Figure 1: Food insecurity scale
What is the Voices of the Hungry Project?
FAO’s Voices of the Hungry (VoH) project is an innovative approach to develop a measure of the severity of food insecurity as experienced by individuals or households. This initiative aims to establish the Food Insecurity Experience Scale (FIES) as a new global standard for measuring food insecurity experience that is valid, endorsed at the international level, and used for global and country monitoring.
For world-wide application of the FIES Survey Module (FIES-SM) consisting of eight questions, the Voices of the Hungry project will leverage on the Gallup® World Poll (GWP), a branch of Gallup, Inc. that has conducted nationally representative surveys in more than 140 countries annually since 2005. Beginning in 2014, the FIES-SM has been incorporated into the World Poll questionnaire. The data will be used to derive estimates of the prevalence of food insecurity at different levels of severity, gathered from a nationally representative sample of adults in all of the countries covered by the World Poll.
The GWP uses face-to-face interviews. In countries with 80% phone penetration, interviews are made via telephone. For the GWP, the FIES-SM will be inserted in a much larger questionnaire that includes questions regarding quality of life, employment and job climate, community infrastructure, civic engagement, and financial, social, physical, and emotional well-being.
What does the Food Insecurity Experience Scale (FIES) measure?
The FIES Survey Module (FIES-SM) used by the Voices of the Hungry project measures the severity of food insecurity of individuals or households facing constraints in their ability to obtain adequate food. It is derived from two widely-used and validated experience-based food security scales: the US Household Food Security Survey Module and the Latin American and Caribbean Food Security Scale (Spanish acronym ELCSA). The FIES-SM assesses the objective condition of food insecurity through information collected from individuals regarding their access to food. Compared to traditional ways of assessing food insecurity indirectly through determinants such as food availability, or consequences such as poor quality diets, anthropometric failures, and other signs of malnutrition, this method represents a significant change in approach to food security measurement.
The FIES-SM consists of a set of eight short questions asked directly to people. Generally, the interviews are conducted face-to-face, but they may be conducted by telephone as well. The questions focus on self-reported food-related behaviors and experiences associated with increasing difficulties in accessing food due to resource constraints. The Survey Module is based on a well-grounded construct of the experience of food insecurity composed of three domains: uncertainty/anxiety, changes in food quality, and changes in food quantity.
The idea is that these domains can be positioned on an underlying scale of severity, as shown for example in figure 1. The measure of food insecurity associated with a respondent can be located on the scale based on the number of positive responses to the questions (number of behaviors or experiences reported). Such measures are then used to classify respondents into categories of food insecurity severity.
What does the FIES not measure?
The FIES does not attempt to quantify food consumption nor does it provide a quantitative assessment of dietary quality. It is not a measure of malnutrition and cannot be used to detect nutritional deficiencies or obesity. Consequently, it is not the appropriate tool for assessing nutrition-specific outcomes of food security programs and policies.
Determinants of food insecurity are many and varied: at the local, regional, national and international levels, including factors as diverse as climatic conditions, food production and availability, food price volatility, poverty/income, nutrition knowledge and practices, and access to public services. The FIES is not designed to measure these determinants but rather to provide population estimates of the prevalence experiencing food insecurity at different levels of severity.
When the FIES module is included in surveys that also collect information on determinants and/or of outcomes of food insecurity, it can help identify respectively the different risk factors for individual or household food insecurity and its consequences in different contexts.
Is the FIES a food security indicator or a nutrition indicator?
The timeliness with which food insecurity indicators can be obtained through use of the FIES improves identification of areas where malnutrition may potentially occur, therefore informing effective preventative measures. While the FIES does not provide specific information on actual food consumption, dietary quality or nutritional status, it provides a valuable tool for the nutrition and food security community to increase knowledge regarding relationships between the experience of food insecurity and the problems revealed by indicators of malnutrition, including anthropometric outcomes.
