Piloting text-mining in a mixed-method evaluation on human rights

Management propositions for small public administrations

Dados Bibliográficos

ID6423911
AutoresNea-Mari Heinonen (0009-0000-3687-8227, Ministry for Foreign Affairs of Finland, autor correspondente)
Ano2024
Volume31
Fascículo3
Páginas352-368
Data de publicação2024-11-28
Peer ReviewedSim
Open AccessSim
TipoARTICLE
PeriódicoEvaluation (JOURNAL)
Identificadores do periódicoISSN: 1356-3890 • E-ISSN: 1461-7153
EditoraSAGE Publishing (PUBLISHER • US)
DOI10.1177/13563890241286508
OpenAlexW4404867636
IdiomaEN
Citações recebidas1
Referências citadas30

Text-mining was piloted in an evaluation of the Human Rights-Based Approach in Finland’s development cooperation. Textual and numerical evidence was extracted from project documents using rules-based, machine learning and topic modelling techniques. This practice-oriented article exemplifies considerations for evaluation managers from a civil servant’s perspective. It discusses obstacles to engaging with these techniques identified by literature. It demonstrates the feasibility of piloting new technologies for public administrations with limited human resources, emphasising guidance from the evaluation manager, staff training and early collaboration with data scientists and partners. Text-mining offers new ways of generating and analysing data. The limitations call for mixed-methods, evaluative judgement by evaluators and human oversight by the evaluation manager. Although resource-intensive, employing these methods does not require one to become a data scientist but to understand what kind of evidence can be generated. Management propositions are provided

Business · Human rights · Political science · Public management · Public relations · Evaluation and Performance Assessment · Law · Public Administration