PORTUGUESE LABOUR FORCE SURVEY 2013 and 2014
Transition of sampling frames
EXECUTIVE SUMMARY
During 2013, two sampling frames for Portuguese LFS coexisted on third and fourth quarters.
A new sampling frame, selected from the National Dwellings Register (NDR) is available, based on the Census 2011 data. The NDR is composed by all the addresses of private dwellings extracted from Census 2011.
The Portuguese LFS sample is a rotating panel in which 1/6 of the sample rotates each quarter and 5/6 are kept from previous waves.
The transition between the old sampling frame (the Master Sample 2001 - MS) and the new sampling frame was established to occur between the third quarter of 2013 and the fourth quarter of 2014: each quarter, anew rotation (1/6)is drawn from the new sampling frame.In the fourth quarter of 2014 all six rotations of the sample will belong to the new sampling frame.
The national sample size of LFS comprises 22554 dwellings until the second quarter of 2013 (based only on Master Sample-MS); after the transition period the national sample size will be comprised by 22572 dwellings.
SAMPLING FRAME
From the third quarter of 2013 to the third quarter of 2014 two sampling frames will coexist: the Master Sample 2001 (MS) and the new sampling frame drawn from the National Dwellings Register (NDR).
The MS was designed and selected using the information of the 2001 Census of Population and Housing (Census 2001). It is constituted by private dwellings and it excludes collective households and institutions since they represent 1% of the total population residing in Portugal.
The new sampling frame was selected from the NDR which in turn uses information collected in the 2011 Census. It is also constituted by private dwellings of usual residence and excludes collective households and institutions.
Both sampling frames are stratified one-stage cluster samples. In each stratum (NUTS 3) the clusters were selected systematically with probability proportional to size (number of private dwellings of usual residence). However the clusters were constructed with slightly differences:
• In MS theareas were geographical areas constituted by one or more contiguous statistical sections (census enumeration areas) with the purpose of having at least near 300 private dwellings of usual residence - 1408 PSUs were selected;
• In the new sampling frame the clusters were composed by one or more contiguous grid INSPIRE (**) cells with 1 Km2 of area, also aiming to have at least near 300 private dwellings of usual residence in each on of them - 1254 PSUs were selected.
The selection of PSU was made with probability proportional to size (in the number of private dwellings of usual residence).
SAMPLING DESIGN
For LFS sample for 2013, there are two different situations:
-First and second quarters– the rotations were selected from PSU of MS;
-Third and fourth quarters – transition period between sampling frames:
- New rotations were selected from PSU belonging to sampling frame of NDR;
- The remaining rotationsweredrawn from MS.
Quarter
1Q2013
2Q2013
3Q2013
4Q2013
1Q2014
2Q2014
3Q2014
4Q2014
Master Sample (MS)
Sampling frame from NDR
In the second stage, dwellings were selected systematically in each PSU. In the new sampling frame and for each rotation and NUTS2 region, 3 dwellings are selected by PSU.
(**)