— Resources for Researchers

Surveys

Adapting and Developing Measures, Including Formative Studies

There are two situations relevant to IPEN investigators when formative work can be very useful or essential in the measurement development process. The first situation is when you want to modify/adapt an existing measure to better fit the environment or population under study. The second situation is when you want to develop a completely new measure. In this section we provide guidance and examples for both situations

Adapting existing measures

This is an essential component of IPEN, because virtually every country needs to translate and adapt the core IPEN measures. The critical principle of adaptation is to retain the core items, then add items as needed to reflect the local environment or culture. Core items need to be retained even if they do not exist in your country (eg, cul-de-sacs), because absence of an attribute in a country is an important data point for international comparisons. As explained below, the goal of translation is to communicate the meaning of the question, not to make a literal translation. Ideally, draft translations will be pre-tested with people not involved with the study, and it is essential to conduct a back-translation and have it certified by IPEN staff.

As part of the translation/adaptation process we recommend conducting formative research. If you are adapting a psychosocial measure, it would be useful to get input from 1 or 2 content experts who could guide you on ensuring the measure is relevant to the culture and the population. Even if measures of the psychosocial construct (eg, social support for physical activity) are available in the native language, to provide comparability with IPEN, the measure used in IPEN should be translated. Pre-test the measure with people not involved in the study to make sure the meaning is clear.

For environmental measures, like the NEWS or an audit/observation tool, it is useful to do a small formative study to get input on (a) communicating the core items clearly and (b) adding items to reflect the local environment. Input should be sought from people who are like participants and from experts. We recommend starting with a panel of experts drawn from fields with expertise related to the content of NEWS. These fields include city planning, transportation, parks and recreation, geography, public health, landscape architecture, and criminology. You may not be able to include representatives of all these fields, and there may be others you want to include. You would discuss questions, such as “How can we clearly communicate the meaning of the current items?”; “What other environmental barriers to physical activity are important in our country?”; “What other environmental facilitators to physical activity are important in our country?”; and “Are there any special circumstances or variations by region that need to be included in the measure?” These discussions will result in candidate items to add to the core measure. It would be useful to ask similar questions of a group of potential participants to refine these ideas. The investigators will make final decisions about which new items to include in the final survey.

An excellent example of an adaptation process is the ALPHA project, which modified the NEWS to reflect common and relevant European environmental attributes, such as bicycle facilities and pedestrian streets. However, they modified the meaning of the core items and deleted some, which we do not recommend. Read paper.

Latin American investigators incorporated many of the new ALPHA items, plus some additional questions, into a Latin American version of the NEWS.

For example, the PLACE study in Australia investigated environmental issues that differed from the U.S. Read Paper

Dr. Oyeyemi led a group that developed an adaptation of NEWS for Sub-Saharan Africa. Read paper.

Dr. Adlakha developed an adaptation of NEWS for India. Read paper.

Developing new environmental measures

New measures are required for the advancement of science, and we encourage investigators around the world to develop new environmental measures for specific contexts or populations. Do not forget the principle of international comparability that requires a core set of measures that are as similar as possible across countries. In some cases, new measures are needed. For example, most built environment measures were developed for urban and suburban settings and do not seem relevant for rural areas. Thus, a group of investigators developed a rural environment assessment.

Our US-based group wanted to conduct a study of built environments and physical activity in youth, but there were no measures available for youth. Therefore, we conducted a formative study, called Active Where?. The ActiveWhere? Study was designed to develop measures to understand how the physical environment impacts physical activity and eating behaviors among children and youth. This study started with individual interviews with children and parents ‘in situ’, allowing researchers to observe and interact with youth while using neighborhood and park environments. We thought informants would be able to give us more useful information if they were commenting on the environments around them, instead of sitting in a room trying to think of environmental barriers and facilitators. This formative work was designed to generate appropriate items of direct relevance to children and adolescents. We summarized the themes from the interviews and generated items for a wide range of scales. A set of quantitative survey measures was then developed to assess home, neighborhood, park, and school environments to be completed by parent-child dyads, parent-adolescent dyads, or adolescents.

A reliability study was conducted in 3 U.S. cities – San Diego, Cincinnati and Boston – which increased the generalizability of findings. Most of the measures performed well. The study resulted in self-report environmental measures that are relevant to youth of a very wide range of ages and supplement the adult instruments currently used in national and international research. Though several scales were newly developed, we found the NEWS-Abbreviated worked well for youth and we only needed to add a small number of items, several of them related to concern about crime.

Dr. Rosenberg led the paper describing the adaptation and evaluation of NEWS for Youth. The IPEN Adolescent studies used this version. Read paper.

Translation Guidelines

Careful translation of surveys in English to other languages is critical to assure comparability of measures.
Generally, we recommend the following translation/back-translation process:

Any translation of survey measures from English to another language should be done by someone who is fluent in writing, reading and spoken English. Your research team should decide how you will verify the quality of the translation (e.g. through independent reviews of the translation, pre-tests, etc.) After a survey has been translated, a second person should conduct the back-translation of all survey items from the translated language to English. This second person does not necessarily need to be a professional translator, but should be someone who is unfamiliar with the survey and again is fluent in writing, reading and spoken English.