The well-documented association between food insecurity and overweight is another factor that complicates analysis of the relationship between food insecurity (food access) and anthropometric outcomes. Many countries are experiencing a nutrition transition characterized by decreased stunting and wasting and a growing prevalence of overweight among the poor. This leads to the need for a different conceptual framework regarding the anthropometric outcomes of food insecurity that reflects the complexity of this relationship. Numerous studies have shown that food insecurity (restricted food access) and overweight/obesity can co-exist, both in children and adults. Explanations for the association between food insecurity and overweight include consumption of less expensive energy-dense foods, stress-induced eating disorders, and metabolic adaptations to periods of going without food. Individuals who experienced hunger in infancy or childhood may be at greater risk of obesity as adults.
It is also worth noting that experiencing food insecurity is not synonymous with malnutrition. Food insecurity can have consequences for other aspects of people’s health and well-being, such as negative psychosocial effects, that are not necessarily associated with nutritional status. Even in the absence of observable negative effects on nutritional status, the experience of food insecurity is a serious problem itself, indicating a violation of the Human Right to Adequate Food.
In summary, the FIES is a measure of food insecurity based on a theoretical construct that has relevance for dietary quality and food quantity considerations. It does not, however, quantify food consumption, dietary quality, or malnutrition per se. It may be considered a nutrition indicator to the extent that a condition of food security is strongly associated with good nutrition, and nutritionists play a role in guaranteeing the Human Right to Adequate Food. Used together with traditional measures of food consumption, dietary quality, and nutritional status, the FIES has the potential to contribute to a more comprehensive understanding of the causes and consequences of food insecurity, including nutritional and dietary impacts.
How does the FIES differ from other indicators of food insecurity?
The FIES measures the severity of food insecurity of individuals or households facing constraints in their ability to obtain adequate food. The tool is designed to have cross-cultural equivalence and validity in both developing and developed countries to produce comparable indicators of the prevalence of food insecurity in a population at various levels of severity.
Compared to the other indicators that have been proposed and used to assess the state of food security at national level, experience-based food insecurity scales like the FIES stand out for the following reasons:
a) Directly asks people about their experience of food insecurity.
b) Ease of administration and timeliness of reporting.
c) Soundness of the statistical basis used to enable cross-country comparisons based on information collected on individuals or households.
d) Ability to reflect the depth of food insecurity by distinguishing between different severity levels.
e) Possibility to disaggregate results by gender when applied at individual level and by sub national groups when applied in surveys with samples that are representative at sub-national level.
Is the FIES a subjective measure?
Experience-based food insecurity scales like the FIES rely on self-reported information regarding food-related behaviors in the face of limited access to food. They reflect respondents´ own perspectives regarding the adequacy of their food consumption rather than that of nutritionists or economists. Sometimes their validity has been questioned using the argument that people may not be good judges of their food security status.
Three questions may be characterized as referring to “subjective” perceptions (Were you worried you would not have enough food? Were you unable to eat healthy and nutritious food? Did you eat less than you thought you should?) while the remaining five questions ask about “objective” behaviors (running out of food, skipping meals, going a whole day without eating) due to a lack of money or other resources.
It is not uncommon for people to question respondents’ ability to be their own judges of whether or not their diet is “healthy and nutritious”. Accumulated experience with food insecurity scales in Latin America and elsewhere, particularly focus group research conducted to adapt scales for different contexts, has shown that people are surprisingly good judges of what constitutes a healthy and nutritious, or balanced diet. Numerous validation studies have shown a close association between severity of food insecurity and dietary quality measured using traditional nutrition indicators.
The FIES is a psychometric scale similar to other widely-accepted psychometric scales designed to measure unobservable traits such as aptitude/intelligence, personality, and a broad range of social psychology and health-related conditions. The items on the scale are based on well-grounded empirical research regarding the experience of hunger and poor food access.
Is the FIES an individual or a household measure of food insecurity?
