İstatistiksel Eşleme Metodolojisi ve Sağlıkta Kullanımı ile İlgili Ampirik Bir Değerlendirme

Year 2021, Volume: 11 Issue: 2, 129 - 136, 07.05.2021

İsmet Doğan , Nurhan Dogan

https://doi.org/10.33631/duzcesbed.784688

Abstract

Amaç: Bu çalışmanın amacı, istatistiksel eşleme yöntemlerini değerlendirmek ve Rubin'in istatistiksel eşleme yöntemini sağlık alanında örnek bir uygulama ile tanıtmaktır. Gereç ve Yöntemler: İstatistiksel eşleme yapay mikro veri setleri oluşturmak için bir yöntem olarak son yıllarda giderek artan bir popülariteye sahiptir. İstatistiksel eşleme, bir araştırmada aynı anda gözlenmemiş (Y,Z) rasgele değişken çiftinden elde edilen taslak bilgi problemini ele almaktadır. Gerçekte Y ve Z birbirinden bağımsız iki farklı araştırmada birbirleri ile örtüşmeyen gözlem birimlerinin oluşturduğu kümelerden elde edilmektedir. Ancak iki araştırmada aynı X değişkeni ortaklaşa gözlenmektedir. İstatistiksel eşleme yöntemleri iki farklı veri kümesinden elde edilen bilginin birleştirilmesini hedeflemektedir. Bulgular: Eşleme işleminde hangi veri setinin alıcı hangisinin donör veri seti olacağı ve kohort değişken kullanmanın söz konusu olup olmayacağı önem arz etmektedir. Çünkü bunlar hem eşlemede hem de eşleme sonucunda hesaplanan uzaklık ölçüsünün değerinin belirlenmesinde belirleyici olmaktadır. Özellikle kohort değişken kullanılması uzaklık ölçüsünün değerini minimum olmaktan uzaklaştırmaktadır. Sonuç: Rubin tarafından önerilen yöntem, diğer yaklaşımlara göre oldukça iyi sonuçlar vermesine rağmen en iyi yöntem veya yöntemler konusunda fikir birliği bulunmamaktadır. En iyi yöntem veya yöntemlere ilişkin görüş birliği bulunmadığından kısıtlı ve kısıtsız yöntemler halen kullanılmaktadır.

Keywords

İstatistiksel eşleme, veri karıştırma, veri birleştirme, dosya birleştirme, yapay eşleme, Mahalanobis uzaklığı

Supporting Institution

-

Project Number

-

Statistical Matching Methodology and An Empirical Evaluation of Its Use in Health

Abstract

Aim: The aim of this study is to evaluate statistical matching methods and introduce Rubin's statistical matching method with an exemplary application in the field of health. Material and Methods: Statistical matching has received increasing popularity in the last decades as a method of creating synthetic microdata sets. Statistical matching tackles the problem of drawing information on a pair of random variables (Y,Z) which have not been observed jointly in one sample survey. In fact, Z and Y are available in two distinct and independent surveys whose sets of observed units are non overlapping. The two surveys observe also some common variables X. Statistical matching techniques are aimed at combining information available in two distinct datasets. Results: In the matching process, it is important that which data set will be the recipient and which will be the donor data set and whether it is possible to use cohort variables. Because these are decisive both in matching and determining the value of the distance measure calculated as a result of the matching. In particular, the use of cohort variables makes the value of the distance measure away from being minimum. Conclusions: Although the method proposed by Rubin gives very good results compared to other approaches, there is no consensus on the best method or methods. No consensus regarding the best method or methods has developed; both constrained and unconstrained methods are still being used.

Keywords

Statistical matching, data fusion, data merging, file concatenation, synthetical matching, Mahalanobis distance

