DOI: https://doi.org/10.14710/j.gauss.15.2.391-401

PERBANDINGAN METODE IMPUTASI DATA KATEGORIK PADA PERILAKU KEUANGAN DIGITAL: SIMULASI MCAR DAN MAR MIKRODATA GLOBAL FINDEX INDONESIA 2025

*Dina Eka Putri  -  Department of Statistics, Universitas Mataram,, Indonesia
Adis Tia Juli Agil Asri  -  Departemen Statistika, Fakultas Sains dan Matematika, Universitas Mataram, Indonesia
Arni Nazira  -  Departemen Statistika, Fakultas Sains dan Matematika, Universitas Mataram, Indonesia
Open Access Copyright 2026 Jurnal Gaussian under http://creativecommons.org/licenses/by-nc-sa/4.0.

Citation Format:
Abstract
Large-scale surveys such as the Global Findex Database are crucial for monitoring financial inclusion, yet their microdata often contain missing values that can bias empirical analyses and weaken inference. This study evaluates three imputation methods for three categorical digital financial behavior variables (Fin3, Con12, and Con26) in the 2025 Global Findex microdata for Indonesia. Missing values were artificially introduced into a complete-case reference dataset (n = 407) under Missing Completely at Random (MCAR) and Missing at Random (MAR) mechanisms at rates of 10%, 20%, and 30%, with 30 Monte Carlo replications per scenario. Mode imputation, Multiple Imputation by Chained Equations (MICE), and random forest imputation (missForest) were compared using the Proportion of False Classification (PFC). Because all target variables are categorical, the simple baseline was mode imputation; no numerical variable was imputed using the mean. Mode imputation yielded the lowest PFC in 17 of 18 combinations of variable, missing rate, and mechanism, followed generally by random forest and then MICE. These findings suggest that, for categorical and imbalanced survey data, simple distribution-based methods can be competitive with model-based approaches, partly because of majority-class dominance and limited predictive structure. The conclusion is limited to the PFC criterion and the scenarios examined.
Fulltext View|Download Email colleagues
Keywords: Global Findex; Imputasi Data Hilang; Data Survei Kategorikal; MICE; Random Forest; Inklusi Keuangan Digital

Article Metrics:

Article Info
Section: Articles
Language : ID
  1. Abdullah, N.A. et al. (2026) “Assessing imputation techniques for missing data in small and multicollinear datasets: insights from craniofacial morphometry,” BMC Medical Research Methodology, 26(1), hal. 54. Tersedia pada: https://doi.org/10.1186/s12874-025-02762-4
  2. Allison, P.D. (2001) Missing Data. Thousand Oaks, CA: Sage Publications
  3. van Buuren, S. (2018) Flexible Imputation of Missing Data. 2 ed. Boca Raton, FL: Chapman & Hall/CRC
  4. van Buuren, S. dan Groothuis-Oudshoorn, K. (2011) “mice: Multivariate imputation by chained equations in R,” Journal of Statistical Software, 45(3), hal. 1–67. Tersedia pada: https://doi.org/10.18637/jss.v045.i03
  5. Cohen, J. (1960) “A coefficient of agreement for nominal scales,” Educational and Psychological Measurement, 20(1), hal. 37–46. Tersedia pada: https://doi.org/10.1177/001316446002000104
  6. Demirgüç-Kunt, A. et al. (2022) The Global Findex Database 2021: Financial Inclusion, Digital Payments, and Resilience in the Age of COVID-19. Washington, DC: World Bank. Tersedia pada: https://doi.org/10.1596/978-1-4648-1897-4
  7. Farah, A.A., Nor, B.A. dan Musse, O.S.H. (2026) “Financial inclusion and green economic growth a systematic review of SCOPUS indexed studies from 2015 to 2025,” Discover Environment, 4, hal. 37. Tersedia pada: https://doi.org/10.1007/s44274-026-00550-5
  8. Jolani, S. (2025) “Hierarchical imputation of categorical variables in the presence of systematically and sporadically missing data,” Research Synthesis Methods, 16, hal. 729–757. Tersedia pada: https://doi.org/10.1017/rsm.2025.10017
  9. Khan, N. et al. (2024) “Financial inclusion as a catalyst for economic growth: evidence from selected developing countries,” Cogent Social Sciences, 10(1), hal. 2387907. Tersedia pada: https://doi.org/10.1080/23311886.2024.2387907
  10. Kuhn, M. (2008) “Building predictive models in R using the caret package,” Journal of Statistical Software, 28(5), hal. 1–26. Tersedia pada: https://doi.org/10.18637/jss.v028.i05
  11. Li, J. et al. (2024) “Comparison of the effects of imputation methods for missing data in predictive modelling of cohort study datasets,” BMC Medical Research Methodology, 24, hal. 41. Tersedia pada: https://doi.org/10.1186/s12874-024-02173-x
  12. Little, R.J.A. dan Rubin, D.B. (2019) Statistical Analysis with Missing Data. 3 ed. Hoboken, NJ: Wiley
  13. Morris, T.P., White, I.R. dan Crowther, M.J. (2019) “Using simulation studies to evaluate statistical methods,” Statistics in Medicine, 38(11), hal. 2074–2102. Tersedia pada: https://doi.org/10.1002/sim.8086
  14. Prihatni, R. et al. (2024) Analisis Literasi Keuangan dan Inklusi Keuangan di Indonesia: Strategi untuk Meningkatkan Kesejahteraan Finansial Masyarakat. 1 ed. Diedit oleh N. Mayasar. Bandung: Penerbit Widina Media Utama
  15. R Core Team (2026) R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing. Tersedia pada: https://www.R-project.org/
  16. Rubin, D.B. (1976) “Inference and missing data,” Biometrika, 63(3), hal. 581–592. Tersedia pada: https://doi.org/10.1093/biomet/63.3.581
  17. Rubin, D.B. (1987) Multiple Imputation for Nonresponse in Surveys. New York: Wiley
  18. Rácz, A. dan Gere, A. (2025) “Comparison of missing value imputation tools for machine learning models based on product development cases studies,” LWT - Food Science and Technology, 221, hal. 117585. Tersedia pada: https://doi.org/10.1016/j.lwt.2025.117585
  19. Sarma, M. (2008) Index of Financial Inclusion. New Delhi. Tersedia pada: https://icrier.org/publications/index-of-financial-inclusion/
  20. Stekhoven, D.J. dan Bühlmann, P. (2012) “MissForest — non-parametric missing value imputation for mixed-type data,” Bioinformatics, 28(1), hal. 112–118. Tersedia pada: https://doi.org/10.1093/bioinformatics/btr597
  21. Thurow, M., Dumpert, F., Ramosaj, B. dan Pauly, M. (2021) “Goodness (of fit) of imputation accuracy: the GoodImpact analysis,” arXiv preprint arXiv:2101.07532. Tersedia pada: https://doi.org/10.48550/arXiv.2101.07532
  22. Wabina, R.S. et al. (2025) “Uncertainty-aware approach for multiple imputation using conventional and machine learning models: a real-world data study,” Journal of Big Data, 12. Tersedia pada: https://doi.org/10.1186/s40537-025-01136-3
  23. World Bank (2025) The Global Findex Database 2025: Connectivity and Financial Inclusion in the Digital Economy, World Bank. Tersedia pada: https://www.worldbank.org/en/publication/globalfindex

Last update:

No citation recorded.

Last update:

No citation recorded.