Menu
 

You are on the Natcen site

Click here for Scotcen

natcen map

You are on the Natcen site

Click here for Scotcen

natcen map
 

Translating personal well-being questions into Sylheti and Urdu

Posted on 13 April 2018 by Catherine Fenton, Researcher .
Tags: Translation, Wellbeing, Cognitive Testing, ONS, QDT, Questionnaire Development and Testing Hub, methodology

How do you translate questions about complex concepts into different languages?

Over the past 10 years, well-being has gained increasing traction as a subject for the social sciences – but measuring this complex concept can be difficult.

In 2010, the Office for National Statistics (ONS) started their Measuring National Well-being programme to track the well-being of the UK population. A national consultation uncovered 10 domains of life which people in the UK said mattered most to their well-being. From here, the ONS chose forty-three indicators as measurements within the various domains.

Personal (or subjective) well-being, referring to people’s thoughts and feelings about their own quality of life, is one of the domains that ONS captures. It was the only area where new questions were developed as part of the programme, as opposed to using existing measures and data.

Currently, there are four personal well-being questions used, and these are included on more than 20 surveys in the UK, as well as a range of smaller scale evaluations.  

 

 Figure 1. ONS harmonised personal well-being questions

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 0 to 10, where 0 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?



Although the four ONS personal well-being questions have been extensively tested in English, they have never been tested in translation among groups that speak little English. As part of a wider interest in ensuring the inclusion of all members of society, ONS wanted to understand the best method for translation, and test any implications this might have for cultural and linguistic interpretations of the questions and the findings.

In previous studies, ONS has found that some speakers of Sylheti and Urdu might be less likely to speak English. These monolingual speakers may also be at risk of reporting lower ‘general health’ than the general UK population and potentially lower well-being.

We teamed up with ONS to translate the harmonised personal well-being questions (see box) into Sylheti and Urdu, and to ensure that the translations were linguistically and culturally appropriate.

 

What the study involved…

The research project functioned as a pilot study, allowing us to design the best method to translate and test these questions in other languages. Our study involved two primary stages to produce the translations in Urdu and Sylheti: translation workshops, and cognitive testing of the translated questions.

 

The main methodological challenges

The project involved several methodological challenges:

  • The languages themselves having many dialects;
  • The potential for some members of the population groups to have low levels of education and literacy;
  • There being no formal written form of Sylheti;
  • Participants being less likely to have any familiarity with research surveys.

 

How did the research team address these challenges?

During the workshops, members from the NatCen Questionnaire Design and Testing Hub and ONS well-being experts teamed up with native speakers of Urdu and Sylheti to discuss and adapt the initial translations of the four personal well-being questions.

Once the questions were finalised from the findings of the workshop, we tested them in cognitive interviews with monolingual speakers in the UK. These interviews aimed to assess the appropriateness of the translations, and to test whether the questions were working as originally intended. The final translated questions were developed following the workshops and cognitive testing, and are available in written form and as audio files.

 

The Project:

The research plan (as outlined in the figure below) revealed complexities relating to each language in the UK context, as well as cultural aspects of the Urdu and Sylheti-speaking communities. 

 

Figure 2. Research plan for the translation of the personal well-being measures 

ONS_wellbeing _1 (1) 

1.     Initial translations

Complexities in translating to Urdu

Complexities in translating to Sylheti

  • Not all of the Urdu speakers participating in the project were comfortable using formal Urdu, which might be associated with higher educational levels and social status
  • Many Urdu speakers use a variation or a dialect of Urdu during day-to-day interactions
  • Sylheti does not have an official written form, and is the umbrella language for different dialects

 

For these reasons, we considered which terms may be most appropriate for monolingual speakers of Urdu in the UK, who may not be comfortable reading, writing or speaking formal Urdu.

Our aim was to develop translations in an accessible form for all Urdu speakers –much as ‘plain English’ is used to ensure English speakers from a range of backgrounds can understand the questions.

The same principle was followed for the Sylheti translations. Additionally, as Sylheti is primarily a spoken rather than written language, we agreed that the questions should be oral rather than written, and we developed audio files for the preamble and each of the well-being questions. This would enable standardised administration of the questions.

With these points in mind, initial translations of the personal well-being questions were developed for both Urdu and Sylheti. Two different Urdu translations were also made in written form. For the Sylheti speakers, female and male audio files of the questions were produced and to supplement this, a written translation was also produced in Bengali.

 

2.     Translation workshops

The workshops aimed to assess the initial translations of Urdu and Sylheti in terms of appropriateness of the language used, clarity of terms, and concepts.