Project Number

-

References

• Marella D, Pfeffermann D. Matching information from two independent informative samples. J Stat Plan Infer, 2019; 203: 70-81.
• Rubin DB. Statistical matching using file concatenation with adjusted weights and multiple imputations. J Bus Econ Stat, 1986; 4(1): 87-94.
• Barry RA, Stewart WH, Turner JS. An empirical evaluation of statistical matching methodologies. Dallas, Texas: Working Paper, Southern Methodist University; 1982. p. 4-5.
• Barry, JT. An investigation of statistical matching. J Appl Stat, 1988; 15(3): 275-283.
• Willenborg L, Heerschap H. Statistics Methods: Matching. 1^st ed. Hague/Heerlen: Statistics Netherlands; 2012. p. 40-46.
• Goel PK, Ramalingam T. The Matching Methodology: Some Statistical Properties. 1^st ed. Berlin Heidelberg: Springer-Verlag; 1989. p. 1-13.
• European Union (eurostat). Statistical matching: a model based approach for data integration. Methodologies and Working Papers, Luxembourg: Publications Office of the European Union; 2013. p. 7-19.
• Radner DB, Allen R, Gonzalez ME, Jabine TB, Muller HJ. Report on exact and statistical matching techniques. Statistical Policy Working Paper 5, U.S. Department of Commerce, Washington, DC., U.S.: Government Printing Office; 1980. p. 15-34.
• Ahfock D, Pyne S, Lee SX, McLachlan GJ. Partial identification in the statistical matching problem. Comput Stat Data Anal, 2016; 104: 79-90.
• D’Orazio M, Zio MD, Scanu M. Statistical Matching: Theory and Practice. 1^st ed. West Sussex, England: John Wiley & Sons Ltd.; 2006. p. 1-2.
• National Statistics. National Statistics Code of Practice Protocol on Data Matching, London: A National Statistics Publication; 2004. p.15-16.
• Moriarity CL, Scheuren F. Statistical matching: Pitfalls of current procedures. Proceedings of the Annual Meeting of the American Statistical Association, Atlanta, Georgia: 2001. p. 5-10.
• Rasner A, Frick JR, Grabka, MM. Statistical matching of administrative and survey data: An application to wealth inequality analysis. Sociol Method Res, 2013; 42(2): 192-224.
• Rodgers WL. An evaluation of statistical matching. J Bus Econ Stat, 1984; 2(1): 91-102.
• Endres E, Augustin T. Statistical matching of discrete data by Bayesian networks. Proceedings of the Eighth International Conference on Probabilistic Graphical Models, PMLR, 2016; 52: 159-170.
• Rassler S. Statistical Matching: A Frequentist Theory, Practical Applications and Alternative Bayesian Approaches. 1'st ed. New York: Springer Science+Business Media, LLC; 2002. p. 2-3.
• Kadane JB. Some statistical problems in merging data files. J Off Stat, 2001; 17(3): 423-433.
• Wiest M, Kutscher T, Willeke J, Merkel J, Hoffmann M, Kaufmann-Kuchta K, Widany S. The potential of statistical matching for the analysis of wider benefits of learning in later life. European Journal for Research on the Education and Learning of Adults, 2019; 10 (3): 291-306.
• Conti PL, Marella D, Neri A. Statistical matching and uncertainty analysis in combining household income and expenditure data. Stat Methods Appl, 2017; 26: 485–505.
• Gessendorfer J, Beste J, Drechsler J, Sakshaug JW. Statistical matching as a supplement to record linkage: A valuable method to tackle nonconsent bias? J Off Stat, 2018; 34 (4): 909–933.
• Conway A, Rolley JX, Fulbrook P, Page K, Thompson DR. Improving statistical analysis of matched case–control studies. Res Nurs Health, 2013; 36 (3): 320–324.
• Faresjö T, Faresjö A. To match or not to match in epidemiological studies-same outcome but less power. Int J Environ Res Public Health, 2010; 7 (1): 325-332.
• Kim K, Park M. Statistical micro matching using a multinomial logistic regression model for categorical data. CSAM, 2019; 26 (5): 507–517.
• Moriarity CL, Scheuren F. Regression based statistical matching: Recent developments. Proceedings of the Section on Survey Research Method, American Statistical Association, 2004; p. 4050-4057.
• Waal Ton de. Statistical matching: Experimental results and future research questions. Den Haag: CBS. 2015. doi: 10.13140/RG.2.1.1969.4161.
• Perchinunno P, Mongelli L, d'Ovidio F. Statistical matching techniques in order to plan interventions on socioeconomic weakness: An Italian case. Socio Econ Plan Sci, 2020; 71. 100836. 10.1016/j.seps.2020.100836.
• Alpman A, Gardes F, Thiombiano N. Statistical Matching for Combining Time-Use Surveys with Consumer Expenditure Surveys: An Evaluation on Real Data. Documents de travail du Centre d'Economie de la Sorbonne 17024, Université Panthéon-Sorbonne (Paris 1), Centre d'Economie de la Sorbonne. 2017. ffhalshs-01529699f.
• Namazi-Rad MR, Tanton R, Steel D, Mokhtarian P, Das S. An unconstrained statistical matching algorithm for combing individual and household level geo-specific census and survey data, National Institute for Applied Statistics Research Australia, University of Wollongong, Working Paper 01-16, 2016, 35.
• Schleicher J, Eklund JD, Barnes M, Geldmann J, Oldekop JA, Jones JPG. Statistical matching for conservation science. Conserv Biol, 2019; doi: 10.1111/cobi.13448.
• Iacus SM, King G, Porro G. A theory of statistical inference for matching methods in causal research. Political Analysis, 2019; 27(1): 46-68.
• Stuart EA. Matching methods for causal inference: A review and a look forward. Stat Sci, 2010; 25(1): 1-21.
• D’Orazio M, Zio MD, Scanu M. Old and new approaches in statistical matching when samples are drawn with complex survey designs. Proceedings of the 45th Riunione Scientifica della Societa Italiana di Statistica, Padua, Italy: 2010. p.1-10.