New Confidence Intervals and Bias Comparisons Show That Maximum Likelihood Can Beat Multiple Imputation in Small Samples

Structural Equation Modeling

Volumn 23, Issue 3, 2016, Pages 422-437

  von Hippel, Paul T ^a  

^a UNIVERSITY OF TEXAS   (United States)

Author keywords

confidence intervals; finite samples; maximum likelihood; multiple imputation; small samples

Indexed keywords

FISHER INFORMATION MATRIX; MAXIMUM LIKELIHOOD; SAMPLING; STATISTICAL TESTS;

BIVARIATE NORMAL; CONFIDENCE INTERVAL; FINITE SAMPLES; FULL INFORMATION MAXIMUM LIKELIHOODS; INCOMPLETE DATA; MULTIPLE IMPUTATION; POINT ESTIMATE; SMALL SAMPLES;

MAXIMUM LIKELIHOOD ESTIMATION;

EID: 84945218777     PISSN: 10705511     EISSN: 15328007     Source Type: Journal    
DOI: 10.1080/10705511.2015.1047931     Document Type: Article

References (53)

1. P.D.Allison, (2002). Missing data. Thousand Oaks, CA: Sage.
2. T.W.Anderson, (1957). Maximum likelihood estimates for a multivariate normal distribution when some observations are missing. Journal of the American Statistical Association, 52, 200–203. doi:10.1080/01621459.1957.10501379
3. R.R.Andridge,, & R.J.A.Little, (2010). A review of hot deck imputation for survey non‐response. International Statistical Review, 78(1), 40–64. doi:10.1111/j.1751-5823.2010.00103.x
4. J.Barnard,, & D.B.Rubin, (1999). Small-sample degrees of freedom with multiple imputation. Biometrika, 86, 948–955. doi:10.1093/biomet/86.4.948
5. S.A.Barnes,, S.R.Lindborg,, & J.W.Seaman, (2006). Multiple imputation techniques in small sample clinical trials. Statistics in Medicine, 25, 233–245. doi:10.1002/sim.2231
6. T.E.Bodner, (2008). What improves with increased missing data imputations? Structural Equation Modeling, 15, 651–675. doi:10.1080/10705510802339072
7. Q.Chen,, & J.G.Ibrahim, (2013). A note on the relationships between multiple imputation, maximum likelihood and fully Bayesian methods for missing responses in linear regression models. Statistics and Its Interface, 6, 315–324. doi:10.4310/SII.2013.v6.n3.a2
8. J.E.Cohen, (2004). Economic perceptions and executive approval in comparative perspective. Political Behavior, 26(1), 27–43.
9. H.Demirtas,, S.A.Freels,, & R.M.Yucel, (2008). Plausibility of multivariate normality assumption when multiply imputing non-Gaussian continuous outcomes: A simulation assessment. Journal of Statistical Computation & Simulation, 78(1), 69–84. doi:10.1080/10629360600903866
10. H.Demirtas,, & D.Hedeker, (2008). Imputing continuous data under some non-Gaussian distributions. Statistica Neerlandica, 62, 193–205. doi:10.1111/j.1467-9574.2007.00377.x
11. N.R.Draper,, & R.C.V.Nostrand, (1979). Ridge regression and James-Stein estimation: Review and comments. Technometrics, 21, 451–466. doi:10.1080/00401706.1979.10489815
12. J.W.Graham,, A.E.Olchowski,, & T.D.Gilreath, (2007). How many imputations are really needed? Some practical clarifications of multiple imputation theory. Prevention Science, 8, 206–213. doi:10.1007/s11121-007-0070-9
13. J.S.Granberg-Rademacker, (2007). A comparison of three approaches to handling incomplete state-level data. State Politics & Policy Quarterly, 7(3), 325–338.
14. D.M.Hawkins,, & R.A.J.Wixley, (1986). A note on the transformation of chi-squared variables to normality. The American Statistician, 40, 296–298. doi:10.2307/2684608
