Indicators and Maturity of Open Science Practices in Tanzania: A Fuzzy Delphi Method

The Fuzzy Delphi Method (FDM) generated multiple categories of analytical outputs corresponding to different stages of the indicator validation process. The first dataset comprised the initial pool of 58 indicators (I1–I58), which were developed through literature review, policy analysis, and synthesis of existing RDM frameworks. These indicators were subsequently organised into nine conceptual dimensions representing major domains of RDM practices. The second analytical output consisted of the linguistic calibration framework, where expert responses expressed through linguistic terms were converted into triangular fuzzy numbers (TFNs). This transformation enabled subjective judgments to be mathematically processed within the fuzzy environment. Following data collection, expert evaluations were aggregated to generate fuzzy scores and defuzzified values for each indicator. These values reflected the collective assessment of indicator importance across the expert panel. In parallel, consensus measures were computed to determine the degree of agreement among experts for each indicator. The resulting outputs formed the empirical basis for applying the data-driven threshold and integrated indicator selection rule. Subsequently, aggregated fuzzy scores were computed across the nine dimensions to determine their relative importance within the overall RDM framework. The dimensional aggregation results enabled comparative interpretation of the structural performance of different RDM domains. Finally, the integrated decision process combining the empirically derived threshold and expert consensus evaluation produced the final validated indicator set. This output classified indicators into retained and rejected categories based on their ability to satisfy both statistical and epistemic validation criteria. To improve readability and avoid repetitive presentation of identical computational procedures, only aggregated and decision-oriented tables are presented in the main findings section, while detailed intermediate computational outputs and extended fuzzy aggregation matrices are deposited in an open-access repository for transparency and reproducibility purposes.