3
security and nutrition monitoring activities.
Tucker et al. (1989) points out that in many
countries food security-related data continue to
be collected for unspecified reasons and with
unknown reliability. A common problem limit-
ing the usefulness of food security and nutrition
data has been the use of indicators which are
ambiguous with respect to the causes of changes
in the level of the indicator. As a result, al-
though such data may be useful for indicating
the extent of problems, they often fail to reflect
causal links to policy decisions needed to make
the information more "actionable." For example,
used alone, anthropometric measures fail to
provide insights regarding appropriate interven-
tions for fighting malnutrition problems, be-
cause they fail to distinguish among various
causes of malnutrition, such as inadequate food,
sanitation, or health care.
Estimating the nature and magnitude of link-
ages between outcome measures of consump-
tion and nutrition and other causally related
variables can help improve understanding of
these processes. Rainfall data, for example, if
rainfall is seen as strongly linked to food con-
sumption, can indicate drought-caused food in-
security crises and perhaps suggest policies to
address supply-side variability (e.g., food im-
ports). Conversely, expenditure data may indi-
cate failures of income and suggest demand-
side measures (e.g., labor-based relief projects)
(Tucker et al. 1989). Understanding the nature
of consumption-nutrition linkages is also im-
portant to test the appropriateness of using food
intake measures (especially those based on food
expenditure data) as proxies for indicating nu-
tritional status. An example is using estimates
of elasticities of food expenditures or intakes in
studies of income-nutrition linkages (Schiff and
Valdes 1990b).
But consensus has not been easy to reach on
the precise nature and magnitudes of these link-
ages, or their implications for policy. Schiff and
Valdes (1990a), for instance, point out that "criti-
cal elements of the pathway from changes in
income to its effect on nutritional status are still
questioned." One reason for this lack of con-
sensus, as just mentioned, is that the nature of
these linkages may differ in different contexts.
This raises the question of the generalizability
of research findings. A second reason for this
lack of consensus has been concerns and dis-
agreements over what the appropriate analyti-
cal approaches and indicators are for analyzing
these linkages.
Issues Regarding Indicators,
Measurement Errors, and Analytical
Methods
The quality of estimates of the nature and mag-
nitude of linkages among food availability, ac-
cess, consumption, and nutrition depends criti-
cally on using appropriate indicators, reliable
data, and valid analytical methods. Disagree-
ments among researchers regarding the degree
to which these conditions have been met in
various empirical studies has led to consider-
able controversy in the food security linkages
literature over the interpretation and meaning-
fulness of certain research findings.
One problem limiting the usefulness and
validity of many research results is that concep-
tual inequivalence inevitably exists between
variables of interest and their associated indica-
tors. For example, while weight/height mea-
surements and nutritional status may be corre-
lated, they are not conceptually equivalent. In
other words, they do not necessarily imply each
other. Despite the obviousness of this point,
such conceptual inequivalence is generally over-
looked in practice. Researchers frequently make
the leap from observations on anthropometric
data to conclusions about nutritional status,
without explicit recognition of the differences.
A second reason for indicators being inad-
equate proxies of underlying variables is the
presence of data measurement errors. Sources
of measurement errors may include imperfect
recall or strategic behaviors by respondents,
imperfect communication between respondents