MCMC Estimation for Random Effect Modelling

1
MCMC Estimation for Random Effect Modelling The
MLwiN experience

• Dr William J. Browne
• School of Math Sciences
• University of Nottingham

2
Contents

• Random effect modelling and MLwiN.
• Methods comparison Guatemalan child health
  example.
• Extendibility of MCMC algorithms
• Cross classified and multiple membership models.
• Artificial insemination and Danish chicken
  examples.
• Further Extensions.

3
Random effect models

• Models that account for the underlying structure
  in the dataset.
• Originally developed for nested structures
  (multilevel models), for example in education,
  pupils nested within schools.
• An extension of linear modelling with the
  inclusion of random effects.
• A typical 2-level model is
• Here i indexes pupils and j indexes schools.

4
MLwiN

• Software package designed specifically for
  fitting multilevel models.
• Developed by a team led by Harvey Goldstein and
  Jon Rasbash at the Institute of Education in
  London over past 15 years or so. Earlier
  incarnations ML2, ML3, MLN.
• Originally contained classical IGLS estimation
  methods for fitting models.
• MLwiN launched in 1998 also included MCMC
  estimation.
• My role in team was as developer of MCMC
  functionality in MLwiN during 4.5 years at the
  IOE.

5
Estimation Methods for Multilevel Models

• Due to additional random effects no simple matrix
  formulae exist for finding estimates in
  multilevel models.
• Two alternative approaches exist
• Iterative algorithms e.g. IGLS, RIGLS, EM in HLM
  that alternate between estimating fixed and
  random effects until convergence. Can produce ML
  and REML estimates.
• Simulation-based Bayesian methods e.g. MCMC that
  attempt to draw samples from the posterior
  distribution of the model.

6
Methods for non-normal responses

• When the response variable is Binomial or Poisson
  then different algorithms are required.
• IGLS/RIGLS methods give quasilikelihood estimates
  e.g. MQL, PQL.
• MCMC algorithms including Metropolis Hastings
  sampling and Adaptive Rejection sampling are
  possible.
• Numerical Quadrature can give ML estimates but is
  not without problems.

7
So why use MCMC?

• Often gives better estimates for non-normal
  responses.
• Gives full posterior distribution so interval
  estimates for derived quantities are easy to
  produce.
• Can easily be extended to more complex problems.
• Potential downside 1 Prior distributions
  required for all unknown parameters.
• Potential downside 2 MCMC estimation is much
  slower than the IGLS algorithm.

8
The Guatemalan Child Health dataset.

• This consists of a subsample of 2,449 respondents
  from the 1987 National Survey of Maternal and
  Child Helath, with a 3-level structure of births
  within mothers within communities.
• The subsample consists of all women from the
  chosen communities who had some form of prenatal
  care during pregnancy. The response variable is
  whether this prenatal care was modern (physician
  or trained nurse) or not.
• Rodriguez and Goldman (1995) use the structure of
  this dataset to consider how well
  quasi-likelihood methods compare with considering
  the dataset without the multilevel structure and
  fitting a standard logistic regression.
• They perform this by constructing simulated
  datasets based on the original structure but with
  known true values for the fixed effects and
  variance parameters.
• They consider the MQL method and show that the
  estimates of the fixed effects produced by MQL
  are worse than the estimates produced by standard
  logistic regression disregarding the multilevel
  structure!

9
The Guatemalan Child Health dataset.

• Goldstein and Rasbash (1996) consider the same
  problem but use the PQL method. They show that
  the results produced by PQL 2nd order estimation
  are far better than for MQL but still biased.
• The model in this situation is
• In this formulation i,j and k index the level 1,
  2 and 3 units respectively.
• The variables x1,x2 and x3 are composite scales
  at each level because the original model
  contained many covariates at each level.
• Browne and Draper (2004) considered the hybrid
  Metropolis-Gibbs method in MLwiN and two possible
  variance priors (Gamma-1(e,e) and Uniform.

10
Simulation Results

• The following gives point estimates (MCSE) for 4
  methods and 500 simulated datasets.

Parameter (True) MQL1 PQL2 Gamma Uniform
ß0 (0.65) 0.474 (0.01) 0.612 (0.01) 0.638 (0.01) 0.655 (0.01)
ß1 (1.00) 0.741 (0.01) 0.945 (0.01) 0.991 (0.01) 1.015 (0.01)
ß2 (1.00) 0.753 (0.01) 0.958 (0.01) 1.006 (0.01) 1.031 (0.01)
ß3 (1.00) 0.727 (0.01) 0.942 (0.01) 0.982 (0.01) 1.007 (0.01)
s2v (1.00) 0.550 (0.01) 0.888 (0.01) 1.023 (0.01) 1.108 (0.01)
s2u (1.00) 0.026 (0.01) 0.568 (0.01) 0.964 (0.02) 1.130 (0.02)
11
Simulation Results

• The following gives interval coverage
  probabilities (90/95) for 4 methods and 500
  simulated datasets.

