Abstract:
While composite indicators have become a valuable tool in policymaking, benchmarking and public communication processes, outliers and outdated data challenge their reliability. Outliers are large or small values in a database that could act as an unintended benchmark. At the same time, outdated data can arise when databases are not frequently updated and current data are missing in annual benchmarking exercises. Outdated data and outliers can render composite indicators less reliable and lead to misleading results and unreliable benchmarking. Outdated data could also hinder countries from tracking the progress of national, international, regional or global commitments, such as the Malabo commitments and SDGs.
This study assessed the effects of outdated data and outliers on Zambia’s 2019 Global Food Security Index (GFSI) score and ranking. The study compared Zambia’s score and rank relative to other countries in the Global Food Security Index before and after updating outdated data and winsorisation of outliers found in the 2019 GFSI dataset. Updated data was obtained from alternative sources to calculate updated scores and rankings. Winsorisation removed outliers and replaced them with the same indicator's net highest or smallest values in the database. Paired t-test and Spearman rank correlation tested the effects of outdated data and outliers.
The study found that the 2019 Global Food Security Index data had ten out of 34 indicators with outlier values from 16 countries. Zambia had an outlier in public expenditure on agricultural research and development indicator. The study also revealed that Zambia had 14 indicators that used outdated data in the 2019 GFSI results. A statistically significant difference was found between the scores after the winsorisation of outliers for the affordability and availability dimensions of and the overall scores. However, despite Zambia’s score and rank improving after updating outdated data, the scores and rankings were not statistically significant.
The study concluded that outliers and outdated data in the 2019 Global Food Security Index impacted Zambia’s scores and ranking. The study recommended that outliers be identified and removed from composite indicators to avoid unreliable benchmarking settings by policymakers. The study also recommended that Zambia enhance timely quality data collection to update databases and improve the food security score and ranking in different regional and global indexes.
