SN:3884         SURVEY OF ENGLISH HOUSING 1996-97

UNIQUELY IDENTIFYING HOUSEHOLDS
-------------------------------
There is no single variable which uniquely identifies a household.
Instead the following variables are needed to match the data from one
file with that of another: ARNUM, ADNUM, HHNUM. However for the 'tenancy
group' files (prgrp96.por and prind96.por) the variable TENNO is also
required if the data are to be analysed by tenancy group.


WEIGHTING
---------
The results presented in the Housing in England 1996/97 report were
weighted to take account of non-response and grossed up to give
population estimates. The percentages were based on these grossed and
weighted figures. The grossed figures are in thousands so need to be
multiplied by 1000 to give the actual population figure.

If files are to be analysed then the data must be weighted.

How to use the weighting variable for each data file.

The household file (hhold96.por) should be weighted by the variable H4A
The variable H4A should be copied from the household file and used to
weight the individual (indiv96.por) and family unit (famun96.por) data.

The tenancy group file (prgrp96.por) should be weighted by the variable
H4AT. The variable H4AT should be copied from the tenancy group file and
used to weight the tenancy group individual file (prind96.por)

However using the weighting variable may cause SPSS to issue warnings
that "on at least one case, the value of the weight variable was zero,
negative or missing." This happens for cases where the depositor has
insufficient information to calculate the weights.


USEFUL INFORMATION
------------------
Please note that in files indiv96.por and prind96.por values for the
variable PNO (person number) greater than 9 relate to any children which
are in that household.