Parameter (True) MQL1 PQL2 Gamma Uniform
ß0 (0.65) 67.6/76.8 86.2/92.0 86.8/93.2 88.6/93.6
ß1 (1.00) 56.2/68.6 90.4/96.2 92.8/96.4 92.2/96.4
ß2 (1.00) 13.2/17.6 84.6/90.8 88.4/92.6 88.6/92.8
ß3 (1.00) 59.0/69.6 85.2/89.8 86.2/92.2 88.6/93.6
s2v (1.00) 0.6/2.4 70.2/77.6 89.4/94.4 87.8/92.2
s2u (1.00) 0.0/0.0 21.2/26.8 84.2/88.6 88.0/93.0
12
Summary of simulations

• The Bayesian approach yields excellent bias and
  coverage results.
• For the fixed effects, MQL performs badly but the
  other 3 methods all do well.
• For the random effects, MQL and PQL both perform
  badly but MCMC with both priors is much better.
• Note that this is an extreme scenario with small
  levels 1 in level 2 yet high level 2 variance and
  in other examples MQL/PQL will not be so bad.

13
Extension 1 Cross-classified models
For example, schools by neighbourhoods. Schools
will draw pupils from many different
neighbourhoods and the pupils of a neighbourhood
will go to several schools. No pure hierarchy can
be found and pupils are said to be contained
within a cross-classification of schools by
neighbourhoods

nbhd 1 nbhd 2 Nbhd 3
School 1 xx x
School 2 x x
School 3 xx x
School 4 x xxx
 
14
Notation
With hierarchical models we use a subscript
notation that has one subscript per level and
nesting is implied reading from the left. For
example, subscript pattern ijk denotes the ith
level 1 unit within the jth level 2 unit within
the kth level 3 unit. If models become
cross-classified we use the term classification
instead of level. With notation that has one
subscript per classification, that captures the
relationship between classifications, notation
can become very cumbersome. We propose an
alternative notation introduced in Browne et al.
(2001) that only has a single subscript no matter
how many classifications are in the model.
15
Single subscript notation
We write the model as
Where classification 2 is neighbourhood and
classification 3 is school. Classification 1
always corresponds to the classification at which
the response measurements are made, in this case
patients. For pupils 1 and 11 equation (1)
becomes
16
Classification diagrams
In the single subscript notation we lose
information about the relationship(crossed or
nested) between classifications. A useful way of
conveying this information is with the
classification diagram. Which has one node per
classification and nodes linked by arrows have a
nested relationship and unlinked nodes have a
crossed relationship.
School
Neighbourhood
Pupil
Cross-classified structure where pupils from a
school come from many neighbourhoods and pupils
from a neighbourhood attend several schools.
Nested structure where schools are contained
within neighbourhoods
17
Example Artificial insemination by donor
1901 women 279 donors 1328 donations 12100
ovulatory cycles response is whether conception
occurs in a given cycle
In terms of a unit diagram
Or a classification diagram
18
Model for artificial insemination data
We can write the model as
Results
Note cross-classified models can be fitted in
IGLS but are far easier to fit using MCMC
estimation.
19
Extension 2 Multiple membership models

• When level 1 units are members of more than one
  higher level unit we describe a model for such
  data as a multiple membership model.
• For example,
•  Pupils change schools/classes and each
  school/class has an effect on pupil outcomes.
• Patients are seen by more than one nurse during
  the course of their treatment.

 
20
Notation
Note that nurse(i) now indexes the set of nurses
that treat patient i and w(2)i,j is a weighting
factor relating patient i to nurse j. For
example, with four patients and three nurses, we
may have the following weights
21
Classification diagrams for multiple membership
relationships
Double arrows indicate a multiple membership
relationship between classifications.
We can mix multiple membership, crossed and
hierarchical structures in a single model.
22
Example involving nesting, crossing and multiple
membership Danish chickens
Production hierarchy 10,127 child flocks
725
houses 304 farms
Breeding hierarchy 10,127 child flocks 200 parent
flocks
As a unit diagram
As a classification diagram
23
Model and results
Note multiple membership models can be fitted in
IGLS and this model/dataset represents roughly
the most complex model that the method can
handle. Such models are far easier to fit using
MCMC estimation.
24
Further Extensions / Work in progress

1. Multilevel factor models
2. Response variables at different levels
3. Missing data and multiple imputation
4. Sample size calculations

25
Multilevel factor analysis modelling

• In survey analysis often there are many responses
  for each individual.
• Techniques like factor analysis are often used to
  identify underlying latent traits amongst the
  responses.
• Multilevel factor analysis allows factor analysis
  modelling to identify factors at various
  classifications in the dataset.
• Due to the nature of MCMC algorithms by adding a
  step to allow for multilevel factor models in
  MLwiN, cross-classified models can also be fitted
  without any additional programming!
• See Goldstein and Browne (2002,2005) for more
  detail.

26
Responses at different levels

• In a household survey some responses refer to
  individuals in a household while others refer to
  the household itself.
• Models that combine these responses can be fitted
  using the IGLS algorithm in MLwiN and shouldnt
  pose any problems to MCMC.
• The Centre for Multilevel modelling in Bristol
  are investigating such models.
• See also Nicky Bests talk on combining datasets
  which is a similar issue.

27
Missing data and multiple imputation

• Missing data is proliferate in survey research.
• One popular approach to dealing with missing data
  is multiple imputation (Rubin 1987) where several
  imputed datasets are created and then the model
  of interest is fitted to each dataset and the
  estimates combined.
• Using a multivariate normal response multilevel
  model to generate the imputations using MCMC in
  MLwiN is described in chapter 17 of Browne (2003)
• James Carpenter (LSHTM) has begun work on macros
  in MLwiN that automate the multiple imputation
  procedure.

28
Sample size calculations

• Another issue in data collection is how big a
  sample do we need to collect?
• Such sample size calculations have simple
  formulae if we can assume that an independent
  sample can be generated.
• If however we wish to account for the data
  structure in the calculation then things are more
  complex.
• One possibility is a simulation-based approach
  similar to that used in the model comparisons
  described earlier where many datasets are
  simulated to look at the power for a fixed sample
  size.
• Bonne Zijlstra will be joining me in Nottingham
  in January on an ESRC grant looking at such an
  approach.

29
Conclusions

• In this talk we have attempted to show the
  flexibility of MCMC methods in attempting to
  match the complexity of data structures found in
  real problems.
• We have also contrasted the methods with the IGLS
  algorithm.
• Although we have used MLwiN, all the models
  considered here could also be fitted in WinBUGS.
• WinBUGS offers even more flexibility in model
  choice for complex data structures however it is
  typically slower in fitting models that can also
  be fitted in MLwiN.
• Genstat and the GLLAMM package in Stata also
  deserve mention as alternatives in the ML/REML
  world for the models considered.

30
References

• Browne, W.J. (2003). MCMC Estimation in MLwiN.
  London Institute of Education, University of
  London.
• Browne, W.J. and Draper, D. (2004). A Comparison
  of Bayesian and Likelihood-based methods for
  fitting multilevel models. University of
  Nottingham Research Report 04-01
• Browne, W.J., Goldstein, H. and Rasbash, J.
  (2001). Multiple membership multiple
  classification (MMMC) models. Statistical
  Modelling 1 103-124.
• Goldstein, H. and Browne, W. J. (2002).
  Multilevel factor analysis modelling using Markov
  Chain Monte Carlo (MCMC) estimation. In
  Marcoulides and Moustaki (Eds.), Latent Variable
  and Latent Structure Models. p 225-243. Lawrence
  Erlbaum, New Jersey.
• Goldstein, H. and Browne, W.J. (2005). Multilevel
  Factor Analysis Models for Continuous and
  Discrete Data. In Maydeu-Olivares, A and McArdle,
  J.J. (Eds.), Contemporary psychometrics a
  festschrift for Roderick P. McDonald, p 453-475.
  Lawrence Erlbaum, New Jersey.
• Goldstein, H. and Rasbash, J. (1996). Improved
  approximations for multilevel models with binary
  responses. Journal of the Royal Statistical
  Society, A. 159 505-13.
• Rodriguez, G. and Goldman, N. (1995). An
  assessment of estimation procedures for
  multilevel models with binary responses. Journal
  of the Royal Statistical Society, Series A, 158,
  73-89.
• Rubin, D.B. (1987). Multiple Imputation for
  Nonresponse in Surveys. New York J. Wiley and
  Sons.