Websites:
http://multilevel.ioe.ac.uk
Includes details of current developments and publications in multilevel modelling.
Other examples of work with multilevel models.
www.ioe.ac.uk/multilevel/publications
These notes are by Mark Tranmer, 2004.
50
Assignment – FOR 5 CREDITS
(nb: people doing Advanced Data Analysis by Short Course do not need to do this
assignment.)
1. Briefly describe a multilevel population indicating the units of interest at each level.
2. Write down a (two level) multilevel model for a response (i.e dependent)
variable of
interest that is:
(i)
a continuous variable – random intercepts model
(ii)
a continuous variable –
random slopes model
(iii)
a binary variable
3. The following results were obtained when a variance components
model was fitted in
MLwiN The population has two levels,
with individuals, indexed by i,
living in areas,
indexed by j, for a variable with a continuous response. Write down the variance
components estimates for the individual level and the area level.
Calculate and interpret
the intra-area correlation
.
[ ]
[
]
[ ]
[
]
)
018
.
0
(
705
.
0
:
)
,
0
(
~
)
029
.
0
(
133
.
0
:
)
,
0
(
~
)
069
.
0
(
653
.
0
~
0
0
0
0
0
=
Ω
Ω
=
Ω
Ω
+
+
=
=
Ω
e
e
oij
u
u
oj
ij
j
j
j
ij
ij
N
e
N
u
e
u
x
y
)
Ν(XΒ,
y
β
β
Hand
in all work to Margaret Martin, Room NG22 Dover Street Building ground floor.
Tel 0161 275 4589 email
margaret.martin@man.ac.uk
Do'stlaringiz bilan baham: