National Records of Scotland

Preserving the past, Recording the present, Informing the future

2001 Census Variables

2001 Census Variables

1. Introduction

Data collected in a Census will inevitably contain omissions and inconsistencies. Also these shortcomings will not be distributed evenly among the data. The aim of the Edit and Imputation process for the 2001 Census was to remove the omissions and inconsistencies in a way that removed any bias that would be left in the data without it. Without Edit and Imputation, individual users would be tempted to make adjustments to the results to compensate. No user would have the same access to the data that the Census Offices had, and each user may well make adjustments different from other users. So, the Edit and Imputation process made good the shortcomings to the data as collected in a way that provided consistent results to all users.

A similar aim underlay the One Number Census (ONC) process which added person and household records to the Census database to compensate for under-enumeration.

Each of the items collected on the Census form has been analysed to measure how often the form-filler’s response was used unchanged, how often it was changed or a missing value imputed, and how often the whole person or household record was made up by the ONC process. The general approach for each item has been to quantify for each category of the item (e.g. male and female for sex), the numbers and proportion of values unchanged from the entry on the form and changed or imputed in various ways.