale showed that the Participation Internal validity For the EMIC stigma scale, 245 per- Scale should be conceived as a two-factor model that sons were included in the analysis. The person reliability consists of “work-related participation” (three items) and and item reliability were 0.74 and 0.90, respectively. The “general participation” (15 items) . Our current study scale was found to be reliable. The person and item also indicates that work-related participation should be fit were both confirmed from the summary statistics. regarded as a distinct factor. Factor 3, which was ex- For person fit, the infit and outfit MNSQs were 1.09 tracted using exploratory factor analysis, consisted of all and 1.11, whereas the infit and outfit ZSTDs were three items related to work and gainful employment. both − 0.1. For item fit, the infit and outfit MNSQs With exploratory factor analysis, three factors were ex- were 1.01 and 1.11, whereas the infit and outfit tracted from the Participation Scale. The extracted fac- ZSTDs were − 0.3 and 0.3, respectively. The Cron- tors encompass the principal constructs related to bach’salpha was0.9,and theitemseparationwas disability and participation. The three factors can be de- 3.04. scribed as (1) activity participation, (2) social engage- ment, and (3) work-related participation. Factor 1 is Dimensionality The results of the principal component activity participation, which refers to the execution of analysis showed that the eigenvalue of the unexplained physical and social activities , and is mainly variance in the first contrast was 1.9 (7.6%). The raw performance-oriented participation . Factor 2 is so- variance explained by items was 25.7%, which showed cial engagement, which is togetherness-oriented partici- that the EMIC stigma scale was unidimensional. pation  that focuses on performing meaningful social roles . Factor 3 is work-related participation. Item hierarchy As for the Participation Scale, the inter- The item difficulties and abilities of people with phys- pretation of the item difficulty is reversed. The higher ical disabilities were revealed by the Rasch model ana- the score of the EMIC, the higher the level of perceived lysis. Item 14 (doing household work) was found to be stigma. The most difficult item was Item 11 Problem in the most difficult for people with physical disabilities. getting married and marriage. The easiest item was Item Items 7 (being as socially active as peers) and 11 (visiting 2 Discuss the problem with others (Fig. 2). others in the community) were ranked high in terms of difficulty for people with physical disabilities. These Scoring category structure The scoring structure of the three items are related to physical activity (doing house- EMIC stigma scale was satisfactory because the average hold work) and physical mobility (visiting others). This measures of the four categories increased monotonically finding supports the construct validity of the scale be- (Table 7). The fit statistics confirmed that the category cause people with physical disabilities have different de- function was good because the outfit MNSQs for all cat- grees of sensori-motor impairment that may limit their egories were less than 2. mobility and performance in physical activities . Fur- thermore the individual experience of shame associated Discussion with physical disabilities may hinder a person from being Participation scale as socially active as peers (Item 7). Shame is regarded as The results of the Rasch model analysis and the explora- a strong emotion in Chinese culture. A person experien- tory factor analysis were complementary, which helped cing shame may feel that he or she has a stain that any- provide a comprehensive perspective on the construct one around them can see . Disability is associated validity of the Participation Scale. with shame in Chinese communities, and the stigmatiz- The results of the Rasch model analysis revealed that ing attitude is obvious . Therefore, people with phys- the Participation Scale is not unidimensional because ical disabilities perceive that being as socially active as the unexplained variance in the first contrast was greater their peers is difficult. Table 6 Category structure of the Participation Scale Category label Observed count Observed count % Observed average Infit MNSQ Outfit MNSQ Threshold 0 1629 41 −15.06 1.25 1.14 None 1 891 22 −8.46 0.58 0.76 −6.60 2 652 16 −3.10 0.76 0.85 −2.14 3 490 12 0.44 0.81 0.87 0.97 4 1 0 0.97 0.97 0.69 62.56 5 337 8 3.40 1.04 1.35 −54.80 Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 10 of 12 Fig. 2 Item hierarchy of the EMIC stigma scale as shown in the construct key map EMIC stigma scale was examined with the aid of exploratory factor ana- The Rasch model analysis and reliability analysis consist- lysis with SPSS. The combined results of Rasch ana- ently showed that the EMIC stigma scale had a high in- lysis and SPSS exploratory factor analysis showed that ternal consistency with a Cronbach’s alpha of 0.9. the EMIC stigma scale is a unidimensional measure Generally, the internal consistency of an instrument is of perceived self-stigma. The reason why Factor 1 strong when its Cronbach’s alpha is higher than 0.70 . stood apart from Factor 2 can be understood in the The detailed statistics of the Rasch model analysis pro- context of culture-bound syndromes, which stress the vided further corroborative support. The general princi- role played by culture in shaping the understanding ples of evaluating the Rasch model are to investigate of illness and health-related issues and places a heavy outfit before infit, and to investigate the MNSQ before emphasis on the relativity of health and illness across the ZSTD . The outfit MNSQ was 1.11 for both per- cultures . People with disabilities in Chinese soci- son fit and item fit. The outfit MNSQ measures the size ety are particularly vulnerable to stigmatization be- of the distortion of the outliers within the measurement cause it is believed that they bring bad luck to the system with expected values of 1 . If the MNSQ re- family and are being punished for immoral behaviors mains between 0.5–1.5, it is evaluated as a productive prior to their disability . Among an adult sample measurement. The evaluation of the EMIC stigma scale in Hong Kong, those with visible disabilities scored and the Participation Scale as two distinct productive significantly lower in self-concept than those without measurements was affirmed by the ZSTD outfit of per- visible disabilities . son fit and item fit with an expected value of 0; the The EMIC evolved from Kleinman’s pioneering ZSTD aims to test a hypothesis of whether the data fit work on an explanatory model of illness that not only the model perfectly . embraces the integrity and complexity of cultural psych- The reliability analysis should be studied in conjunc- iatry and medical anthropology but also recognizes the tion with the construct validity. The construct validity necessity of incorporating an interaction between “emic” Table 7 Category structure of the EMIC Category label Observed count Observed count % Observed average Infit MNSQ Outfit MNSQ Threshold 0 1911 52 −13.09 0.97 1.01 None 1 754 21 −5.56 0.82 0.82 −1.56 2 524 14 −1.44 0.94 1.01 −0.15 3 484 13 3.08 1.10 1.45 1.70 Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 11 of 12 (an understanding of an illness within a cultural context) Authors’ contributions EC contributed to develop the theoretical framework, research design, and “etic” (medical professionals’ understanding of an ill- coordinate data collection, preform data analysis and interpret results, as well ness without a strict adherence to cultural beliefs), which as the writing of the manuscript. GL participated in the EFA analysis and was first introduced by Pike . interpretation of results and drafting of the manuscript. Both authors read and approved the final manuscript. This study has limitations. A gold standard for meas- uring participation restriction and self-stigma has not Ethics approval and consent to participate been set. Both the EMIC stigma scale and the Participa- Ethical approval from the Committee on the Use of Human and Animal tion Scale are feasible and robust in administration and Subjects in Teaching and Research of Tung Wah College was obtained measurement. This validation study confirms that the (HASC1415H04). All participants were given informed consent for participation in this study. translated versions of both the EMIC stigma scale and the Participation Scale can effectively measure the level of participation and self-stigma for people with physical Publisher’sNote disabilities. However, the interrater and test-retest reli- Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ability have not been tested in this study. The generalizability of this study may be affected by the un- Author details even age distribution of the participants. Furthermore, Department of Special Education and Counseling, The Education University of Hong Kong, 10 Lo Ping Road, Tai Po, New Territories, Hong Kong. School caution should be exercised when interpreting the total of Medical and Health Sciences, Tung Wah College, 31 Wylie Road, score of the Chinese version of the Participation Scale. 3 Homantin, Kowloon, Hong Kong. School of Arts and Humanities, Tung Wah The original version’s cutoff score could not be used College, 31 Wylie Road, Homantin, Kowloon, Hong Kong. here because the standards are based on other popula- Received: 4 August 2017 Accepted: 21 May 2018 tions. This problem is compounded by the issue of the Chinese version’s dimensionality; therefore, further study is required if a norm or a cutoff score is required to dif- References ferentiate grades of participation restriction. 1. Aziz AA. Rasch model fundamentals: scale construct and measurement structure. Kuala Lumpur: Perpustakaan Negara Malaysia; 2010. 2. Bedford O, Hwang K. Guilt and shame in chinese culture: a cross-cultural Conclusion framework from the perspective of morality and identity. J Theory Soc Behav. 2003;33(2):127–44. This study translated and validated the two scales for use 3. Bond, T., & Fox, C. M. (2015). Applying the rasch model: Fundamental with people with physical disabilities in Chinese commu- measurement in the human sciences Routledge. nities. Results of the validation showed that both the Par- 4. Chan HH, Wong ET, Yeung CK. Psychosocial perception of adults with onychomycosis: a blinded, controlled comparison of 1,017 adult Hong Kong ticipation Scale and the EMIC stigma scale were valid and residents with or without onychomycosis. BioPsychoSocial Med. 2014;8(1):15. reliable. Although the interrater and test–retest reliability 5. Chen W, McLeod L, Coles T. Rasch first? Factor first? Value Health. 2014; were not tested, this study sufficiently tested the internal 17(7):A569. 6. Cohen, J. (1988). Lawrence Earlbaum Associates. Hillsdale, 20–26. reliability and construct validity of the Chinese Participa- 7. Costello AB, Osborne JW. Best practices in exploratory factor analysis: four tion Scale and the EMIC stigma scale. The translated recommendations for getting the most from your analysis. Pract Assess Res scales enable further development of the evidence-based Eval. 2005;10(7):1–9. 8. Daniels MJ, Rodgers EBD, Wiggins BP. “Travel tales”: an interpretive analysis practice of CBR because the effect of participation restric- of constraints and negotiations to pleasure travel as experienced by tion and self-perceived stigma on people with disabilities persons with physical disabilities. Tour Manag. 2005;26(6):919–30. can be accurately measured and documented. 9. Deepak S, Sharma M. A participatory evaluation of community-based rehabilitation programme in north Central Vietnam. Disabil Rehabil. 2001; Acknowledgements 23(8):352–8. The research team would like to thank all participants in their contribution in 10. Fabrigar LR, Wegener DT, MacCallum RC, Strahan EJ. Evaluating the use of this study. exploratory factor analysis in psychological research. Psychol Methods. 1999; 4(3):272. Completing interests 11. Field A. Discovering statistics using IBM SPSS statistics: London: Sage; 2013. The authors declare that they have no competing interests. 12. Franchignoni F, Giordano A, Sartorio F, Vercelli S, Pascariello B, Ferriero G. Suggestions for refinement of the disabilities of the arm, shoulder Funding and hand outcome measure (DASH): a factor analysis and rasch The design of the study and data collection was supported by the validation study. Arch Phys Med Rehabil. 2010;91(9):1370–7. departmental research grant of the Tung Wah College (2014–00-74- 13. Furr M. Scale construction and psychometrics for social and personality RGC140301). The analysis, interpretation of data and the writing of psychology. London: SAGE Publications Ltd; 2011. manuscript was supported by the Research Support Scheme 2016/2017 of 14. Glass TA, De L, Mendes CF, Bassuk SS, Berkman LF. Social engagement and the Department of Special Education and Counselling at the Education depressive symptoms in late life: longitudinal findings. J Aging Health. 2006; University of Hong Kong. 18(4):604–28. 15. Goffman E. Stigma: notes on the management of spoiled identity. New Availability of data and materials York: Simon and Schuster; 2009. The datasets generated and/or analysed during the current study are not 16. Imms C, Adair B, Keen D, Ullenhag A, Rosenbaum P, Granlund M. publicly available because consent and approval was not obtained for ‘Participation’: a systematic review of language, definitions, and constructs sharing of subject data from participants or the Committee on the Use of used in intervention research with children with disabilities. Dev Med Child Human and Animal Subjects in Teaching and Research. Neurol. 