modeling achievement trajectories when attrition is informative

25
Modeling Achievement Trajectories When Attrition is Informative Betsy J. Feldman & Sophia Rabe-Hesketh

Upload: aman

Post on 14-Jan-2016

36 views

Category:

Documents


0 download

DESCRIPTION

Modeling Achievement Trajectories When Attrition is Informative. Betsy J. Feldman & Sophia Rabe-Hesketh. Dropout and missing data in longitudinal educational data; Traditional approaches (listwise deletion, mean imputation) - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Modeling Achievement Trajectories When Attrition is Informative

Modeling Achievement Trajectories When Attrition is Informative

Betsy J. Feldman & Sophia Rabe-Hesketh

Page 2: Modeling Achievement Trajectories When Attrition is Informative

• Dropout and missing data in longitudinal educational data;

• Traditional approaches (listwise deletion, mean imputation)

• Mean imputation: variance, covariance and standard error will be underestimated

• (a) imputed data are constant at a given time, hence, the variance will be underestimated;

• (b) regression coefficients for covariates (other than time) will be biased toward zero

• (c) treating imputed data as real ignore the variability of response, SE is underestimated (multiple imputation)

Page 3: Modeling Achievement Trajectories When Attrition is Informative

Missingness Mechanisms & Modeling Techniques

• 1. Models for Growth

• Level 1: time specific response (within-person)• Level 2: between-person variability in growth

trajectories

Page 4: Modeling Achievement Trajectories When Attrition is Informative

• Missing completely at random (MCAR): missingness does not depend on any other observed or unobserved variables

• Xgi is a time-varying and time-invariate covariate matrix• ygi is observation for individual i in group g at time t• ugi is a vector of intercept and slope residuals for

individual i in group g

Page 5: Modeling Achievement Trajectories When Attrition is Informative

• Covariate dependent missingness

• Missing at random (MAR)

• MCAR or MAR are referred to as ignorable or noninformative.

Page 6: Modeling Achievement Trajectories When Attrition is Informative
Page 7: Modeling Achievement Trajectories When Attrition is Informative

• Not missing at random (NMAR)• (a) missingness depend on missing values

• (b) missingness depend on random coefficient

Page 8: Modeling Achievement Trajectories When Attrition is Informative
Page 9: Modeling Achievement Trajectories When Attrition is Informative

• Survival-Process NMAR model

• Single-indicator NMAR model

Page 10: Modeling Achievement Trajectories When Attrition is Informative

• Two basic approach of modeling nonignorable missingness:

• (a) selection model: assuming the missing either depend on outcome-dependent missingness or random-coefficient-dependent missing;

• (b) pattern mixture model

Page 11: Modeling Achievement Trajectories When Attrition is Informative

Muthén-Roy Pattern-Mixture Model

Page 12: Modeling Achievement Trajectories When Attrition is Informative

Simulation

• Five analysis• (a) a traditional growth model (HLM)• (b) a dual-process single-indicator model in

which a single-indicator variable for dropping out at any time after the first time point

• (c) a survival-process model (true model)• (d) listwise deletion• (e) mean imputation

Page 13: Modeling Achievement Trajectories When Attrition is Informative

• Three manipulate variable • (a) sample sizes (n=300, 1000)• (b) percentage of missing data (10%, 40%)• (c) levels of dependence (weak and strong) of

the drop-out process weak: -0.1, -0.2 strong: -0.5, -1.4

Page 14: Modeling Achievement Trajectories When Attrition is Informative

Results

If missingness was NMAR but treated as ignorable, the slope means, slope variance, and covariance were biased only when missing percentage and the dependence were both high. The fit statistics is not likely to indicate the incorrect treatment. Standard error were found not to be affected much but the coverage were poor.

Page 15: Modeling Achievement Trajectories When Attrition is Informative

The slop parameters were biased when the missing percentage and the dependence were high, but better than growth model (MAR). SE were not affected.

The parameters and the coverage were good for the survival-process model which was the true model. The results were not reported.

Page 16: Modeling Achievement Trajectories When Attrition is Informative

Listwise deletion resulted in upward bias for slop and intercept mean, but the variance were underestimated.

Page 17: Modeling Achievement Trajectories When Attrition is Informative

The slope means and variance were upward biased.

? The residual variance should be underestimated. Consequently, the standard error will be underestimated.

Page 18: Modeling Achievement Trajectories When Attrition is Informative

Empirical Data

• National Education Longitudinal Study of 1988 (NELS: 88)

• Analysis:• (a) linear growth model with and without the

covariates• (b) dual-process survival-process NMAR model• (c) Single-indicator NMAR model

Page 19: Modeling Achievement Trajectories When Attrition is Informative
Page 20: Modeling Achievement Trajectories When Attrition is Informative

Ethical & Grade as dummy variables

Page 21: Modeling Achievement Trajectories When Attrition is Informative
Page 22: Modeling Achievement Trajectories When Attrition is Informative
Page 23: Modeling Achievement Trajectories When Attrition is Informative
Page 24: Modeling Achievement Trajectories When Attrition is Informative

Discussion & Conclusions

• Treating NMAR as ignorable (depend on the random coefficients) can results in biased estimates, especially for the estimated variances and covariances.

• The simulations showed what proportion of missingness and how strong the dependence will begin to result in serious bias.

Page 25: Modeling Achievement Trajectories When Attrition is Informative

Comments & Questions

• regression coefficients are given in Figure 1.• Use another imputation other than mean

imputation method• The y was substituted by estimate of theta

score, it may be risky to ignore the standard error of theta.

• It will be better to use plausible values in the growth model or to use multilevel IRT approach to estimate growth.