3. Types of variables

The consequences of a change in dietary energy intake should be observable in one or more of the following dimensions:

(a) Body size and composition
(b) Time allocation to different activities
(c) Intensity of activity

The measurement of these changes can be classified in terms of the following taxonomy:

1. Intermediate behavioral variables
2. Intermediate biological variables
3. Outcome variables

(a) Investigator-determined
(b) Community- and subject-determined

The types of variables that can be measured within this taxonomy include the following:

1. Intermediate behavioral variables

1.1. Activity patterns (time-shares among activities classified by two dimensions)

(a) Activity type
(b) Energy cost

1.1.1. Instrumental (task-oriented)

- market production
- home production
- education and training (human capital formation)

1.1.2. Expressive (relational)

- leisure time
- social interaction

1.2. Social support network
1.3. Intrahousehold economic roles and decision-making

2. Intermediate biological variables

2.1. Body composition/growth and development
2.2. Energy intake
2.3. Health status

3. Outcome variables

3.1. Investigator-determined (etic)

3.1.1. Economic
Household income (monetary and non-monetary) expenditures (secondary effects, e.g., gardening; home improvement)

3.1.2. Psychological: cognitive and psycho-motor development and social-emotional functioning

3.2. Community- and subject-determined (emic)

3.2.1. Individual level
3.2.2. Community level

4. External (community) variables

4.1. Economic; e.g., market prices of final and intermediate goods
4.2. Community infra-structure (autonomous or as a secondary effect)

It should be clear that the identified outcome variables may have secondary or multiplier effects (e.g., on the intermediate variables of energy intake, health status). The study design should fully anticipate measurement of the secondary effects over time.

In order to define what community/subject (culturally or emically) determined outcomes should be included, the study design must anticipate a period of interaction with community members before actual, pre-intervention measurements are initiated.