The FIES being applied through the Gallup® World Poll, as part of the Voices of the Hungry project, is a measure of food insecurity at the individual level. This choice is determined by the Gallup® World Poll methodology which is a survey of individuals. However the FIES can be easily applied at the household level with minor adaptations. The Voices of the Hungry Project aims to promote inclusion of the FIES Survey Module in national-level household surveys, as well, such as Household Income and Expenditure Surveys, Household Budget Surveys and Living Standard Measurement Studies, and health and nutrition surveys.
Measurement of people´s personal experiences with food insecurity has many advantages, including the capacity to monitor violations of the human right to adequate food. Another major advantage of measuring food insecurity at the individual is that results can be meaningfully disaggregated by gender, so that different and comparable measures of the prevalence of food security among women and men can be calculated. As the Gallup® World Poll is designed to be representative only at the national level, based on relatively small samples, further gender disaggregation by age, ethnic group, livelihood group, etc. is not feasible. This is why the Voices of the Hungry Project is promoting inclusion of the FIES Survey Module in sub-national population surveys.
Why does the FIES have a 12 month recall period?
For the Gallup® World Poll, the recall period for each of the questionnaire items is the “past 12 months”, a decision that was taken to avoid the influence of seasonal effects and to improve comparability of the measure across different areas. However, the FIES Survey Module is a flexible tool even with respect to recall period. The choice of the recall period depends on the objectives for collecting the data. Where seasonal changes are of interest, a three-month period might be the best choice to investigate how food insecurity varies in different seasons. However, this would limit comparability if data are collected across different areas that do not have the same climatic or agricultural calendar. In the case of national surveys with the objective of estimating overall prevalence of food insecurity in different parts of the country, or multi-national surveys that include countries with different environmental and climatic zones, In this case, the 12-month period is recommended. When a survey is being carried out in settings undergoing a humanitarian crisis, the investigators may be interested in capturing the most recent time period, in which case a one-month recall would be appropriate.
What are the potential uses of the FIES?
The unique contribution of the FIES is that it is a measure of food insecurity as experienced by individuals or households obtained through direct administration of a set of related questionnaire items. It is simple and inexpensive to apply yet theoretically sound and reliable. The FIES Survey Module can be easily incorporated into many kinds of individual and household surveys. It can be used together with other indicators to deepen our understanding of the determinants and consequences of individual and household food insecurity.
The information resulting from the uses of the FIES in different settings is relevant for a variety of audiences, including advocates, community leaders and activists, researchers, programme managers and government officials at all levels. FAO is committed to more effectively linking information and action. Prevailing models of policy-making concede a potentially influential role for survey-based information. However, there is also consensus that policy-making is a messy process characterized by negotiation among competing interests and diffuse decision making processes (Jones, 2009; Weiss, 1983). Those responsible for producing the information are therefore called upon to strive to improve their political savvy and reflect on the uses and appropriations of the resulting information.
Through the Gallup® World Poll, the Voices of the Hungry project will use the FIES to estimate the prevalence of food insecurity of individuals at different levels of severity in the national populations of most countries and regions of the world. Global information collected annually will be used for monitoring trends within and across regions and for tracking progress towards achieving international goals for eradicating hunger. The FIES has great potential to contribute to monitoring progress related to the Post 2015 Sustainable Development Goals Initiative.
With repeated application of the FIES Survey Module in the same population at either individual or household level, it is possible to monitor trends and changes in food insecurity levels over time. The Module can potentially be used to assess the impact of a food security program or policy. If applied before and after an intervention, it is possible to detect changes in the severity of food insecurity, taking into consideration other variables that might have influenced the changes besides the specific intervention itself.
When included in surveys that are representative beyond the national level, for example at sub-national levels, and regions of a country, states or provinces, the prevalence of food insecurity can be estimated for sub-groups of the population, such as rural or urban residence, age groups, and different racial or ethnic groups.
