Creating a 2000 IES-LFS Database in Stata

PROVIDE Project Technical Paper 2005:1

Table 1: Percentage differences (employment and wages)

Also reported in Figure 1 and Table 1 are employment figures reported in the IES and
LFS. The LFS reports higher employment figures for almost all occupation groups. The
patterns of employment in the LFS and IES are, however, similar. The unspecified category
shows the largest difference. This is probably due to the fact that the LFS uses two separate
questions to determine the occupation code, while the IES has only one question. Table 2 is a
cross-tabulation of the two occupation code variables. Only 45 persons report their
occupation code as ‘unspecified’ in the LFS, compared to the 1,798 unspecified workers in
the IES. Many of these workers (1,253) reported no income in the LFS, which is why they are
classified as ‘not applicable’ in the LFS.

As for the remainder of the occupation codes there is a fairly high correlation between the
LFS and IES. In Table 2 those observations on the diagonal of the cross-tabulation show
those persons who report the same occupation code in both the LFS and the IES. The
percentage ‘correctly categorised’ observations is fairly high. The ‘correctly categorised’ row
is defined as the number of observations on the diagonal divided by the column total (IES
number of observations), while the column is defined as the number of observations on the
diagonal divided by the row total (LFS number of observations). Section 4.2.4 continues the
discussion of the new employment figures that are used together with inclabp_new and
inclabp_old.

13

© PROVIDE Project

<<   <   15   16   17   18   19   20   21   >   >>