An analytically tractable latent Markov chain with correlated random effects from Gamma distributions is developed by combining techniques developed in latent class modeling and random effects modeling of survival and recurrent binary events data. Continuous unobserved heterogeneity in both classification error and latent class membership is allowed for. The model is used to compare a number of instruments that measure male labor market status in the British Household Panel Survey. Instrument specific, correlated random effects for classification errors are specified. Although the presented methodology is limited to binary events and specific distributional forms for the random effects, the analysis illustrates the importance of trying to understand and allow for potential complexities in classification error processes, if measurement error adjustments are to be relied on. The standard result of such analyses, that nontreatment of dependent classification errors leads to underestimation of the number of “changers,” is emphasized here.
Get full access to this article
View all access options for this article.
References
1.
Abowd, John M.
and Arnold Zellner. 1985. “Estimating Gross Labor Flows.”Journal of Business and Economic Statistics3:254-283.
2.
Clark, K. B.
and L. H. Summers. 1982. “The Dynamics of Youth Unemployment.” Pp. 199-234 in The Youth Labor Market Problem: Its Nature, Causes and Consequences, edited by R. Freeman and D. Wise. Chicago: University of Chicago Press for NBER.
3.
Clayton, David
and Jack Cuzick. 1985. “Multivariate Generalizations of the Proportional Hazards Model (With Discussion).”Journal of the Royal Statistical Society A148:82-117.
4.
Collins, Linda M.
and Stuart E. Wugalter. 1992. “Latent Class Models for Stage-Sequential Dynamic Latent Variables.”Multivariate Behavioural Research27(1):131-157.
5.
Conoway, Mark R.1990. “A Random Effects Model for Binary Data.”Biometrics46:317-328.
6.
Crouchley, Robert
and Andrew R. Pickles. 1995. “Multivariate Survival Models for Repeated and Correlated Events.”Journal of Statistical Planning and Inference47:95-110.
7.
Gönül, Füsun
. 1992. “New Evidence on Whether Unemployment and Out of the Labor Force Are Distinct States.”Journal of Human Resources27:329-361.
8.
Goodman, Leo A.
1976. “The Relationship Between Modified and Usual Multiple Regression Approaches to the Analysis of Dichotomous Variables.” Pp. 83-110 in Sociological Methodology, edited by D. R. Heise. San Francisco: Jossey-Bass.
9.
Graham, John W.
, Linda M. Collins, Stuart E. Wugalter, N. K. Chung, and William B. Hansen. 1991. “Modelling Transitions in Latent Stage-Sequential Processes: A Substance Use Prevention Example.”Journal of Consulting and Clinical Psychology59:48-57.
10.
Hagenaars, Jacques A.
1988. “LCAG-Loglinear Modeling With Latent Variables: A Modified LISREL Approach.” Pp. 111-130 in Sociometric Research Volume 2. Data Analysis, edited by W. E. Saris and I. N. Gallhofer. London: Macmillan.
11.
Heckman, James J.
and Bo E. Honoré. 1989. “The Identifiability of the Competing Risks Model.”Biometrika76(2):325-330.
12.
Heckman, James J.
and Burton Singer. 1985. Longitudinal Analysis of Labor Market Data. Cambridge: Cambridge University Press.
13.
Hougaard, Philip
. 1986. “A Class of Multivariate Failure Time Distributions.”Biometrika73(3):671-678.
14.
Humphreys, Keith. 1994. “Estimation of Longitudinal Flows in the Presence of Classification Error: The Use of Models With Latent Classes and Random Effects.” Technical Report No. 94-10, Centre for Survey Data Analysis, University of Southampton.
15.
Humphreys, Keith. 1996. “The Latent Markov Chain With Random Effects From Multivariate Distributions: Some Simulation Evidence of Robustness to Mis-Specification of Distributional Form.” Technical Report No. 96-01, University of Southampton.
16.
Johnson, Norman
and Samuel Kotz. 1972. Distributions in Statistics: Continuous Multivariate Distributions. New York: John Wiley.
17.
Jöreskog, Karl G.
and Dag Sörbom. 1989. LISREL 7: A Guide to the Program and Applications. Chicago. SPSS.
18.
Kristiansson, K. E.1984. “Gross-Flow Estimates in the Swedish Labor Force Surveys.”Statistical Journal of the U.N. Economic Commission for Europe2:357-367.
