Intro

A 24 hour-research to find, read and craft a plan for

How to adapt a scale/instrument from Language A to B? In our case, from English to Turkish.

Sources

A. World Health Organization (WHO)

B. Antunes (2012)

C. Secer (2015)

D. Widenfelt et al. (2005)

A) World Health Organization (WHO)

Here is the page. The document is titled “Process of translation and adaptation of instruments”

Main Points

The translated instrument should be equally natural and acceptable and should practically perform in the same way. The focus is on crosscultural and conceptual, rather than on linguistic/literal equivalence. Do not prioritize the word-to-word translation

Steps to follow

  1. Translation: 1 translator preferably a field expert whose mother tongue Turkish, but familiar with English-speaking culture.
  • aim at the conceptual translation, again, do not prioritize word-to-word
  • be simple
  • translate for layman, not for professionals
  • do not use technical terms, idioms or vernacular terms
  • items should not be offensive
  1. Expert panel: A bilingual team should be gathered by a designated editor-in-chief.
  • the number of experts may vary, but should include the original translator, field experts, an instrument-development expert.
  • the final product of this step is a complete translated version

  1. Back-Translation: an independent translator whose mother tongue is English and has no knowledge of the instrument. Compare back translated form to original form and examine conceptual equivalence, not word-to-word equivalence.

  2. Pre-testing and cognitive interviewing: Pre test the instrument on the target population

  • representative but preferably not from the main study participants.
  • Debrief ask respondents what they thought.
  • make changes if needed after debrief
  • write a report of the pretest process.
  1. Final version and documentation: Procedure should be traceable.

B) Antunes (2012) and Sousa & Rojjanasrirat (2010)

cite

Antunes B, Daveson B, Ramsenthaler C, Benalia H, Ferreira PL, Bausewein C and Higginson IJ; The Palliative care Outcome Scale Manual for cross-cultural adaptation and psychometric validation, London, March 2012 (http://pos-pal.org/Resources.php)

Sousa, V. D., & Rojjanasrirat, W. (2011). Translation, adaptation and validation of instruments or scales for use in cross‐cultural health care research: A clear and user‐friendly guideline. Journal of Evaluation in Clinical Practice, 17(2), 268-274. doi:10.1111/j.1365-2753.2010.01434.x

Steps to follow

  1. Conceptual definition: the research team follows;
  • review the literature to familiarize with concepts
  • identify key concepts item by item
  • use a focus group to ensure the concepts can be recognized by target population
  1. Forward Translation: Use three bilingual translators (Turkish mother tongue)
  • Forward Translator 1, should be familiar with concepts
  • Forward Translator 2, should NOT be familiar with concepts
  • The third translator compares above 2.
  1. (Blind) Backward Translation: Blinded to original scale, preferably 2 back translators whose mother tongue is English. Blindness is important to avoid information bias.
  • Backward Translator 1 should be familiar with concepts
  • Backward Translator 2 should NOT be familiar with concepts

  1. Expert Committee Review: compare two backward translation in terms of conceptual, semantic, experiental and content equivalence. Produce final format for pretesting

  2. Cognitive Debriefing: 2 interviewers recommended (a coordinator and an assistance). 5 - 8 participants might be sufficient.

  3. Proof Reading: each step should be evaluated

  4. Psychometric testing: Aim is to ensure validity and reliability. The sample size for this step depends on the complexity of model. 10 participants per item recommended but not a must. Recommended approaches;

  • internal consistency
  • Test-retest reliability
  • Construct validity

C) Seçer 2015

Dr. Seçer is the author of “psychological test development and adaptation process, SPSS and Lisrel exercises”.

Main points

  1. Language (translation) validity is a big deal.
  2. Scale adaptation is common.
  • advantage: saves time and effort by skipping ‘theory and item development’ process.
  • disadvantage: cultural disagreements. (i.e if Supervisor means the same in Turkey compared to USA?)
  • disadvantage: different factor structures, measurement invariance.

Steps to follow

  1. Identify the need, and the tool
  2. gather up a team for translation (at least 2 experts in Turkish, at least 6 in English.)
  3. gather up another team of field experts (at least 2)
  4. translate the tool (at least by 2 English experts)
  5. compare translated forms (at least 2 field experts, 2 Turkish experts and 2 English (cannot be those in step 4))
  • examine linguistic, cultural and construct differences.
  • try to use a Likert type form to evaluate language validity for each item
item English score Turkish
1 0-1-2-3-4-5
2 0-1-2-3-4-5
  1. do back translation (at least by 2 Eng. experts)
  2. compare back translated form to original form (use a bilingual sample of at least 30, administer forms in both languages. Check correlation, larger than .7?. If bilingual sample is not available, ask experts)
  3. pilot study
  • a sample of at least 50 to study reliability and validity
  • if the first pilot study reports misbehaving items, try to make them behave! and use another pilot study.
  1. Use the translated tool and study its validity and reliability.

D) Widenfelt et al 2005 (this is not a guideline)

Article named Translation and Cross-Cultural Adaptation of Assessment Instruments Used in Psychological Reseach with Children and Families reports authors’ experience on translating instruments from English to Dutch. Their aim is to summarize important points they faced. This is not a guideline.

  • word-to-word translation is risky. i.e. Chinese culture does not have an equivalent term for self-esteem. swearing does not have the same meaning in Thai. Saying I am a winner is not welcomed by Dutch kids.

  • Contact the original author, and make sure the tool was not adapted before.

  • have at least 3 translator for forward translation

  • use short and simple sentences, active voice. Use not, metaphors, vague words such as probably, two verbs at the same time.

  • Slight alterations might be defensible i.e translating the term bullying will probably require slight alteration.

  • if entirely new items are generated, its good practice to contact the original author.

  • Deleting items might be defensible, but replacing them with new ones is a good idea to keep the scoring range the same. i.e in Thailand special classes are not common, asking “is your child in a special class?” does not make sense.

  • if translators disagree on any item, different forms of the same item should be included in pretest.

  • after the pilot studies, conduct the field test and look for; skewness-kurtosis of items, Differential Item Functioning, Internal consistency and factor analyses.

Bonus, how to develop a scale

In their ‘mplus short courses topic 1’ Muthen and Muthen (2009) suggests following steps to develop an instrument;

  1. Pilot Study
  • Small n, EFA
  • Revise, delete, add items
  1. Pilot study 2
  • Small n, EFA
  • Formulate tentative CFA model
  1. Pilot study 3
  • Larger n, CFA
  • Test model from Pilot study 2 using random half of the sample
  • Revise into new CFA model
  • Cross-validate new CFA model using other half of data

  1. Large scale study, CFA

  2. Investigate other populations