Results from surveys that include the FIES Survey Module can also be used to inform decisions regarding priorities for targeting programs and resources. While it is not appropriate to use the FIES to identify individual beneficiaries for programs, the information provided by population surveys that include the FIES Survey Module can serve to identify vulnerable sub-populations or geographic areas that are more affected by food insecurity. Results from project or program impact research that include the FIES can be used to inform decisions regarding investments for up-scaling successful projects and programs.
The FIES Survey Module can be used together with other indicators to identify risk factors and consequences of food insecurity. The phenomenon of food insecurity encompasses much more than what the FIES captures; it includes aspects ranging from social, economic and agriculture policies at the international and national levels to livelihood strategies, access to public services, basic sanitation, food habits, and nutritional status at the household level. Since the FIES directly assesses individual or household food insecurity through administration of a number of related questionnaire items, it can be applied in broad-based studies together with indicators of these additional aspects collected on the same units to build a better understanding of the complex phenomenon of food insecurity and to inform policy aimed at improving the well-being of the population and ending hunger.
Is there a potential response bias due to respondent expectations of receiving assistance/benefits?
Response bias due to expectations of receiving assistance is a potential problem for many types of surveys. The risk can be addressed through sound survey methodology to ensure that respondents are explicitly informed that their replies will have no influence on potential benefits and that their personal identification will not be provided to authorities or aid agencies.
It would be advisable to avoid using interviewers known to be associated with assistance programs to collect data. If information is being collected specifically for programmatic purposes, an independent team not identified with the agency responsible for service delivery should be employed to carry out the survey.
Can the FIES items be applied and analyzed individually?
Together, the items of the FIES Survey Module compose a psychometric scale designed to cover a range of severity of food insecurity and must be analyzed together as a scale, never as separate items. This is because the items are meant to collectively scan the entire severity range of food insecurity, as shown in Figure 1.
An additional consideration is that assessing food insecurity by analyzing percentages of positive answers to individual items is much less reliable from a measurement point of view compared with using the measures obtained from the joint analysis of the 8 items, each of which contributes to assess the level of food insecurity of the respondent.
Each FIES item, when considered individually, would have substantial measurement error due, for example, to possible inconsistencies in understanding and recall. Taken together, the questions compose a unitary psychometric scale in which differences at the question level tend to cancel out, reducing measurement error for the scale as a whole.
Is there a universal order in which the eight FIES questions fall along the scale of severity?
The FIES is grounded in Item Response Theory, which is applied to establish the severity of each item on the scale. The idea is rather simple, yet powerful: by looking at the way in which the many respondents report on one of the experiences, one can establish a measure of the severity associated with that experience, i.e., where that experience is located on the scale (intuitively, experiences reported by a larger number of subjects are deemed less severe, and vice versa).
Figure 1 shows how the general domains of food insecurity tend to flow along the underlying scale of severity, with uncertainty/anxiety being the least severe, followed by compromising food quality, reducing food quantity, and finally, experiencing hunger. However, an essential aspect of the FIES is that the severity of each question (its position on the scale) for a specific population is established based on the number of positive responses to that item, as described above. Therefore, people’s responses in different countries may result in a different ordering of the items along the scale. This may be due to different translations reflecting somewhat different objective conditions, or from differences in the ways in which food insecurity is experienced or managed in different cultures and livelihood systems.
What are the different classifications/levels of severity of food insecurity and what do they mean?
The classifications of food insecurity severity are: mild, moderate, and severe. The experience of food insecurity, according to the theoretical construct on which the FIES is based, is characterized initially by uncertainty and anxiety regarding food access (mild food insecurity), followed by changes in the quality of the diet as the situation worsens, such as a less balanced, more monotonous diet (mild-to-moderate food insecurity). With increasing severity, the quantity of food consumed decreases as portion sizes are reduced or meals are skipped (moderate-to-severe food insecurity). Severe food insecurity may be considered essentially equivalent to hunger, as it is characterized by not eating for an entire day, or experiencing hunger and not eating, due to lack of money or other resources.
The classifications presented here are approximate because they are yet to be defined based on ongoing international data collection, and they may differ slightly from one country or culture to another. Thresholds will be determined based on data collected from many countries.