19.
Langeheine, Rolf
and Frank Van de Pol. 1990. “A Unifying Framework for Markov Modeling in Discrete Space and Discrete Time.”Sociological Methods and Research18:416-441.
20.
Lawless, Jerald F.
1982. Statistical Models for Lifetime Data. New York: John Wiley.
21.
Marsh, Herbert W.1989. “Confirmatory Factor Analysis of Multitrait-Multimethod Data: Many Problems and a Few Solutions.”Applied Psychological Measurement13:335-361.
22.
Mathai, A. M.
and P. G. Moschopoulos. 1991. “On a Multivariate Gamma.”Journal of Multivariate Analysis39:135-153.
23.
Mathiowetz, Nancy A.
and Greg J. Duncan. 1988. “Out of Work, Out of Mind: Response Errors in Retrospective Reports of Unemployment.”Journal of Business and Economic Statistics6:221-229.
24.
McCutcheon, Alan L.
1987. Latent Class Analysis. Sage University Paper Series on Quantitative Applications in the Social Sciences, 07-064. Beverly Hills, CA: Sage.
25.
Petersen, J. H.
, P. K. Andersen, and R. D. Gill. 1996. “Variance Components Models for Survival Data.”Statistical Neerlandica50(1):193-211.
26.
Pickles, Andrew
, Robert Crouchley, Emily Simonoff, Lindon Eaves, Joanne Meyer, Michael Rutter, John Hewitt, and Judy Silberg. 1994. “Survival Models for Developmental Genetic Data: Age of Onset of Puberty and Antisocial Behavior in Twins.”Genetic Epidemiology11:155-170.
27.
Qu, Yinsheng
, Ming Tan, and Michael H. Kutner. 1996. “Random Effects Models in Latent Class Analysis for Evaluating Accuracy of Diagnostic Tests.”Biometrics52:797-810.
28.
Saris, Willem E.
and A. Van Meurs, eds. 1990. Evaluation of Measurement by Meta-Analysis of Multitrait-Multimethod StudiesAmsterdam: North-Holland.
29.
Shockey, James W.1988. “Adjusting for Response Error in Panel Surveys: A Latent Class Approach.”Sociological Methods and Research17:65-92.
30.
Singh, Avinash C.
and Georges E. Lemaitre. 1989. “On an ε-Contamination Model for Multinomial Response Error With Application to Labor Flows.” In Proceedings of the Section on Survey Research Methods, August 6-10. Washington, DC: American Statistical Association.
31.
Skinner, Chris J.
, David Holt, and T.M.F. Smith, eds. 1989. Analysis of Complex Surveys. New York: John Wiley.
32.
Skinner, Chris J.
and Nicola Torelli. 1993. “Measurement Error and the Estimation of Gross Flows From Longitudinal Economic Data.”Statistica53(3):391-405.
33.
Taylor, Marcia F.
, ed. 1992. British Household Panel Survey User Manual Volume A: Introduction, Technical Report and Appendices. Colchester: University of Essex Press.
34.
Van de Pol, Frank
and Rolf Langeheine. 1990. “Mixed Markov Latent Class Models.” Pp. 213-247 in Sociological Methodology, edited by C. C. Clogg. Oxford: Blackwell.
35.
Van de Pol, Frank
and Rolf Langeheine. Forthcoming. “Separating Change and Measurement Error in Panel Surveys (With an Application to Labor Market Data).” in Survey Measurement and Process Quality, edited by P. P. Biemer.
36.
Van de Pol, Frank
, Rolf Langeheine, and W. de Jong. 1991. “PANMARK User Manual: Panel Analysis Using Markov Chains, Version 2.2.”Netherlands Central Bureau of Statistics, Department of Statistical Methods, Voorburg.
37.
Wiggins, Lee M. 1955. “Mathematical Models for the Analysis of Multi-Wave Panels.” Ph.D. dissertation, Columbia University, Ann Arbor, MI.
38.
Wiggins, Lee M.
1973. Panel Analysis: Latent Probability Models for Attitude and Behavioural Processes. San Francisco: Jossey-Bass/Elsevier.
39.
Yashin, Anatoli I.
and Ivan A. Iachine. 1995. “Genetic Analysis of Durations: Correlated Frailty Model Applied to Survival of Danish Twins.”Genetic Epidemiology12:529-538.