15. Y.He,, & T.E.Raghunathan, (2012). Multiple imputation using multivariate gh transformations. Journal of Applied Statistics, 39, 2177–2198. doi:10.1080/02664763.2012.702268
16. D.F.Heitjan,, & S.Basu, (1996). Distinguishing “missing at random” and “missing completely at random”. The American Statistician, 50, 207–213. doi:10.2307/2684656
17. K.Y.Hogarty,, C.V.Hines,, J.D.Kromrey,, J.M.Ferron,, & K.R.Mumford, (2005). The quality of factor solutions in exploratory factor analysis: The influence of sample size, communality, and overdetermination. Educational and Psychological Measurement, 65, 202–226. doi:10.1177/0013164404267287
18. Hoogendoorn, A. (2009, November 18). Mulitple [sic] imputation in a very simple situation: Just two variables. Retrieved from http://www.mail-archive.com/impute@listserv.it.northwestern.edu/msg00455.html
19. K.Kelley,, & S.E.Maxwell, (2003). Sample size for multiple regression: Obtaining regression coefficients that are accurate, not simply significant. Psychological Methods, 8, 305–321. doi:10.1037/1082-989X.8.3.305
20. J.K.Kim, (2004). Finite sample properties of multiple imputation estimators. The Annals of Statistics, 32, 766–783. doi:10.1214/009053604000000175
21. S.Kunovich,, & P.M.Paxton, (2003). Women’s political representation: The importance of ideology. Social Forces, 82(1), 87–113. doi:10.1353/sof.2003.0105
22. S.Lipsitz,, M.Parzen,, & L.P.Zhao, (2002). A degrees-of-freedom approximation in multiple imputation. Journal of Statistical Computation and Simulation, 72, 309–318. doi:10.1080/00949650212848
23. R.J.A.Little, (1976). Inference about means from incomplete multivariate data. Biometrika, 63, 593–604. doi:10.1093/biomet/63.3.593
24. R.J.A.Little, (1988). Approximately calibrated small sample inference about means from bivariate normal data with missing values. Computational Statistics & Data Analysis, 7, 161–178. doi:10.1016/0167-9473(88)90090-4
25. R.J.A.Little, (1992). Regression with missing X’s: A review. Journal of the American Statistical Association, 87, 1227–1237.
26. R.J.A.Little,, & D.B.Rubin, (1989). The analysis of social science data with missing values. Sociological Methods & Research, 18, 292–326. doi:10.1177/0049124189018002004
27. R.J.A.Little,, & D.B.Rubin, (2002). Statistical analysis with missing data. Hoboken, NJ: Wiley.
28. R.C.MacCallum,, K.F.Widaman,, S.Zhang,, & S.Hong, (1999). Sample size in factor analysis. Psychological Methods, 4, 84–99. doi:10.1037/1082-989X.4.1.84
29. J.M.Marszalek,, C.Barber,, J.Kohlhart,, & C.B.Holmes, (2011). Sample size in psychological research over the past 30 years. Perceptual and Motor Skills, 112, 331–348. doi:10.2466/03.11.PMS.112.2.331-348
30. X.-L.Meng, (1994). Multiple-imputation inferences with uncongenial sources of input. Statistical Science, 9, 538–558.
31. D.F.Morrison, (1973). A test for equality of means of correlated variates with missing data on one response. Biometrika, 60, 101–105. doi:10.1093/biomet/60.1.101
32. L.K.Muthén,, & B.O.Muthén, (2002). How to use a Monte Carlo study to decide on sample size and determine power. Structural Equation Modeling, 9, 599–620. doi:10.1207/S15328007SEM0904_8
33. T.E.Raghunathan,, J.M.Lepkowski,, J.Van Hoewyk,, & P.W.Solenberger, (2001). A multivariate technique for multiply imputing missing values using a sequence of regression models. Survey Methodology, 27(1), 85–95.
