Back to guide home

2 Understanding MICS data

MICS microdata are collected through face-to-face interviews in carefully selected nationally or subnationally representative sample of households. The survey primarily collects data on issues that may affect children’s lives. It routinely collects household and individual background data so that differences associated with age, gender, education, wealth, location of residence, ethnicity and other characteristics can be analysed. To this end, MICS uses several questionnaires (at most 8) which come as several microdata files with different units of observation. A useful way to think about MICS microdata is as a set of linked microdata files below:

HOUSEHOLD (hh)
   │
   ├── household members (hl)
   │
   └── one randomly selected child aged 5–17 (fs)
           │
           └── ages 7–14 may take the Foundational Learning Skills (FLS) assessment

The household questionnaire (hh) tells us about the household and every member listed in it. Additional questionnaires collect more detailed information for particular populations, including women (wm) , men (mn), children under five (ch) and one randomly selected child aged 5–17 per household (fs).

2.1 Common MICS microdata files, keys, and who is in fs

MICS6 usually contains 8 questionnaires (or, microdata files). However, each country’s national steering committee, i.e., National Statistics Office, education, labour and health ministries, and UNICEF, decides on the final set of questionnaires to be included in their survey based on an assessment of a country’s data gaps. Shown below are the typical microdata files researcher will encounter.

File	Unit of observation	Examples of information
hh	Household	Household characteristics, wealth, water and sanitation
hl	Household member	Age, sex, relationship, education on the roster
wm	Woman aged 15–49	Women’s individual questionnaire
mn	Man aged 15–49, where administered	Men’s individual questionnaire
ch	Child under five	Under-five questionnaire
fs	One randomly selected child aged 5–17 per sampled household	Child background, schooling, child labour, child functioning, parental involvement, FLS
bh	Birth	Birth-history records from the women’s questionnaire

For early-foundational learning

The Foundational Learning Skills (FLS) assessment is in fs microdata file. Household characteristics often still need to be added from the hh and hl microdata files.

2.2 Key identifiers in each microdata file

MICS microdata files do not share one key identifier, because the unit of observation differs. The HH1 variable is the cluster, HH2 is the household within the cluster, and a line number (HL1/LN/FS4) then picks out one person inside that household.

Household:              HH1 + HH2
Person on the roster:   HH1 + HH2 + HL1
Selected child:         fs: HH1 + HH2 + LN
                        hl: HH1 + HH2 + HL1     (LN corresponds to HL1)
Mother or caretaker:    fs: HH1 + HH2 + FS4
                        hl: HH1 + HH2 + HL1     (FS4 is another person's line)

The household (hh) microdata file stops at HH1 + HH2. The household can then be splitted in two: left is the household member roster (hl microdata file that adds HL1 variable for every member), right is the selected child (fs microdata file that adds LN variable). It is important to note that LN is the child’s own line number in the fs, which can be the same number in hl, but using HL1 . FS4 variable in the fs microdata file is the mother or caretaker’s identifier.

                         HOUSEHOLD
                            hh
                         HH1 + HH2
                             │
              ┌──────────────┴──────────────┐
              │                             │
              ▼                             ▼
       HOUSEHOLD MEMBERS              SELECTED CHILD
              hl                           fs
       HH1 + HH2 + HL1              HH1 + HH2 + LN
              ▲                             │
              └──────── LN = HL1 ──────────┘

       SELECTED CHILD                       HOUSEHOLD MEMBER
              fs                                  hl
       HH1 + HH2 + FS4  ─────────────►  HH1 + HH2 + HL1
                    mother/caretaker

2.3 Who is in fs microdata file

The fs microdata file contains information on one randomly selected child aged 5–17 per sampled household. Please note that it is not every child of that age in the household listing that can be found in this file.

Sampled household
      │
      ▼
List all members (hl)
      │
      ▼
Children aged 5–17
      │
      ▼
Randomly select one child → fs record
      │
      ├──────── age 5–6 ────────► no FLS assessment 
      ├──────── age 7–14 ───────► eligible for FLS assessment 
      └──────── age 15–17 ──────► no FLS
      

After the random selection of the 5-17 aged child, they can either be eligible or ineligible for the FLS module that sits inside the fs microdata file. The FLS module directly assesses reading and numeracy among children aged 7-14. UNICEF describes the module as measuring skills expected at approximately Grade 2/3 level in a household-survey setting.

