• Home
• |
• Contact Us
• |
• User Guide

Translation Tools Guidelines

VI. Translation: Translation Tools

Webpage last modified: 2008-Nov-12

Introduction

This section discusses tools to aid survey translation, including:

• Standard reference sources
  • Dictionaries and other reference books and hardcopy materials
  • Web-based reference materials
• Standard aids
  • Checklists
  • Listservers and newsgroups
  • Translator procedures
• Templates for the translation process and translation output
• Technological support, such as translator software
  • Translation Memory (TM)
  • Terminology tools
  • Concordances

Increasingly, large-scale international survey translation efforts combine source document production with that of translated versions. That is, the source text is entered into a management system which anticipates the needs and documentation of later production steps in other languages [1]. Our guidelines do not assume such a system; they do, however, include consideration of the technological components that would be available in an integrated document production and management system.

Tools and aids for translation can be provided by the translation project coordinator or can be a normal part of a translator’s own toolkit. Who provides what may vary by project. A project might, for example, require translators to provide their translations using project-specific software, as is the case with SHARE and PIAAC.

(See Appendix A for a list and description of various translation tools.)

Guidelines

1. Identify relevant materials, provide them to translators, and instruct translation team members on their use.
   Rationale

   The more relevant information that competent translators receive, the better they can meet the needs of the project.

   Procedural steps

   • Materials may include:
     • The website (intranet and/or internet) of the survey project providing background information and documentation of the project.
     • The entire questionnaire, even if only parts of it require translation. This enables translators to:
       • See the context in which the parts to be translated belong.
       • Plan for consistency.
     • Any available sections already translated that have been vetted for quality.
       • This contributes to consistency.
       • Material not yet vetted for quality may also be provided but must be used with great caution.
     • A bilingual glossary for any terms or phrases whose translation has already been established.
       • This helps to ensure compliance and consistency.
       • It also benefits the review and copy editing phases.
     • A style sheet guide, if relevant, detailing how to treat standard components of the text (e.g., strategies for emphasis in the source text or exclamation marks).
     • Tracking documents that list major recurring elements and their location.
       • Ideally these are produced automatically as part of a content management system and can be begun during development of the source questionnaire. In this case, project coordinators set the parameters for what matches should be included.
       • They may also be part of translation software.
       • In many cases, usually in modestly funded projects, they may be developed manually.
       • Create quality checklists to be incorporated into each country’s final copy editing effort. Include frequent or likely oversights in the list (e.g., "Check the order of answer categories").
   Lessons learned

   • If existing translated material that has not been vetted for quality is made available to translators for consideration in a new translation, coordinators must decide whether the translators will be able to accurately assess its quality. These issues may also arise when translators access "parallel texts" (e.g., texts from other surveys) in the target language. They might include very similar questions, such as frequently translated health questions, or offer translations for standard components, such as answer scales. Surveys need to be aware that existing translations may not be appropriate for their new purposes.
   • The purpose of various tools and procedures may not be self-evident to translators; the translation staff may need to be briefed regarding their purpose and implementation.
2. Provide translators with documentation tools and require the translators to use them.
   Rationale

   Providing thorough documentation of decisions, ongoing problems, and adaptations at each step of the translation process guides and enhances subsequent steps. Documentation tools and specifications can ensure that each participating unit provides systematic, comparable documentation.

   (Translation documentation may be part of the development of the source document if the project uses a text content management system.)

