The households were contacted using a random digit dialing
technique whereby telephone numbers in existing exchanges

in the district were systematically selected from the Vancouver
telephone directory and then altered by a digit to reach new
and/or unlisted numbers. Once the household was contacted,

the individual to be interviewed was also selected at random.
All eligible voters residing in the household were listed

and then the respondent was selected by a pre-set marking

on the questionnaire. Up to two attempts (one callback} were
made to each household to obtain an interview, in an effort

to reduce potential bias due to non-response.

The final sample was weighted by the person's chance of being
included from the household, to compensate for the under-~
representation of those from larger househclds that results

by taking only one telephone interview per household.

The questionnaire consisted of basically two questions after

‘a brief explanation of the referendum: the expected outcome
and how the person thinks they will vote. The questionnaire

used is appended.

Interviewing was conducted on September 18 and 19, 1981 in
Vancouver from a central location telephone facility where

interviewers are continuously supervised and monitored.

15