The Linguist

16 The Linguist Vol/63 No/2 ciol.org.uk/thelinguist FEATURES during remote interpretation. We are more likely to notice when a patient signals confusion by frowning, makes a bewildered expression or turns to look directly at relatives in the room for help with medical decisions. We can cultivate human connections and build rapport, making patients feel more comfortable asking questions. We often hear statements from patients and colleagues such as "You're a lifesaver!" and "Can you just stay with me all day?" EMPLOYEE BENEFITS One benefit of working in-house is that I receive a regular, bi-weekly paycheck. When I worked full-time as a freelancer, I had to be very organised and keep track of the invoices I sent to various language agencies. I now work 40 hours a week in 8-hour shifts across 2 hospitals, with 2 paid 15-minute breaks and a 30-minute unpaid lunch break. The main employee benefits include health insurance plans (including vision and dental), retirement plans, accrued paid time off (vacation, sick time), gym membership and tuition discounts. There are also green spaces and meditation rooms for us to use. Communication plays a key role in our team. We help each other when changes in our daily schedules occur (e.g. cancellations, delays, additional time needed with a patient/ provider). It is a supportive environment and every day is different, which makes me want to come to work the next day! Special thanks to Allison deFreese and Glenda Carelhue for reviewing initial drafts. Better use and understanding of translation are key to improving global research, says Celine Garbutt Textbooks about how to design research surveys put great emphasis on the importance of using suitable language. Recommendations include "avoid ambiguity and bias" and "be exact and simple". There is also a need to ensure that language is 'faithful' to the theory behind each question. In other words, will the question elicit a response that is relevant to what is being 'measured'? Understanding the do's and don'ts of question development is essential, because answers from faulty surveys can distort the results of statistical analysis. For example, a leading question, 'On a scale of 1 (not really) to 5 (a lot), how much do you enjoy learning foreign languages?', is likely to produce a different response to the question 'How do you feel about the following statement on a scale of 1 (strongly disagree) to 5 (strongly agree): "I enjoy learning new words in a foreign language"?' Writing questions in one language is challenging, so what happens when surveys are carried out across several countries at once? There are various ways in which questionnaires are translated. In 'forward translation', one person (often the researcher or research assistant) translates the survey into the target language. This approach is common because it is considered time and cost efficient, but there are disadvantages: the quality of the translation depends entirely on the skill and ability of one person, and if the translation is poor, the quality of the data suffers. The 'forward-backward' approach aims to screen out errors which appear in the back translation. However, back-translation has major drawbacks as well. One recent study investigated the quality of translation in multilingual, multicultural and multiregional research and concluded that back-translation is not a hallmark of quality. 1 Dorothée Behr gives us an example of what can go wrong: in one case, the expression 'care services' was translated into German as Pflegedienste. This back-translated as 'care services', indicating a good translation. However, Pflegedienste refers to care of the ill or elderly, which in a survey about general childcare services, would be inappropriate. This type of error can only be detected by someone with contextual knowledge of the subject in both languages. A five-stage approach The failure of these two methods to deliver reliable translations led to the development of a third: the TRAPD model. TRAPD stands for five key phases in team translation: translation, revision, adjudication, pre-test and documentation. Established in the early 2000s, it is part of a broader set of guidelines for the development of multinational, multicultural and multiregional surveys. 2 The first and second phases require translation of questions by at least two accredited translators, followed by revision. Adjudication refers to the selection of one translated version. Pre-testing is then carried out to assess readability and to see if people Is research working? © PEXELS

• Cover