   Procedural steps

   • Clearly identify what requires translation and what does not.
     • Some work platforms allow the user to freeze sections that should not be translated.
   • Produce translation templates that align source text segments, target text fields, and comments fields (see Appendix B for (1) an extract from an ESS Round 4 template and (2) an example of how a merged review template based on such a translation template might look).
     • Questions, instructions, and answer scales are examples of "natural" source text segments.
     • Subdivisions at least to sentence level are often useful.
     • A simple MS Word or Excel table, produced manually, may suffice.
     • Translation software and content management systems may produce templates automatically.
   • Provide instructions for translators and any other users on how to use the templates and how to document. For example, clearly explain the kinds of information expected in the comments field.
   • Since individual team members fill their templates, it is advisable to hold meetings to merge template inputs. In this way, people can compare options, notes, or comments (see Translation). Final output from one phase can be passed on in a modified template to the next phase of work.
   Lessons learned

   • The following issues apply in particular to the manual production of templates:
     • The manual production of templates is labor-intensive and calls for care but, in many cases, may be the only option. As relevant, budget for the time and effort to produce translation templates manually. Involve at least two suitable people with adequate bilingual proficiency and proofreading skills for the final proof-reading effort (one reading out, the other checking).
     • Remember to check layout and format issues, not just wording.
     • Although, ideally, template production should begin after the source text is finalized, this may not always be feasible. If production starts prior to finalization, a tracking system for version control is essential to check modifications at either the source or target text levels.
     • Remember that copy and paste mistakes occur frequently. Technology (e.g., use of translation memory) may or may not increase their likelihood.
     • Working between different source versions of a question and different translated versions within or across languages can be complicated. A tracking system is essential to identify which elements should or do differ across versions.
     • A procedure and protocol for alerting locations or different teams to changes in either source documents or translation requirements will be needed. For example, in a centrally organized project, the source text may be modified after templates have been sent out to translating locations (countries) and countries need to be able to recognize unambiguously what needs to be changed and then incorporate these changes into their templates (or at least into their translations). The ESS alert system, as an example, is accurate but not particularly user-friendly; it has to be produced manually and changes have to be manually transferred from source version to source version.
3. Provide translators with appropriate task instructions and briefing (see Finding, Selecting, and Briefing Translators).
4. Consider networking translation teams within the project.

   This refers to different countries involved in one project confering with one another on translation challenges. Networking sometimes occurs unofficially. For example, some Scandinavian countries in the ISSP confer often on problems. Despite speaking and writing different languages, they find that, many times, there are similarities regarding the kinds of solutions open to them. In the ESS, on the other hand, countries sharing a language are officially encouraged to confer.

   Rationale

   Consultation within a language family can be helpful for all. It is equally possible that sharing across language families can also be of benefit, since some generic issues are shared by rather diverse languages and cultures. This is not a richly researched area in survey research, although past experience and some research suggest its usefulness. Harkness [2] found that a reasonably wide range of languages and cultures face similar translation challenges. Thus, information can be compiled to assist many different language translation efforts.

   Procedural steps

   • Decide whether collaboration is to be unofficial or official and whether or not it must be officially documented.
     • If it is to be documented, decide on the template and detail required.
     • Both official collaboration and official documentation help to unify practices across and within projects.
   • Set up a protocol and schedule for sharing experiences or solutions and documenting these.
   Lessons learned

   • The publication of collaborative benefits, procedures and successful outputs experienced within one group may inspire other groups that have not have considered such collaboration. For example, the ISSP Scandinavian experiences, in which countries speaking similar languages confer on problems faced in translation, might inspire other collaborations. While Asian languages, for instance, differ considerably from one another, their common differences from the source language out of which they are translating (often English) may be far greater.
   • If researchers fielding in different regional forms of a "shared" language do not collaborate, many differences across versions may result that could otherwise have been avoided.

© 2008 The authors of the Guidelines hold the copyright. Please contact us if you wish to\n publish any of this material in any form.

View Chapters

• Introduction
• Study Structure
• Tenders, Bids, & Contracts
• Ethical Considerations
• Sample Design
• Questionnaire Design
• Translation
•      Translation Assessment
•      Language Harmonization
•      Translators
•      Translation Budgeting
•      Translation Scheduling
•      Translation Tools
• Adaptation
• Survey Instrument Design
• Pretesting
• Interviewer Recruitment & Training
• Data Collection
• Harmonization of Data
• Data Processing
• Dissemination
• Assessing Quality
• Glossary
• Bibliography