The different levels of severity of food insecurity (i.e. mild, moderate, and severe) are sometimes combined for analysis and reporting. However, it is important to analyze and report them separately whenever possible, as the differences between them are meaningful, theoretically as well as empirically.
How are the classes (‘cut-points’) of food insecurity severity determined?
Raw scores can and have been used for classification (in the US and in Latin America, for example) by assigning each raw score to a food security status category. Cut-points, or thresholds, defining the categories have mostly been defined based on the underlying conceptual framework together with face validity.
This most simple of classification procedures may work well within a single country, or across countries sharing a common language and culture, but potentially introduces substantial bias in comparisons across countries with substantially different cultures and languages. When cross-country comparability is an objective, it is necessary to resort to more sophisticated scoring methods, based on “probabilistic assignment” of cases to food security status categories, and taking account of the fact that in any given country, one or two items may indicate different levels of severity than the corresponding questions in most countries.
Are the questions and results of the FIES comparable across different cultures and countries?
There is considerable evidence pointing to the validity and reliability of experienced-based food insecurity scales in diverse contexts. Various scales have been developed and validated in different regions of the world, the majority of which measure food insecurity in essentially the same way.
The FIES is a global adaptation of the household-referenced and adult-referenced items in the Latin American and Caribbean Food Security Scale (ELCSA), whose origins derive from the US Household Food Security Survey Module, the Brazilian Food Insecurity Scale, and a similar scale adapted for Colombia. In countries where the ELCSA has been applied, as well as the more than 145 countries in which the FIES was applied in the 2014 Gallup World Poll®, the scientific evidence indicates that experience-based food insecurity scales can measure food insecurity accurately and in a culturally-comparable manner.
In the context of the VoH project, where cross-country comparability is the objective, food insecurity measures will be made comparable in terms of severity along the continuum defined by the arrow in Figure 1, through the use of statistical techniques borrowed from the toolkit of Item Response Theory (IRT) models, commonly used in the educational and psychological testing fields. Cut points will be established based on the cognitive content of the scale questions, which are located on the same severity-of-food-insecurity scale as the raw-score-based respondent measure.
Nevertheless, further research is needed to confirm the validity of a global experience-based food insecurity scale and ensure cross-cultural comparability of results. FAO´s Voices of the Hungry project aims to undertake this challenge based on global data made available for the first time to enable such analyses.
Why are frequencies (number of time each experience occurred) not collected by the FIES? Isn’t frequency a sign of severity?
Indeed frequency of occurrence of food-insecure conditions can be considered as a sign of severity. However, previous studies in US have revealed that for a 12-month referenced scale, frequency of occurrence and greatest severity during the year are somewhat distinct dimensions that cannot be effectively measured along a one-dimensional continuous scale (Figure 1).
Additional considerations include the risk of compromising comparability of answers across settings, as the frequency options “rarely”, “sometimes” and “often” are not standardized. These potential complications suggest that, while it is a worthwhile issue for future research, it is best to begin with the more straightforward never (no) vs. ever (yes) response categories.
In countries that already have their own nationally-developed food security scales, is it possible the FIES will produce different results?
Some countries in the world already have existing nationally-validated food insecurity scales that have been applied in national surveys and/or are incorporated into national monitoring systems. Most of these scales are based on the same theoretical framework and construct of food insecurity as the FIES Survey Module although the number and wording of the questions differs, as well as the reference period. Almost all of them refer to the household level and include questions referring to children living in the household.
To compare results obtained with another scale with those of the FIES, the two scales will need to be placed on a common metric, and comparisons will have to be made at the same severity levels. (The standard thresholds used for each of the scales may not be at the same level of severity). Because of their similarity and shared origins, FAO believes that, once these adjustments are made, it is unlikely that results of the FIES will differ significantly from results of surveys using other, nationally developed food insecurity scales. Nevertheless, FAO is working with national governments to determine the best way to address this potential problem.
| || |