These workshops were crucial in identifying where refinements were needed in the translations, as well as cultural and linguistic issues to be explored further during cognitive testing. Our key findings and suggestions from the workshop included the following:

 

Findings

Recommendations

Both initial translations were too formal for monolingual Urdu or Sylheti speakers.

The questions were adjusted based on the feedback, and the new translations were circulated to all the panellists for review.

Some participants showed limited or no experience or understanding of the research process. Panellists thought it was likely that members from the Urdu or Sylheti community may not have taken part in a survey before, or used a numeric response scale.

It was agreed that the interviewers would record this issue if it arose during the cognitive interviews, and attempt to explore how participants may find a suitable comparison to answer the question. It was suggested that visual scales could be used instead of the 0 to 10 scale.

The suitability of using showcards with numerals in Urdu and Bengali script was raised, since they are different from the Latin numerals.

The preference was to be assessed during the cognitive interviews.

The standard introduction to the four well-being questions may not be sufficient for respondents who have never participated in a survey before.

Before the cognitive interview began, interviewers would spend extra time explaining the purpose and nature of large-scale surveys in an attempt to make participants feel more comfortable.

Potential cultural differences were raised – the importance of religious beliefs and their possible impacts on people’s responses to the personal well-being questions.

We decided to explore this during the cognitive interviews. It was agreed that the gender of the interviewers and respondents should be matched to avoid participants feeling uncomfortable and allow them to speak freely.

 

3.     What happened during the cognitive interviews?

Once the new translations of the personal well-being questions were agreed, we conducted ten cognitive interviews in each language with monolingual speakers. Translations were made into Urdu, and both female and male audio recordings into Sylheti. Interviewers were matched with respondents of the same gender, to ensure that participants were comfortable taking part in this study.

These interviews aimed to explore both participants’ understanding of the personal well-being questions, as well as their ease of participation in the overall research process. Additionally, the cognitive interviews intended to assess the cultural appropriateness of both the Urdu and Sylheti questions.

The cognitive interviews identified that the translations were generally well understood by participants. However, participants had some difficulty understanding:

  • The purpose of the personal well-being questions;
  • The use of the response scales;
  • The relationship between a quantified answer on a scale and everyday life experience. For instance, religious-minded participants might give a quantified answer about their life satisfaction – but qualify that by accepting their situation as given by a divine power.

 

4.     What was done to ensure the final translations were appropriate?

To address these issues, we provided more detail in the preamble to explain to participants the purpose of the study and to make them feel more at ease. For people who are not used to taking part in research projects, this is an important aspect: understanding why they are asked these questions and who will have access to their responses.

We suggested that a showcard should be used to encourage respondents to answer on a scale. Although doing so is not always intuitive for novice research participants, people get used to it quite quickly and eventually find it easy to use. A more detailed preamble and a visual showcard can thus ease the interviewers’ task of explaining the research process to the future participants.

We also found the women who took part in the cognitive interviews said they would not be as comfortable answering these questions and so forthcoming with their answers had the interviewer been male. It was also felt that it would not be appropriate for a woman to be interviewed by a man. We thus decided that it would be best to have audio files with both a male and a female voice, for both languages. The audio files could supplement the use of the written translation in Urdu, and could ensure that even participants with lower literacy skills are able to take part. And of course, they would allow the researchers to ask the questions in a standardised way, such that the responses can be compared to responses collected from the general UK population.

The translations in both Urdu and Sylheti were purposefully done using colloquial language and sentence structure, rather than choosing the more formal language. We believe that it will help future participants engage more with the research process and answer these questions truthfully.

 

What did we learn from this project?

There’s no doubt that the iterative nature of the project, with multiple stages for testing and refinement, contributed to a better outcome at the end.

Sufficient time was given to each stage of the research project to allow for discussion around the meaning and purpose of the personal well-being measures – as well as the cultural and religious characteristics of the communities involved, and how that might affect responses to the questions and interpretations of the findings.

If you’d like to learn more about the project, the full published report is available here, along with the written translations and audio files - you can also contact me at 

comments powered by Disqus
Blog filters
Year
  • 2018
  • 2017
  • 2016
  • 2015
  • 2014
  • 2013
  • 2012
  • 2011
  • 2010
Month
  • Jan
  • Feb
  • Mar
  • Apr
  • May
  • Jun
  • Jul
  • Aug
  • Sep
  • Oct
  • Nov
  • Dec
NatCen/ScotCen
Clear Filters

Subscribe to the RSS Feed: