Indicator Guidelines

 

Indicator Pages

 

The below description provides a summary of the contents of the indicator pages that follow in this guide. Each indicator page provides a description of the indicators, the population group and time period they represent, how the indicators are calculated, including the definition of numerators and denominators and the datasets and variables used, how missing and “don’t know” responses are handled, any changes over time, and useful references and resources.

 

①           Brief subject title for the indicators

 

②           Title of the indicators in a single statement

 

③           Definition

 

Provides a more detailed definition of the indicators, listing each indicator separately.

 

④         Coverage: Summarizes the population covered by the indicators and the time period to which the indicators apply:

⑤                         Population base: The population group to which the indicators apply (e.g. households, household members, women, men, births, children, etc.), and the datasets to use for this population (e.g. HR, PR, IR, MR, BR, GR, KR, NR, SR, etc.)

⑥                         Time period: The period preceding the survey to which the indicators apply

 

⑦         Numerator(s): Number of the population meeting the criteria for inclusion in the numerator(s) of the indicators. Criteria for selection for the numerator(s) are given based on the variables in the recode file. Unless noted, the criteria for the numerator(s) are a subset of those selected for the denominator(s) below, and the selection criteria for the denominator(s) is not usually repeated in the numerator(s).

 

⑧         Denominator(s): Number of the population meeting the criteria for inclusion in the denominator of the indicators. Criteria for selection for the denominator are given based on the variables in the recode file. If no criteria is specified, the denominator is formed of all cases in the dataset.

 

⑨         Variables: Dataset(s) used to produce the indicators and the list of variables needed to calculate the indicators. If indicators are to be disaggregated, the background variables to use are not included in this list.

v???

Variables needed for the indicator

?005

Sample weight variable

 

⑩           Calculation

 

Explanation of the calculation of the indicators, typically based on the numerator(s) and denominator(s) defined above. For most indicators that are simple percentages this is numerator divided by denominator multiplied by 100. For more complicated indicators a detailed description of the calculation method is provided.

 

⑪           Handling of Missing Values

 

Explanation of how missing values and “Don’t know” responses are handled for the indicators.

 

⑫           Changes over Time

 

Brief description of the ways in which the indicators have changed over time between different phases of The DHS Program.

 

⑬           References

 

List of key references to international indicator guidelines further describing the indicators or cross-national analyses using the indicators.

 

⑭           Resources

 

                A list of resources relevant to the indicators, including links to DHS videos and webpages.

 

⑮           DHS-8 Tabulation plan: Table number reference to the DHS-8 Tabulation Plan, found at https://www.dhsprogram.com/publications/publication-DHSM6-DHS-Questionnaires-and-Manuals.cfm

 

⑯           API Indicator IDs: IDs for use with the DHS Application Programming Interface (API) described at https://api.dhsprogram.com/#/index.html. The full list of API Indicator IDs can be found at http://api.dhsprogram.com/rest/dhs/indicators?f=html

 

API link – link to present national indicator data for surveys from the API in a browser.

STATcompiler link – link to national indicator data in STATcompiler.

 

 

 

Example Indicator Page

 

①           Knowledge of HIV or AIDS

 

②           Percentage of women and men who have heard of HIV or AIDS

 

③           Definition

 

Percentage of women and men age 15-49 who have heard of HIV or AIDS.

 

④         Coverage:

⑤                         Population base: Women and men age 15-49 (IR file, MR file)

⑥                         Time period: Current status at time of survey

 

⑦         Numerator: Number of women (or men) age 15-49 who respond “Yes” that they have heard of HIV or AIDS (women: v751 = 1; men: mv751 = 1)

 

⑧         Denominator: Number of women age 15-49 or number of men (age range varies by survey)

 

⑨         Variables: IR file, MR file.

v751

Ever heard of HIV or AIDS (women)

v005

Woman’s individual sample weight

mv751

Ever heard of HIV or AIDS (men)

mv005

Man’s individual sample weight

 

⑩         Calculation

 

Numerator divided by denominator, multiplied by 100.

 

⑪         Handling of Missing Values

 

Missing values are assumed to reflect a lack of knowledge about HIV or AIDS and are therefore not included in the numerator.

 

⑫         Changes over Time

 

Prior to DHS-7, the standard questionnaires asked if women and men had “ever heard of an illness called AIDS”. Due to improved knowledge about HIV/AIDS and the decrease in AIDS cases as treatment is more common, some respondents may be more familiar with the term “HIV” than “AIDS”. The purpose of the question is to assess basic knowledge of either HIV or AIDS.

 

In DHS-8 this has become an optional question. It may be deleted in countries with near universal knowledge of HIV in which case everyone is assumed to have knowledge of HIV.

 

⑬         References

               

Bankole, A., S. Singh, V. Woog, and D. Wulf. 2004. "Risk and protection: youth and HIV/AIDS in sub-Saharan Africa.”. https://www.guttmacher.org/news-release/2004/new-resource-documents-hiv-risk-and-protection-african-youth

 

Khan, M.A.2002. "Knowledge on AIDS among female adolescents in Bangladesh: evidence from the Bangladesh demographic and health survey data." Journal of health, population and nutrition: 130-137.

