Data Sources and Integration

Module 1: Nutrient Intakes

Overview

In the Guatemalan data context, estimating nutrient intake from household survey data requires integrating multiple data sources: consumption records from a nationally representative survey, food composition tables with nutrient content, and expert-provided conversions to translate local measurement units into standardized weights.

This module loads and integrates three primary sources to create a unified analytical dataset:

  1. ENCOVI 2023 — Guatemala’s National Survey of Living Conditions provides household food consumption records, demographic characteristics, and socioeconomic indicators for over 10,000 households.

  2. INCAP 2014 — The Instituto de Nutrición de Centro América y Panamá (INCAP) Food Composition Table for Central America contains nutrient content (energy, protein, iron, zinc) and amino acid profiles (lysine, tryptophan, methionine, cysteine, threonine) for foods common in the Guatemalan diet.

  3. Expert protein-quality and conversion data — Two contributions that complete the INCAP base:

    • Protein quality scores (PDCAAS) — Amino acid profiles and Protein Digestibility Corrected Amino Acid Scores compiled from the literature by nutrition expert María José Soto Méndez (FINUT) — principally (Boye et al., 2012), (Suárez López et al., 2006) and FAO amino acid data, following the FAO/WHO/UNU reference methodology (World Health Organization et al., 2007). Where no literature value was available for a food, the score was derived from comparable food groups. These scores are delivered as an additional column in the composition table for direct application in the framework.
    • Unit conversion factors — Weights in grams for non-standard units (bags, packages, cups) compiled through on-site market research in Guatemalan supermarkets and markets, addressing the gap left by the original composition tables where most entries lacked complete gram equivalences.

The integration process standardizes food quantities to grams and links consumption records to nutritional composition. The distribution of household intake to individuals is documented in Intake Calculation.

ENCOVI 2023: Household Survey Data

ENCOVI (Encuesta Nacional de Condiciones de Vida) is Guatemala’s primary source for household-level socioeconomic data. The 2023 survey includes a detailed food consumption module where households report foods purchased and obtained during the previous 15 days, along with quantities and measurement units.

Three datasets are loaded from ENCOVI:

  • Food consumption — Individual food items with quantities, units, and acquisition method (purchased vs. obtained without purchase).
  • Household characteristics — Housing conditions, basic services, and assets used as socioeconomic predictors in transfer models.
  • Individual demographics — Age, sex, education, and household relationship for each person, enabling individual-level intake calculations.

INCAP 2014: Food Composition Tables

INCAP maintains the reference food composition database for the region. The 2014 edition provides nutrient content per 100g for foods commonly consumed in Guatemala.

The integrated composition database combines three provenance layers, each documented at the value level:

  • INCAP composition base — Energy, protein, iron, zinc, and the edible-portion factor, transcribed from the INCAP food composition report for Central American foods.
  • Amino acid profiles — Essential amino acids including lysine and tryptophan, critical for assessing protein quality in maize-based diets, compiled from the literature by María José Soto Méndez (FINUT).
  • PDCAAS scores — Protein Digestibility Corrected Amino Acid Scores indicating protein quality relative to human requirements (Boye et al., 2012). Literature-based scores were compiled by FINUT; where no literature value was available, the score was derived from comparable food groups (Suárez López et al., 2006).

The composition tables are keyed by food code (ali) and linked to the ENCOVI survey food code (codine), which is the operative join key to consumption records. All nutrient values are expressed on a per-gram basis to facilitate calculations with varying portion sizes.

