The effects of outdated data and outliers on Kenya's 2019 Global Food Security Index score and rank

Show simple item record

dc.contributor.advisor Hendriks, Sheryl L.
dc.contributor.postgraduate Atieno, Prisca
dc.date.accessioned 2021-04-22T09:04:50Z
dc.date.available 2021-04-22T09:04:50Z
dc.date.created 2021-09
dc.date.issued 2021-04
dc.description Mini Dissertation (MSc Agric (Agricultural Economics))--University of Pretoria, 2021. en_ZA
dc.description.abstract While composite indicators are considered robust in measuring food security, outdated data and outliers challenge their reliability. Outdated data can occur when national databases are not frequently updated while outliers are extremely small or large values in a study. Outdated data could be referred to as missing current data in composite indicators used for annual benchmarking exercises, where data must be frequently updated. Besides hindering useful information within an index, outdated data could also result in outliers in a database, especially when the outdated or missing current data are imputed by estimation. Studies that have assessed the robustness of composite indicators highlight that outdated data and outliers could bias results, thereby hindering an index's reliability. However, depending on the methods used when constructing a composite indicator, some methods can be considered robust even with the presence of outliers in a data point. Outdated national data could hinder countries from tracking the progress of international, national or regional commitments, such as the Sustainable Development Goals, while outliers could act as an unintended benchmark. This study assessed the impacts of outdated data and outliers on Kenya's scores and rankings in the Global Food Security Index (GFSI). The study objective was achieved by assessing Kenya's performance in the 2019 GFSI result before and after removing outliers from the GFSI data points and updating Kenya's outdated indicators. Winsorisation was used to remove the outliers from the GFSI database, while the Spearman correlation and Paired t-tests were used to test for the statistical significance of the outdated data and outliers. The study revealed that while Kenya's 2019 GFSI database did not have outliers, outliers in other countries' data points impacted Kenya's score and rank. For example, the winsorisation of outliers for other countries reduced Kenya's 2019 overall GFSI score by six points. Moreover, thirteen indicators in Kenya's 2019 GFSI database were found to be outdated. However, despite Kenya's score improving from updating the outdated data, the impact was minimal to increase the GFSI's mean score for all countries. That is, updating Kenya's outdated indicators was found not to differ significantly from zero. The study concluded that Kenya's score and rank in the 2019 GFSI were affected by the outdated data in Kenya's database and outliers in other countries' data. The study, therefore, recommended that Kenya should update its national database and allow open access to the national data while the GFSI should identify and remove outliers from the data points. en_ZA
dc.description.availability Unrestricted en_ZA
dc.description.degree MSc Agric (Agricultural Economics) en_ZA
dc.description.department Agricultural Economics, Extension and Rural Development en_ZA
dc.description.sponsorship Mastercard Foundation Scholarship en_ZA
dc.identifier.citation * en_ZA
dc.identifier.other u18394613 en_ZA
dc.identifier.uri http://hdl.handle.net/2263/79569
dc.language.iso en en_ZA
dc.publisher University of Pretoria
dc.rights © 2019 University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria.
dc.subject UCTD en_ZA
dc.subject Food security en_ZA
dc.title The effects of outdated data and outliers on Kenya's 2019 Global Food Security Index score and rank en_ZA
dc.type Mini Dissertation en_ZA


Files in this item

This item appears in the following Collection(s)

Show simple item record