34. J.P.Reiter, (2007). Small-sample degrees of freedom for multi-component significance tests with multiple imputation for missing data. Biometrika, 94, 502–508. doi:10.1093/biomet/asm028
35. C.Rose,, & M.D.Smith, (2002). Mathematical statistics with Mathematica. New York, NY: Springer.
36. D.B.Rubin, (1987). Multiple imputation for nonresponse in surveys. New York, NY: Wiley.
37. V.Savalei,, & P.M.Bentler, (2009). A two-stage approach to missing data: Theory and application to auxiliary variables. Structural Equation Modeling, 16, 477–497. doi:10.1080/10705510903008238
38. V.Savalei,, & M.Rhemtulla, (2012). On obtaining estimates of the fraction of missing information from full information maximum likelihood. Structural Equation Modeling, 19, 477–494. doi:10.1080/10705511.2012.687669
39. J.L.Schafer, (1997). Analysis of incomplete multivariate data. London, UK: Chapman & Hall.
40. G.A.F.Seber,, & A.J.Lee, (2003). Linear regression analysis. Hoboken, NJ: Wiley-Interscience.
41. R.Tibshirani, (1996). Regression shrinkage and selection via the Lasso. Journal of the Royal Statistical Society, Series B (Methodological), 58, 267–288.
42. S.Van Buuren,, & K.Oudshoorn, (1999). Flexible multivariate imputation by MICE (TNO Report No. PG/VGZ/99.054). Leiden, The Netherlands: TNO Institute of Prevention and Health.
43. P.T.von Hippel, (2005). Teacher’s corner: How many imputations are needed? A comment on Hershberger and Fisher (2003). Structural Equation Modeling, 12, 334–335. doi:10.1207/s15328007sem1202_8
44. P.T.von Hippel, (2007). Regression with missing Ys: An improved strategy for analyzing multiply imputed data. Sociological Methodology, 37, 83–117. doi:10.1111/j.1467-9531.2007.00180.x
45. P.T.von Hippel, (2009). How to impute interactions, squares, and other transformed variables. Sociological Methodology, 39, 265–291. doi:10.1111/j.1467-9531.2009.01215.x
46. P.T.von Hippel, (2013a). The bias and efficiency of incomplete-data estimators in small univariate normal samples. Sociological Methods & Research, 42, 531–558. doi:10.1177/0049124113494582
47. P.T.von Hippel, (2013b). Should a normal imputation model be modified to impute skewed variables? Sociological Methods & Research, 42(1), 105–138. doi:10.1177/0049124112464866
48. D.A.Wagstaff,, & O.Harel, (2011). A closer examination of three small-sample approximations to the multiple-imputation degrees of freedom. Stata Journal, 11, 403–419.
49. N.Wang,, & J.M.Robins, (1998). Large-sample theory for parametric multiple imputation procedures. Biometrika, 85, 935–948. doi:10.1093/biomet/85.4.935
50. E.J.Wolf,, K.M.Harrington,, S.L.Clark,, & M.W.Miller, (2013). Sample size requirements for structural equation models: An evaluation of power, bias, and solution propriety. Educational and Psychological Measurement, 73, 913–934. doi:10.1177/0013164413495237
51. K.-H.Yuan,, & P.M.Bentler, (2007). Structural equation modeling. In C.R.Rao & S.Sinharay (Eds.), Handbook of statistics: Vol. 26. Psychometrics (pp. 297–358). Boston, MA: North Holland.
52. K.-H.Yuan,, & L.Lu, (2008). SEM with missing data and unknown population distributions using two-stage ML: Theory and its application. Multivariate Behavioral Research, 43, 621–652. doi:10.1080/00273170802490699
53. K.-H.Yuan,, F.Yang-Wallentin,, & P.M.Bentler, (2012). ML versus MI for missing data with violation of distribution conditions. Sociological Methods & Research, 41, 598–629. doi:10.1177/0049124112460373