Advanced search
Start date
Betweenand
Related content
(Reference retrieved automatically from Web of Science through information on FAPESP grant and its corresponding number as mentioned in the publication by the authors.)

A propensity score approach in the impact evaluation on scientific production in Brazilian biodiversity research: the BIOTA Program

Full text
Author(s):
Colugnati, Fernando A. B. [1, 2] ; Firpo, Sergio [3] ; Drummond de Castro, Paula F. [1] ; Sepulveda, Juan E. [4] ; Salles-Filho, Sergio L. M. [1]
Total Authors: 5
Affiliation:
[1] Univ Estadual Campinas, Inst Geosci, Dept Sci & Technol Policy, Lab Studies Org Res & Innovat GEOPI, BR-13083970 Campinas, SP - Brazil
[2] Univ Fed Juiz de Fora, Dept Clin Med, BR-36036330 Juiz De Fora, MG - Brazil
[3] Sao Paulo Sch Econ FGV, BR-01332000 Sao Paulo - Brazil
[4] Univ Estadual Campinas, Inst Econ, BR-13083857 Campinas, SP - Brazil
Total Affiliations: 4
Document type: Journal article
Source: SCIENTOMETRICS; v. 101, n. 1, p. 85-107, OCT 2014.
Web of Science Citations: 6
Abstract

Evaluation has become a regular practice in the management of science, technology and innovation (ST\&I) programs. Several methods have been developed to identify the results and impacts of programs of this kind. Most evaluations that adopt such an approach conclude that the interventions concerned, in this case ST\&I programs, had a positive impact compared with the baseline, but do not control for any effects that might have improved the indicators even in the absence of intervention, such as improvements in the socio-economic context. The quasi-experimental approach therefore arises as an appropriate way to identify the real contributions of a given intervention. This paper describes and discusses the utilization of propensity score (PS) in quasi-experiments as a methodology to evaluate the impact on scientific production of research programs, presenting a case study of the BIOTA Program run by FAPESP, the State of So Paulo Research Foundation (Brazil). Fundamentals of quasi-experiments and causal inference are presented, stressing the need to control for biases due to lack of randomization, also a brief introduction to the PS estimation and weighting technique used to correct for observed bias. The application of the PS methodology is compared to the traditional multivariate analysis usually employed. (AU)