The construction of the SIGI and its sub-indices involved three main stages, from updating the country profiles and constructing the database to producing the composite index and its sub-indices, which provide the SIGI ranking. For more information, please refer to the methodological background paper or the technical note.

Step 1: Updating the country profiles
  • Updating the country profiles
    The SIGI country profiles contain fully-referenced qualitative information relevant to discriminatory social institutions. They were drafted by gender and development experts following a standardised structure and format to ensure comparability across countries. Then, the country profiles were validated by external gender experts or institutions with knowledge of the policy and legal landscape for gender equality and women’s rights at a national level (see list of experts).
  • Assigning a score to qualitative variables
    Scoring the qualitative information is the process by which the qualitative information is given a discrete value corresponding to each category it can assume. Categorical variables are based on a scale of three categories (0, 0.5 and 1) or five categories (0, 0.25, 0.5, 0.75 and 1), where 0 represents no discrimination and 1 represents a high level of discrimination (see the coding manual that sets out the detailed coding guidelines for categorising the qualitative SIGI variables and provides examples).
Step 2: Building the Gender, Institutions and Development Database
  • Collecting quantitative information
    Quantitative gender statistics are collected from national and international data sources (see the indicator dashboard describing the definition and the sources)
  • Constructing indicators
    Some indicators are based on one variable while others on several. In the latter case, the indicator is the average of its available variables. For example:
    Parental authority =1/2(Parental authority during marriage+Parental authority after divorce)Where data is available for at least one variable of an indicator, the score is calculated based on the variables available.
Step 3: Aggregating indicators to build the sub-indices
  • Measuring association between indicators
    Each sub-index combines indicators that are assumed to belong to one dimension of discrimination in social institutions. The statistical association between the indicators is tested using a Kendall Tau b rank correlation analysis and a multiple joint correspondence analysis (MCA).
  • Constructing the sub-indices
    The sub-indices aim to provide a summary measure of each dimension of discrimination. Constructing a sub-index consists of aggregating the indicators with a reasonable weighting scheme through a polychoric principal component analysis (PCA).
Step 4: Computing the SIGI

The SIGI is a composite indicator built as an unweighted average of a non-linear function of the sub-indices:



Why use polychoric PCA to build sub-indices?
Polychoric PCA has numerous advantages for constructing a sub-index:
• aggregating continuous and categorical indicators;
• analysing the underlying structure of the data, thus avoiding poor construction and misinterpretation
• grouping individual indicators according to their degree of correlation without presuming a specific underlying structure of the data but allowing the data to speak for themselves
• identifying endogenous and reasonable weighting schemes
• summarising the underlying trend and common information captured by raw variables by correcting for statistical bias and redundancy and by preserving the maximum possible proportion of the total variation in the original data set.
Why square each SIGI sub-index?
The non-linear function arises because:
• The partial compensation means that very high inequality in one dimension can be only partially offset by low inequality in another dimension.
• The SIGI measures gender inequalities that correspond to deprivation experienced by the affected women; and that deprivation increases more than proportionally when inequality increases.
• The SIGI has an aversion to high values of sub-indices: reducing inequalities from 1 to 0.75 is not equivalent to reducing from 0.25 to 0.
Why are the sub-indices equally weighted?
Equal weights for each sub-index offer two benefits:
• Each dimension of discriminatory social institutions has equal value.
• No dimension is more important than another in terms of deprivation experienced by women.
How is the SIGI classified?
The SIGI classification clusters 108 countries into five levels of discrimination in social institutions: very low, low, medium, high and very high. It is based on the Jenks Natural Breaks Classification. This method of classifying data optimally arranges values into the five levels, or classes. It aims to minimise the average deviation from the class mean, while maximising the deviation from the means of the other classes. Hence, this method reduces the variance within classes and maximises the variance between classes.