Stratum
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
Unit ID
7
9
3
13
14
1
18
17
12
5
2
20
15
8
16
10
4
11
6
19
Random Sample
Stratum
0.70142 B
0.688363 B
0.122008 B
0.163691 B
0.142514 B
0.205718 B
0.742348 B
0.656411 B
0.982414 B
0.149085 B
0.767822 B
0.0600
ST SCI 3100 Final Exam
Fall 2011
1. A general theorem states that for sample designs with a xed sample size,
the sum of the rst-order inclusion probabilities over the population is equal
to the sample size. That is, N i = n, where as usual N and n are the
On Turings denominator for the
capture-recapture HTE
John Bunge
November 2, 2012
In our capture-recapture setup, we have the true population size = N , number
of trapping occasions k , total number of captures m, and number of individuals
captured n. In m
Computationally Intensive
Statistics
John Bunge
jab18@cornell.edu
Department of Statistical
Science
1
What is Computationally Intensive Statistics?
Journal Computational Statistics and Data Analysis:
Bayesian computing, bioinformatics, computational econo
Natasha Saidikowski
HW5
STSCI 3100
19 October 2011
(1) If the strata are perfectly homogeneous then the variance of the stratified-sample-based estimator of
the mean is zero; use equation (3) to show that this is true under post-stratification.
(2) Verify
ST 3100 Notes
John Bunge
Stratied random sampling: H strata, with sizes N1 , . . . , NH , H=1 Nh = N .
h
Draw SRS of size nh from stratum h, independently across strata. Estimand
1
is := N entire popn xi .
Stratied estimator or stratied mean:
xst :=
H
N
Adaptive cluster sampling point estimator
Dene
k
=1
N mk
n1
N
n1
,
where N is the total number of units in the population, n1 is the number of units in the
initial simple random sample, and mk is the number of units in the network containing the
k th samp