2016;58(1):29–38. Chung and Lam Health and Quality of Life Outcomes (2018) 16:105 Page 12 of 12 17. International Federation of Anti-Leprosy Association. (2011). Guidelines to reduce stigma. Retrieved from file:///C:/users/chunge/downloads/ilep- stigma-guidelines-2-how-to-assess-health-related-stigma.Pdf. 18. Ismail R, Noh M. A tale of two construct validation analysis: Rasch model and exploratory factor analysis approach for three-factor eating questionnaire (TFEQ-R21) among malaysian male workers. Med J Malaysia. 2015;70(3):169–76. 19. Kleinman A. Patients and healers in the context of culture: An exploration of the borderland between anthropology, medicine, and psychiatry. Berkeley: Univ of California Press; 1980. 20. Koo L. A journey into the cultural aspects of health and ill-health in chinese society in Hong Kong: the importance of health and preventive medicine in chinese society. J Comm Dent Health. 1989;11:51–8. 21. Lebel S, Castonguay M, Mackness G, Irish J, Bezjak A, Devins GM. The psychosocial impact of stigma in people with head and neck or lung cancer. Psycho-Oncology. 2013;22(1):140–52. 22. Lim K, Tan C. Epilepsy stigma in asia: the meaning and impact of stigma. Neurol Asia. 2014;19(1):1–10. 23. Linacre, J. (2017). A user’s guide to winsteps ministeps rasch-model computer programs: Program manual 3.93. 0. Retrieved from http://www. winsteps.com/manuals.htm 24. Mauro V, Biggeri M, Grilli L. Does community-based rehabilitation enhance the multidimensional well-being of deprived persons with disabilities? A multilevel impact evaluation. World Dev. 2015;76:190–202. 25. Montesinos AH,Rapp M,Temur-Erman S,Heinz A,HegerlU,Schouler- Ocak M. The influence of stigma on depression, overall psychological distress, and somatization among female turkish migrants. Eur Psychiatry. 2012;27:S22–6. 26. Peters RM, Van Brakel WH, Zweekhorst MB, Damayanti R, Bunders JF. The cultural validation of two scales to assess social stigma in leprosy. PLoS Negl Trop Dis. 2014;8(11):e3274. 27. Pike KL. Language in relation to a unified theory of the structure of human behavior. Walter de Gruyter GmbH & co KG; 1967. 28. Portney LG, Watkins MP. Foundations of clinical research: Applications to practice. NJ: Prentice Hall Upper Saddle River; 2000. 29. Radomski MV, Latham CAT. Occupational therapy for physical dysfunction. Philadelphia: Lippincott Williams & Wilkins; 2008. 30. Rensen C, Bandyopadhyay S, Gopal PK, Van Brakel WH. Measuring leprosy- related stigma–a pilot study to validate a toolkit of instruments. Disabil Rehabil. 2011;33(9):711–9. 31. Schumacker R, Linacre J. Factor analysis and rasch analysis. Rasch Meas Trans. 1996;9(4):470. 32. Stevelink S, Hoekstra T, Nardi S, van der Zee C, Banstola N, Premkumar R, van Brakel W. Development and structural validation of a shortened version of the participation scale. Disabil Rehabil. 2012;34(19):1596–607. 33. Tam S, Chan MH, Lam HW, Lam LH. Comparing the self-concepts of Hong Kong chinese adults with visible and not visible physical disability. J Psychol. 2003;137(4):363–72. 34. Tsang HW, Tam PK, Chan F, Cheung W. Stigmatizing attitudes towards individuals with mental illness in Hong Kong: implications for their recovery. J Commun Psychol. 2003;31(4):383–96. 35. Rockville MD. Mental health: A report of the Surgeon General. US Department of Health and Human Services; 1999. 36. Van Brakel WH, Anderson AM, Mutatkar R, Bakirtzief Z, Nicholls PG, Raju M, Das-Pattanayak RK. The participation scale: measuring a key concept in public health. Disabil Rehabil. 2006;28(4):193–203. 37. Van Brakel, W. H., Sihombing, B., Djarir, H., Beise, K., Kusumawardhani, L., Yulihane, R.,. .. Wilder-Smith, A. (2012). Disability in people affected by leprosy: the role of impairment, activity, social participation, stigma and discrimination. Glob Health Action, 5(1), 18394. 38. Van de Velde D, Bracke P, Van Hove G, Josephsson S, Vanderstraeten G. Perceived participation, experiences from persons with spinal cord injury in their transition period from hospital to home. Int J Rehabil Res. 2010;33(4):346–55. 39. WHO, I., & UNESCO, I. (2010). Community-based rehabilitation: CBR guidelines. 40. World Health Organization. International classification of functioning, disability and health: ICF. Geneva: World Health Organization; 2001.
Health and Quality of Life Outcomes
– Springer Journals
Published: May 29, 2018