The National Institutes for Mental Health in England provided a small grant to begin to embark on the process of cross-cultural validation of WEMWBS in England. After deliberation, we chose to focus on two distinct cultural groups: Muslims from Pakistan and the Chinese. We were influenced in this decision by a recent report from Scotland in which a review of the literature relating to concepts of mental well-being had been undertaken, focusing on these two groups (Newbiggging et al. 2008) which allowed our limited resources to be devoted to data gathering and analysis. At the same time, we recognized that both groups represented relatively distinct cultures, influenced by specific spiritual and philosophical traditions, which were therefore likely to reveal diversity in the conceptualization of mental well-being. However, neither group is homogeneous. While both have been established in the UK for many decades, they include recent and long-term immigrants who live in a variety of socioeconomic circumstances, university students, young people who have been born and educated in the UK, and elderly who socialize almost exclusively within their own community. The Chinese community in the UK includes a high proportion of students and encompasses Chinese from both Hong Kong and mainland China. While we were limited by the resources available for the project, to working with members of these minority groups who spoke English to complete the WEMWBS in English (the group most relevant from the point of view of large-scale population surveys with English questionnaires), We worked with two community workers who were able to translate for participants when this was requested or seemed appropriate.

Measures of positive mental health are in demand from diverse parts of the world. We have now given permission for WEMWBS to be used in Australia, Canada, the United States, Italy, Spain, Germany, France, the Netherlands, Belgium, Iceland, India, Pakistan, Malaysia, and South Africa. So far, two of these research groups have completed formal quantitative evaluations of WEMWBS. Both Paola Gremigni in Italy and Marie Wissing in South Africa (with the Setswana community) have given permission for their results to be included in this chapter. These studies add much to the current topic due to their valuable translations of WEMWBS into other languages and therefore begin to provide a picture of how the instrument might work with black African cultures.

We undertook both quantitative and qualitative investigations of WEMWBS in two minority ethnic English communities. We worked with one of the Primary Care Trusts (the organizations which manage primary care and public health services in England) and a local charity in Birmingham, a city with a high proportion of minority ethnic inhabitants and a high level of social inequalities. The Trust employed community workers for both of the communities we were interested in and gave invaluable help in designing the study. Neither of these communities had been known to be typically easy to access. Few people, especially in the Pakistani community, were familiar with or interested in university research, and many were reluctant to give personal details to strangers. In both groups and among Chinese men in particular, there is reluctance to discuss mental health. With the help of the community workers, we developed an approach that was tailored to each community and involved door-to-door assessment in streets with high proportions of Pakistani residents, most of whom knew each other having immigrated from one community in Pakistan; direct contact with people through fitness gyms, access to which was at that time being provided free of charge in deprived communities; and contact through taxi bases, sewing groups, youth groups, and from social network groups. The Chinese community members were approached by the community worker at places where they typically gathered on Sundays or at Chinese restaurants, supermarkets, travel agencies, hairdressers, local colleges, local housing associations, and by word of mouth. We offered £20 to focus group participants to cover their expenses and a mobile phone voucher worth £2.00 to those who completed the questionnaires. For the quantitative evaluation, we were able to supplement the data we had collected with data derived from a representative population survey undertaken in Coventry, a smaller, multiculturally deprived city close to Birmingham. Public health practitioners in this locality had undertaken a general population survey, which included WEMWBS alongside 44 other health-related questions, in a survey that took 15 min to complete. The data collection team used a quota sampling method to achieve a representative population and undertook interviews on doorsteps and on the street using computer-assisted technology. Overall, the respondents were representative of the population of the city in terms of demographic characteristics, with 44 % engaged in paid work, nearly half of which were male (48 %), and with a good spread across age range. Out of a total population of 3,750 (from a sampling frame of 8,500—a response rate of 44 %), data were collected on 43 Chinese and 94 Pakistani respondents, and these were included in the quantitative evaluation.

For the Birmingham study, a booklet containing WEMWBS together with two comparator scales—the General Health Questionnaire (GHQ-12; Goldberg et al. 1997) and the WHO-5 Well-Being Index (Bech 2004) with demographic details—was designed by the research team at Warwick University and approved by representatives of the Chinese and Pakistani community in Birmingham, by public health practitioners working in the Primary Care Trust, and by the University Ethics Committee. The latter involved considerable negotiation because committee members felt that in order to get effective consent, participants should be offered a week between first contact and taking part in the study, an approach which was not considered practical or necessary by community workers. Letters of invitation for door-to-door and on-the-spot recruitment, information sheets, thank you cards, and consent forms for both the survey and the focus groups were also approved.

Community workers distributed 120 questionnaires in each community using a quota sampling method so that equal numbers of men and women were represented and so that the age mix reflected that of the UK population. Many people who were approached declined to complete the questionnaire. Most participants answered the questionnaire on the spot, but some preferred to take it away to complete at home. Of those who agreed to take a questionnaire, for the Pakistani group, 107 out of 120 (89%) questionnaires were returned, while for the Chinese group, 116 out of 120 (97%) were returned. The mean age of respondents was 48 (SD 9.0) for the Pakistani community, with a preponderance of young and middle-aged respondents. In the Chinese community, the mean age for men was 47 (SD 9.3), and for the women, the mean age was 50 (SD 11.2); there was a good spread across the age range. In both groups, even though just over half of the participants had been born in the UK, three quarters were less confident in English than they were in their first language. Half of those in the Pakistani community and two thirds of those in the Chinese community were engaged in paid work. Furthermore, all participants in the Pakistani group reported their religion to be Islam, and three quarters of the Chinese reported no religion.

Five age- and gender-specific focus groups were held in the Pakistani community: men aged 16–24 years, men aged 25–49 years, men aged 50–75 years, women aged 16–24 years, and women aged 25–49 years. Three age-specific groups were held in the Chinese community: ages 16–24 years, ages 25–49 years, and ages 50–75 years. Ten people were invited to each group. Participants were all able to speak English, with the exception of one older Chinese woman and three middle-aged Pakistani women for whom the community worker or other young women interpreted, respectively. We defined English speaking as the capacity to complete the questionnaire since this is how it would be defined in any population survey.

Rasch modeling was developed to investigate the psychometric properties of scales and instruments. Based initially on examination data, it assumed a hierarchy among items with questions being more or less difficult to answer. It also assumed that each item retained its hierarchical order in all cases (or students). Instruments which meet Rasch criteria have the important property of numerical scaling, so the difference between 5 and 10 can be assumed to the same as the difference between 20 and 25. These assumptions are implicitly made about most instruments used to measure mental health, but they are not necessarily justifiable from a mathematical point of view. Lack of fit to the Rasch model does not invalidate a scale, but it does violate many of the assumptions of the statistical tests which are used to assess their significance and the implicit assumptions about the scores. Such scales can show respondents to be higher or lower on a scale than other respondents but are not clear as to how much better or worse they are.