Page 337 - Contributed Paper Session (CPS) - Volume 4
P. 337
CPS2253 Gabrielle Palermo
The GWSM can be applied at Geres’s data to estimate weights for all waves.
The links between the base wave b and a given next wave t could be
(b)
represented by li,kj, where pupil i ∈ U and pupils j of school . The
GWSM steps are:
(t)
Stage 1: Compute the initial weight , given school k ∈ Ω .
(t)
(t)
And Ω = {j ∈ U | ∃i ∈ s and Li,k > 0}:
b
(0.1)
Where:
(b)
ti = 1 if i ∈ s , and 0 otherwise. li,kj = 1 if pupil i ∈ U (b)
corresponds to pupil , and 0 otherwise.
(t)
Stage 2: Sum of the links into each school k ∈ Ω :
(0.2)
Stage 3: The final weight is:
(0.3)
Where: for all pupil .
(t)
Therefore, the estimated total of pupils Y with a given characteristics is:
(0.4)
3. Result
The estimations weights for wave 4 were computed for schools, having the
population U1 as the base. The same calculations were done, but now having
as base the estimation weights of wave 3.
The figures 2 and 3 show the first results. The sampling weights are quite
different from the estimation weights.
326 | I S I W S C 2 0 1 9