Translation and cognitive testing of personal well-being questions into Sylheti and Urdu

This summary presents the findings of a pilot research project commissioned by the Office for National Statistics (ONS) and carried out by the National Centre for Social Research (NatCen). The project involved translation of the ONS personal well-being questions into two minority languages spoken in the UK, Urdu and Sylheti, and subsequent cognitive testing of the translated questions.

An important goal was to understand more about the linguistic and cultural relevance of the personal well-being questions to people living in the UK who may have a limited understanding of the English language. The research also sought to establish methods for appropriate translation and understanding of cultural issues that may shape responses and implications for use of the questions both in translation and in English. The study was conducted in two main stages:

• translation workshops

• cognitive testing of the translations

The full report (PDF, 1.3MB) summarises the process and findings of each stage and how they led to the final recommended translations for use with monolingual speakers of Urdu and Sylheti in the UK. The translations may also be used as an aid in conducting interviews with bilingual speakers of English, and Urdu and Sylheti, respectively. The project highlighted a range of issues to be considered in relation to different cultural and linguistic interpretations of the questions, which may have wider relevance to interpretations of findings about personal well-being in the UK.

The personal well-being questions are part of a wider programme in which Office for National Statistics (ONS) regularly monitors changes in 10 areas of life that people in the UK said mattered most to their well-being in a National Debate held in 2010. Since 2011, ONS and other government departments have included the well-being questions on many national surveys and national estimates of personal well-being are now updated quarterly.

The questions have been designated as National Statistics and are considered a harmonised standard, recommended for use on surveys to enable comparisons between different data sources. The questions are as follows:

I would like to ask you four questions about your feelings on aspects of your life. There are no right or wrong answers. For each of these questions I’d like you to give an answer on a scale of nought to 10, where nought is ‘not at all’ and 10 is ‘completely’.

• Overall, how satisfied are you with your life nowadays?

• Overall, to what extent do you feel the things you do in your life are worthwhile?

• Overall, how happy did you feel yesterday?

• Overall, how anxious did you feel yesterday?

Urdu and Sylheti were the languages selected for this pilot work because they were identified in a 2011 Census detailed analysis as among the most common languages spoken by people in the UK with a limited understanding of the English language. Monolingual speakers of these languages in the UK have also been identified as being at risk of lower general health and potentially also lower well-being. Urdu is the national language of Pakistan. Although it has a standardised spelling system, spoken Urdu varies depending on the geographical origins of its speakers. In the UK, Urdu is the main language of 0.5% of the population.

Sylheti was originally spoken in parts of Bangladesh and India, however, Bengali is now more commonly spoken in these areas. The Nagri Siloti script, which was used for writing Sylheti, has fallen out of use and there is no longer a formal written script. In the UK, Bengali (through the Sylheti and Chatgaya dialects) is the main language of 0.4% of the population.

The research programme for the translation and testing of the personal well-being questions was carried out in four iterative stages:

• initial translations

• translation workshops

• revised translations and cognitive interviews

• final translations

The following section will discuss each of these in more detail.

Initial translations

Two initial translations for each language were prepared. For Urdu, one translation was carried out by a translation agency and the other was the translation currently used by the Civil Service People Survey, which is already being used to collect personal well-being data from Urdu speakers. For Sylheti, a written translation in Bengali was done by the translation agency, as well as a male and female audio file version in Sylheti.

Translation workshops

These were attended by two bilingual Office for National Statistics (ONS) survey interviewers for each language, bilingual National Centre for Social Research (NatCen) researchers, questionnaire development and cognitive testing experts from NatCen, and ONS personal well-being experts. The workshops were used to discuss the initial translations, understand the cultural context, and amend the translated questions accordingly. Several issues needed to be considered including:

• specific words and phrases that best captured the meaning of the original questions in English

• cultural issues that might influence interpretations of the questions and responses to them

• whether questions asked verbally or in writing would be best understood

The workshops highlighted several issues with the initial translations of the personal well-being questions including:

• formality of the language used
• potential difficulties with use of response scales
• comprehension of script used for show cards
• cultural and religious beliefs requiring respondents to be matched with interviewers of the same sex and potentially also affecting question responses

Revised translations and cognitive interviews

New translations were developed based on the workshop findings. Four bilingual interviewers were trained in cognitive testing methods and were asked to conduct cognitive interviews with five monolingual speakers of their language using a topic guide developed by the research team. The exact wording of the questions was discussed in the interviews with monolingual participants and where needed, alternative words and phrases were sought.

The cognitive testing aimed to assess:

• respondents’ understanding of the terms used to describe the personal well-being measures

• how comfortable they felt answering the questions

• their understanding of the scale used and interpretations of time period covered in the questions

The interviews were audio recorded with the participants’ permission and interviewers reported the findings back to the research team in the form of detailed notes for each interview. Any issues that arose were compared with the results of previous cognitive testing of the questions in English. Where the same issue emerged both in the English and the Urdu or Sylheti testing, this was deemed to be rooted in the source question, rather than the translation or distinctive cultural issues.

Final translations

All interviews provided feedback from the testing and suggested final wordings for the questions. The research team summarised and analysed the detailed notes and made final recommendations for the wording of the questions in Urdu and Sylheti. A final version of the translations was provided, including a male and female voice audio file for each language, transliterations in the Latin script of these final translations and a written translation in Urdu (PDF, 114.7KB), Bengali (PDF, 19.3KB) and Nagri Siloti (PDF, 19.2KB) script respectively; these are now all available through NatCen.

This pilot research programme has shown that both the workshops and the cognitive testing stages of the process were vital to the project. There were many cultural and linguistic issues that needed to be considered when translating the personal well-being questions into Urdu and Sylheti. Linguistic considerations included:

• the fact that Sylheti is a spoken language with no official written form

• not all Urdu speakers are comfortable using the formal version of the language

• written versions of the questions in both languages may be hard to use

• there are a multitude of dialects for both languages, which are more or less mutually understandable

Apart from cultural and religious considerations that must be considered, there are also issues relating to literacy, numeracy and education levels, gender of interviewer and respondent, and lack of familiarity with survey questions and response scales.

For these reasons, it is important to allow enough time at each stage of the project for discussion, testing and refinement to arrive at the best possible outcome. The final outputs of the project include a methodological report providing a template for future translations of the personal well-being questions as well as both written and audio translations of the personal well-being questions in Urdu and Sylheti. These can be used in interviewing monolingual speakers of these languages in the UK as well as with those who have some proficiency in English, but would prefer to answer the questions in their native languages.

The full report, Translation and cognitive testing of the harmonised personal well-being questions (PDF, 1.3MB) is available. This summarises the process and findings of each stage and how they led to the final recommended translations for use with monolingual speakers of Urdu and Sylheti in the UK.

All translations, transcripts, transliterations and recordings of the Urdu and Sylheti files are also available.