TipProtein Quality Score Assignment

Each food in the composition database carries a PDCAAS value assigned through one of three routes:

  • Literature-based and derived scores — Most foods receive a score compiled by María José Soto Méndez (FINUT) from the literature (Boye et al., 2012; Suárez López et al., 2006), or derived from comparable food groups where no literature value was available.
  • Negligible protein contribution — Foods that provide no meaningful dietary protein (sugars, fats and oils, salt, soft drinks) are assigned a conservative value of 0.1, reflecting their negligible contribution to protein quality.
  • Insufficient composition data — Where composition data are too sparse to assign a value, the same conservative 0.1 default is applied so that protein quality estimates are not inflated for foods with undetermined digestibility.
Table 1: PDCAAS assignment across the composition database
Category N Percentage
Literature-based or derived score 90 81.8%
Negligible protein contribution (0.1) 19 17.3%
Insufficient composition data (0.1 default) 1 0.9%
Total 110 100.0%

Food Unit Equivalences

Household surveys record food quantities in local measurement units — bags (bolsas), packages (paquetes), trays (bandejas), dozens (docenas) — that vary by food type and regional packaging conventions. Converting these to grams requires local expertise.

The food equivalence table was compiled by María José Soto (FINUT) through on-site market research in Guatemala during 2024. The table provides:

Data Element Purpose
Gram conversions Weight in grams for each food-unit combination
Plant-based classification Identifies foods containing phytates that affect mineral bioavailability
Survey frequency Number of times each unit appears in ENCOVI responses

The conversions reflect the retail packaging that survey respondents report against. A bolsa (bag) of rice, for example, has a different weight than a bolsa of beans, and both differ from standardized INCAP reference weights.

Household Socioeconomic Variables

Household characteristics from ENCOVI are recoded into interpretable factor variables. These serve as predictors in the transfer models that estimate nutrient intake for SIVESNU children (source of anthropometric data) based on their household conditions.

The variables capture multiple dimensions of socioeconomic status:

  • Housing quality — Wall, roof, and floor materials.
  • Basic services — Water source, sanitation, electricity.
  • Assets — Television, telephone, computer ownership.
  • Geographic context — Department and urban/rural classification.

Individual nutrient intake is derived from household intake using Adult Male Equivalent (AME) factors. See Intake Calculation for the methodology and application.

Summary

ENCOVI 2023 Survey Data

Table 2
ENCOVI 2023 Survey Data. Loaded datasets overview
ENCOVI 2023 Survey Data
Loaded datasets overview
Dataset N Variables
Food consumption 569,503 33
Households 10,964 129
Individuals 46,017 626

INCAP 2014 Food Composition Database

Table 3
INCAP 2014 Food Composition Database. Integrated composition tables with amino acids and PDCAAS
INCAP 2014 Food Composition Database
Integrated composition tables with amino acids and PDCAAS
Component Available
Integrated nutrient database 110 foods
Energy (kcal) ✓
Protein (g) ✓
Iron (mg) ✓
Zinc (mg) ✓
Lysine (g) ✓
Methionine (g) ✓
Cysteine (g) ✓
Threonine (g) ✓
Tryptophan (g) ✓
PDCAAS scores ✓

External Expert Data

Table 4
External Expert Data. María José Soto (FINUT) contributions
External Expert Data
María José Soto (FINUT) contributions
Dataset N Description
Food equivalences 718 Expert-provided unit conversions
Unique foods 116 Foods tracked in survey
Measurement units 95 Survey measurement codes
PDCAAS values 109 Protein quality scores

Processed Base Data

Table 5
Processed Base Data. Cleaned and prepared datasets for nutrient intake calculations
Processed Base Data
Cleaned and prepared datasets for nutrient intake calculations
Dataset N %
Household socioeconomic 10,964 —
Individuals (demographic) 46,017 —
Back to top

References

Boye, J., Wijesinha-Bettoni, R., & Burlingame, B. (2012). Protein quality evaluation twenty years after the introduction of the protein digestibility corrected amino acid score method. British Journal of Nutrition, 108(S2), S183–S211. https://doi.org/10.1017/S0007114512002309
Suárez López, M. M., Kizlansky, A., & López, L. B. (2006). Evaluación de la calidad de las proteínas en los alimentos calculando el escore de aminoácidos corregido por digestibilidad. Nutrición Hospitalaria, 21(1), 47–51.
World Health Organization, Food and Agriculture Organization of the United Nations, & United Nations University. (2007). Protein and Amino Acid Requirements in Human Nutrition: Report of a Joint WHO/FAO/UNU Expert Consultation. World Health Organization.