 

⑭         Resources

 

⑮           DHS-8 Tabulation plan: Table 13.0.w

 

⑯           API Indicator IDs:

HA_KAID_W_HRD, HA_KAID_M_HRD

(API link, STATcompiler link)

 

 

Using Indicator Pages

 

Example code

 

The indicator pages can be used as a guide in constructing code in statistical software for tabulating indicators. Below are examples in Stata and SPSS for producing indicators based on the information in the indicator pages. Most indicators are simple percentages computed as the proportion meeting some numerator condition that is a subset of cases meeting the denominator condition. The below code provides a generic example of code, applied to the indicator page example above. The code below provides examples of simple tabulations that do not take into account the complex sample design, as well as examples that take into account the complex sample design and produce confidence intervals and standard errors for the indicators.

 

Stata

* open dataset, using type of dataset from variables list, e.g. IR file in this case

use "ZZIR62FL.dta", clear

 

* select denominator

* gen denom = ( denominator criteria )

* or denom = 1 if all cases are selected

gen denom = 1

 

* select numerator, based on subset of denominator

* gen numer = ( numerator criteria ) if (denom == 1)

gen numer = (v751 == 1) if (denom == 1)

 

* Label numerator variable

label variable numer "Knows of HIV/AIDS"

label define YesNo 1 "Yes" 0 "No"

label values numer YesNo

format numer denom %1.0f

 

* set up weight

* gen wt = weightvar / 1000000

gen wt = v005/1000000

 

* simple tabulation

tab numer [iw=wt] if (denom == 1)

 

* or by background characteristic, e.g. by region

tab v024 numer [iw=wt] if (denom == 1), row

 

* or for complex samples use svyset and svy: tab

 

* first set up strata variable - can vary from survey to survey

* example using v023, but may be a combination of v024 and v025

gen strat = v023

* or egen strat = group(v024 v025)

 

* set up svyset command

svyset v021 [pw=wt], strata(strat)

 

* tabulate using svy: tab

svy: tab numer if (denom == 1), ci se

 

* or by background characteristic

svy: tab v024 numer if (denom == 1), row ci se

 

SPSS

* open dataset, using type of dataset from variables list, e.g. IR file in this case.

get file = "ZZIR62FL.SAV".

 

* select denominator

* compute denom = ( denominator criteria ).

* or denom = 1 if all cases are selected.

compute denom = 1.

 

* select numerator, based on subset of denominator.

* if (denom = 1) numer = ( numerator criteria ).

if (denom = 1) numer = (v751 = 1).

 

* Label numerator variable.

variable labels numer "Knows of HIV/AIDS".

value labels numer 1 "Yes" 0 "No".

formats numer denom (f1.0).

 

* set up weight.

* compute wt = weightvar/1000000.

compute wt = v005/1000000.

weight by wt.

 

* filter for selected cases.

filter by denom.

 

* simple tabulation.

frequencies variables=numer.

 

* or by background characteristic, e.g. by region.

crosstabs tables=v024 by numer /count=asis /cells=row.

 

* or for complex samples use csplan and csdescriptives.

 

* first set up strata variable - can vary from survey to survey.

* example using v023, but may be a combination of v024 and v025.

compute strat = v023.

* or compute strat = v024*2 + v025.

 

* turn off weight to avoid warning message, as weight comes from csplan.

weight off.

 

* set up csplan command.

csplan analysis /plan file='DHS.csaplan'

 /planvars analysisweight=wt

 /srsestimator type=wor

 /design strata=strat cluster=v021

 /estimator type=wr.

 

* tabulate using csdescriptives.

csdescriptives /plan file='DHS.csaplan'

 /summary variables = numer

 /mean

 /statistics se cin.

 

* or by background characteristic.

csdescriptives /plan file='DHS.csaplan'

 /summary variables = numer

 /subpop table=v024

 /mean

 /statistics se cin.

 

Numerator and denominator conditions

 

The indicator pages provide generic logic for the definitions of the numerators and denominators that is not specific to any particular software and needs to be translated to the software to be used for analysis. Below is a brief explanation of the generic code used and how to implement it in Stata or SPSS.

 

Generic code

Stata code

SPSS code

var = n

var == n

var = n or var EQ n

var ≠ n

var != n

var <> n or var NE n

var ≠ a,b

var != a & var != b

var <> a AND var <> b

var in a,b,c

var == a | var == b | var == c

or

inlist(var,a,b,c)

var = a OR var = b OR var = c

or

any(var,a,b,c)

var in x:y

var >= x & var <= y

or

inrange(var,x,y)

var >= x AND var <= y

or

range(var,x,y)

var in a,b,c,x:y

var == a | var == b | var == c | (var >= x & var <= y)

or

inlist(var,a,b,c) | inrange(var,x,y)

var = a OR var = b OR var = c OR (var >= x AND var <= y)

or

any(var,a,b,c) OR range(var,x,y)

x <= var <= y

var >= x & var <= y

or

inrange(var,x,y)

var >= x AND var <= y

or

range(var,x,y)

x < var < y

var > x & var < y

var > x AND var < y

condx & condy

e.g.

vara = x & varb = y

condx & condy

e.g.

vara == x & varb == y

condx & condy or

condx AND condy

e.g.

vara = x & varb = y

or

vara = x AND varb = y

 

condx or condy

e.g.

vara = x or varb = y

condx | condy

e.g.

vara == x | varb == y

condx | condy or

condx OR condy

e.g.

vara = x | varb = y

or

vara = x OR varb = y

 

any of vara – x = 1

vara = 1 | varb = 1 | varc = 1 | ... varx = 1

vara = 1 | varb = 1 | varc = 1 | ... varx = 1