«Minimum Dietary Diversity for Women A Guide to Measurement Minimum Dietary Diversity for Women A Guide to Measurement Published by the Food and ...»
Steps for adapting the model questionnaire in Section 3 include: making a decision on the data collection approach and on how many of the optional categories to include; creating a basic translation of the MDD-W model questionnaire in the survey language(s); and reviewing and adjusting the translated enumerator instructions and introductory text using words and phrases that are easily understood by both the enumerators and the respondents. Following this, it will be necessary to adapt the foods listed in each row of the MDD-W model questionnaire to include seasonally and locally available foods commonly consumed, using local names where appropriate.
This preparatory work can be undertaken at a basic level or, if time and resources permit, may be expanded to include qualitative research on different population subgroups to ensure completeness of the food and beverage examples listed in each questionnaire row and to gather local recipes for mixed dishes. While this second level of adaption is highly recommended so that the list of foods consumed by survey populations will be as accurate as possible, it may not be feasible to carry out in all cases.
Decision on method for collecting the information The survey team will need to decide whether to follow the recommended method of soliciting an open recall of all foods and beverages consumed from when the respondent awoke the previous day through the day and night, or whether to use the list-based method of inquiring about each food group (see Section 1, Methodological approaches to measurement of food group diversity, and Appendix 3 for discussion of the list-based method). This decision will determine the type of enumerator instructions and the opening text to read to respondents, as well as enumerator training content.
27 Section 4 Preparing the MDD-W questionnaire
Content of the survey-specific MDD-W questionnaire The model questionnaire provided in Section 3 includes 22 mutually exclusive food groups and categories, 14 of which will be aggregated to create the MDD-W 10 food group indicator. Among the remaining eight categories, at a minimum the “Condiments and seasonings” and “Other beverages and food” categories (rows “U” and “V” on the model questionnaire) must be included.
Survey designers should decide whether or not to include the other optional categories (rows “O” through “T” on the model questionnaire) in addition to those required for constructing the MDD-W indicator. For example, they need to determine if insects and other small protein foods and red palm oil are consumed in the survey area and if so, to decide whether to include these categories on the questionnaire.
In addition to the categories listed in the model questionnaire, other foods of interest, including fortified and/or biofortified foods, can be added. Also, model questionnaire rows could be further disaggregated for data collection if survey designers want to capture consumption of one or more specific foods within an MDD-W food group or other food category. This may be the case when specific food items are promoted during interventions.
When modifying the questionnaire in these ways, survey designers should take care to ensure that it is still possible to construct the MDD-W indicator.
In the case when questions are added to capture consumption of fortified or biofortified foods, it is best to train enumerators to mark items in their “home” food group (e.g. grains, for fortified flour or golden rice) in addition to eliciting answers for added questions specific to fortified/biofortified foods. Note this “breaks the rule” of mutually exclusive food groups and categories as found in the model questionnaire. But marking in “home” food groups allows the standard construction of the MDD-W indicator.
When further disaggregating food groups to capture specific targeted foods, the questionnaire rows should be mutually exclusive. The tabulation instructions need to be modified to reflect the disaggregation.
When using the open recall method, further disaggregation of questionnaire rows does not affect responses or results, because the enumerator does not read the lists of items/examples to the respondent. However, when the list-based method is used, disaggregation of groups is likely to result in recall of more food items and can change survey results.
Translation and adaptation of the questionnaire
INITIAL TRANSLATION OF THE MDD-W MODEL QUESTIONNAIRE INTO SURVEYLANGUAGES This guide, which is in English17, contains an MDD-W model questionnaire. In countries where English is not the predominant language, the MDD-W model questionnaire with its generic food examples should be translated first into one of the major survey languages as a starting point for the linguistic and cultural adaptation work that follows. There are several methods of ensuring good translation, including group work to reach consensus on translation and back-translation to English18. This initial
There are plans to make Spanish and French versions available in the future.17
Although back-translation of survey questionnaires has been a common practice, some now emphasise focusing 18 attention on first producing the best possible translation and then directly evaluating the translation produced in the target language, rather than indirectly through a back-translation. See, for example, http://ccsg.isr.umich.edu/ translation.cfm.
translation into the predominant survey language will reflect the generic examples of foods provided in the MDD-W model questionnaire. The initial translation will be the starting point for adapting the questionnaire in other survey languages, if any.
BASIC LEVEL OF LINGUISTIC AND CULTURAL ADAPTATION OF THE MDD-W
QUESTIONNAIRELinguistic and cultural adaptation means modifying the translated MDD-W questionnaire to reflect cultural norms, vocabulary and usage (words and phrases) that will be easily understood and to include locally available and commonly consumed foods. This step will be carried out by the survey designers in consultation with a nutritionist and involves customising the introductory text and the food items in each questionnaire row.
Involvement of field staff (enumerators and supervisors) is also ideal, because in an open recall enumerators will be required to correctly classify reported foods into the food categories listed in each row on the questionnaire. Their ability to do this will be influenced by their comprehension of the objectives and by the quality and comprehensiveness of the examples filled in for each food category. Thus, involving field staff early in the adaptation process provides additional assurance of their ability to collect accurate information during the survey. Their input is also useful in reviewing the translation of the introductory text to make sure the language and terms will be understood by the respondent. Particular attention should be taken to carefully translate terms used to describe key concepts (such as “meal”, “snack” or “main meal”).
During this step, the names of the food categories should be reviewed and translated, and the list of examples in each filled in with a comprehensive list of seasonally and locally available foods, using local names where appropriate. (Appendix 2 provides guidance on how to classify individual food items into the questionnaire rows.) At some time during the process, it will be necessary to consult with individuals familiar with commonly consumed foods in the target population and with a nutritionist to review the draft questionnaire and advise on correct classification of the food items into the rows. The review can be done either by the nutritionist alone or, ideally, together with the survey designers and the field staff. Other resources may be consulted as well, such as questionnaires from nutrition modules in previous surveys, if available.
If the survey team is unable to carry out the second level of adaptation (below) because of time and/ or resource limitations, the work to identify ingredients of commonly consumed mixed dishes and foods likely to be consumed in small quantities, as described in the next section, should be done during the basic adaptation, using information from team members and knowledgeable persons.
SECOND LEVEL OF ADAPTATIONWhen time and resources permit, it is strongly recommended that a second level of adaptation be carried out to complete the lists of example food items for the questionnaire rows. This step involves consultations with members of the target population in the form of key informant interviews and focus group discussions in different subpopulations or locations within the survey area(s)19. These
conversations provide critical information on:
• Seasonally and locally available food items (including foods gathered in the wild) and their common names • Commonly consumed mixed dishes and ingredients used in these local dishes This step may be incorporated into enumerator training, as was done in Tajikistan, when the training was 19 focused only on collecting dietary diversity information (see http://www.fao.org/fileadmin/templates/nutrition_ assessment/Workshops/ Training_Report_Khujand_April_2015__03062015_.pdf). However, this would not be practical when training on the MDD-W is incorporated into a broader enumerator training session for multi-topic surveys.
• Foods that are typically consumed in small amounts that should not count as part of an MDD-W food group (see Box 1) but rather should be classified as condiments or seasonings • Commonly consumed “street foods” and prepared foods purchased outside the home Informants for this process can include experts at the local and national levels, community leaders, agricultural or health extension workers at the local level and women in the community who are responsible for food planning and preparation for their households. Informants from various communities in the survey area whose food patterns may differ should also be included: urban, rural and peri-urban residents, and different ethnic and livelihood groups. For surveys that will be carried out over large geographic areas with distinct dietary practices by location, it may be necessary to produce more than one adapted questionnaire.
OUTPUTS FROM THE LINGUISTIC AND CULTURAL ADAPTATIONOutputs of the adaptation process always include the adapted MDD-W questionnaire and can include job aids/guidance sheets on common ingredients of mixed dishes and on foods to place in the “Condiments and seasonings” category. Such guidance sheets are extremely useful for enumerator training, as well as for use as reminders during data collection. They may be in the form of cards, photographs or printed sheets.
Once this preparatory work has been done in a specific geographic area, subsequent surveys could use the same adapted questionnaires, enumerator instructions and guidance sheets, greatly reducing the preparation time.
Field testing, finalisation and piloting
FIELD TESTINGBefore finalising the MDD-W questionnaire, it is recommended that a small field test with a limited number of respondents (5–10 may be sufficient) be carried out to make sure that the examples in each questionnaire row are complete and that the respondents understand the script and the probing that elicit their open recall. The respondents are usually informed that this is a trial to improve the data collection instrument, and they may be interviewed afterwards to get their views on how well they understood the questions and were able to answer them. Members of the survey team would conduct interviews with a small convenience sample in locations similar to where the survey will be conducted and compare notes afterwards to identify any gaps or potential problems in comprehension by the respondents. Fine-tuning may be required to complete the food lists for each row or to modify the script and probing questions to improve clarity.
This last step in the adaptation process does not replace the standard practice of piloting the complete and final questionnaire when the MDD-W is incorporated into larger multi-topic surveys (see below).
FINALISATION OF THE MDD-W QUESTIONNAIREOnce these steps have been carried out in the major survey language, the final version of the MDD-W questionnaire is ready for use. If the survey is to be carried out using multiple languages, it will be necessary to repeat the steps above for each language to ensure that instructions are clearly understood and that the questionnaire includes the correct names of foods in each language.
However, if the process of translation and adaptation into the first language is thorough, adaptation to additional languages can be more rapidly accomplished.
30 Section 4 Preparing the MDD-W questionnaire
PILOT STUDIES IN THE CONTEXT OF THE LARGER SURVEYIn most situations, MDD-W will be measured in the context of surveys that include multiple topics and modules. A pilot study is a practice of all the survey steps, from start to finish, including all survey modules and procedures. Often a convenience sample of approximately 50 respondents is interviewed and their responses coded and analysed. Questions that are not clearly understood are modified, problems administering the questionnaire are addressed and the final revisions of the questionnaire are made20.
See, for example, http://www.tools4dev.org/wp-content/uploads/how-to-pretest-and-pilot-a-survey-questionnaire.pdf.
Enumerator selection Since the open recall method recommended in this guide places the burden on the enumerator to correctly classify reported foods into the rows on the questionnaire, it is strongly recommended to use enumerators who have some training in nutrition surveys or who have participated in the questionnaire adaptation. Ideally, enumerators will also have direct personal experience in shopping for and preparing local foods; in many cultures, this means they will be female.