This age range is one of the major analytical strengths of MICS. Researchers can examine foundational skills among children currently attending school, among children who have never attended, among children who attended previously but are no longer in school, by age rather than only grade, across grades, and in relation to household and family characteristics.

Because MICS is household based, the assessment is not restricted to children who are present in school on an assessment day.

Name the denominator

“Percentage with foundational reading skills” is incomplete until the reader knows among whom: children aged 7–14, children of Grade 2/3 age, children attending Grade 2/3, or another population you define.

A child aged 10 in Grade 2 can enter a grade-based Grade 2/3 estimate and still be over-age for that grade. A child of Grade 2/3 age may be in another grade or out of school. Therefore, it is advisable to use these denominators separately because they answer different questions.

When the rows are selected children, use fsweight. Adding a variable from hh or hl does not change the weight. The analysis population does.

2.4 Foundational Learning Skills Module

The Foundational Learning Skills (FLS) module assesses reading and numeracy among children aged 7–14 in the household, whether or not they are in school. Skills are those expected in early primary (about Grade 2/3).

The official “foundational skills” indicators are conjunctive: the child must meet every component. The microdata still hold the items, so you can look at domains, totals, and distributions as well as the headline yes/no.

2.4.1 Reading

Eligible child age 7–14
          │
          ▼
Permission / child participation
          │
          ▼
Assessment language (is a booklet available?)
          │
          ▼
Does the child agree to read?
          │
          ▼
Practice / routing where it applies
          │
          ▼
Read story → words attempted, words missed
          │
          ▼
Comprehension: 3 literal, 2 inferential

The interviewer assigns a language from what the child speaks at home and, for children with schooling, the language teachers use. If no suitable booklet is available, the child never reaches the story. In surveys with more than one passage, use the UNICEF files, not the IPUMS reading variables (Getting IPUMS MICS files).

A blank story result is not one kind of missing. A child with no suitable language, a child who declined to read, and a child who attempted the story and could not read it are different cases. Do not treat them all as NA.

The official foundational-reading indicator requires all three:

≥ 90% of passage words correct
           AND
3/3 literal questions correct
           AND
2/2 inferential questions correct
           ▼
FOUNDATIONAL READING SKILLS

You cannot compute the 90% rule without the passage length. Attempted and incorrect words are not enough; story length varies by country and language.

Reading comparisons across countries need more caution than numeracy. Differences can reflect language of instruction, language at home, the assessment language, the adapted story, and the child’s schooling, not only reading skill. The assessment is for national monitoring with country adaptation, not a language-neutral test.

2.4.2 Numeracy

Twenty-one items in four tasks. There is no language-specific passage of varying length.

Task	Items	What the child does
Number reading	6	Reads numbers shown by the interviewer
Number discrimination	5	Identifies the larger of two numbers
Addition	5	Solves addition problems
Pattern recognition	5	Supplies the missing number in a sequence
Total	21

A task counts as completed only if every item in it is correct. Foundational numeracy requires all four tasks:

6/6 number reading
          AND
5/5 number discrimination
          AND
5/5 addition
          AND
5/5 pattern
          ▼
FOUNDATIONAL NUMERACY SKILLS

The binary hides large differences in performance:

                    Number   Discrim.   Addition   Pattern    Total
Child A               6/6      5/5        5/5       4/5      20/21
Child B               3/6      1/5        0/5       0/5       4/21

Foundational numeracy?  NO                            NO

Use the official indicator when the question is mastery of all four tasks. Use task scores or the 21-item count when the question is how many items children get right. Neither is always better.

The questionnaire has stop rules. Item responses can be correct, incorrect, or not attempted. Do not treat every blank as the same without the skip pattern.

Numeracy avoids passage-language problems. Age, grade, current attendance, and when in the school year the survey ran still shape cross-country comparisons. A Grade 3 child assessed at the start of the year has had less Grade 3 instruction than one assessed at the end.