Documentation for the Workshop / Country Papers : New Zealand

Survey and census sources of national disability data

Country: New Zealand

1. Overview of survey or census

Title of survey or census - New Zealand Disability Survey
Date of latest data collection - 2001
Periodicity of survey or census - ad hoc
Contact person(s) for the survey/census - Denise Brown, Chief Analyst Social Statistics, PO Box 2922, Wellington, New Zealand; denise.brown @ stats.govt.nz, phone 64 04 495 4600; www.stats.govt.nz

2. Brief summary of survey or census

Describe what information on disability is collected
- nature, duration, cause and severity of disability, underlying condition, main disability (for those with multiple disabilities)
- characteristics of people with disabilities e.g age, sex, ethnicity, marital status, labour force status, highest educational qualification etc
- access to support services and special equipment, including unmet needs.
What uses are made of the data and who are the users?
NGOs in the disability sector use the data for advocacy. Government departments use the data to inform policy development and service delivery, and to monitor the government's disability strategy.
Were users involved in deciding what information to collect?
A stakeholders committee comprising representatives from government departments (education, health, social development, labour, Maori development etc) and the disability sector assisted in defining the survey objectives and information needs.
What definition of disability was used?
A functional concept of disability was used.
"Any restriction or lack (resulting from impairment) of ability to perform an activity in the manner or within the range considered normal for a human being." (WHO)
Using this concept, a disability was defined as any limitation in activity resulting from a long-term health condition or problem. People were not considered as having a disability if an assistive device (eg a hearing aid) completely eliminated their limitation. A concept of time was an additional filter; the disability must have lasted or be expected to last for six months or longer.
Were data collected and compiled according to international standards or national classifications and definitions? (e.g.ICIDH or ICF)
Data were collected and compiled according to the ICIDH.

3. Scope of survey or census

Describe the population covered in the survey (age, private dwelling population, institutionalised population, etc.)
The survey covered children, adults and older people living in private dwellings and group homes and people aged 15 years and over living in residential facilities.

4. Coverage and sampling

Coverage of survey or census (how well was the scope achieved)
The survey undercovered the target population, as it excluded
1. people living in off-shore islands and remote locations inaccessible by private car because of the high cost of sending interviewers to these places; and
2. selected types of non-private dwellings, such as hospices and acute psychiatric wards, forensic psychiatric wards, prisons, boarding houses etc.
  The exclusions are not considered to have a significant impact on estimates of the total prevalence of disability.
Sample frame(s) (sample size? Sample frame based on lists, area, register, etc.)
There were separate frames for private households and residential facilities (institutions). The sample frame for private households was the Population Census. Two disability questions were included in the census to help identify people with disabilities to take part in the survey. As well as drawing a sample of census respondents who said they had a disability, a sample was drawn from those who said they did not. The reason for including the latter group is that testing and previous surveys have shown that some people who have a disability do not record it in the census.
The sample-frame for residential facilities was a two-stage frame. The frame for the first stage was a list of all long-stay units (with 10 or more people) for people with an intellectual, psychiatric or physical disability, homes for the elderly and public and private hospitals with long-stay beds. The frame for the second stage was a list of all eligible residents living in sampled facilities. The administrator of each facility provided the list.

5. Are there other data sources which you use to benchmark your survey results? (e.g. previous surveys to compare with? what evaluations did you make of the results?)

The 1996/97 Disability Survey was used to benchmark the results of the 2001 Disability Survey.

6. Response and non-response

What was the response rate?
The achieved response for the household component of the 2001 Disability Survey was 73 percent, and 90 percent for the facilities component.
Did you evaluate the non-response? (e.g. break it up into full and partial? impact of non-response? follow-up studies? imputations for missing items?)
Unit non-response was accounted for in the weighting methodology. Item non-response for some variables for respondents living in private dwellings was accounted for by imputing their census records.

7. Data collection (Attach or mail us a copy of your questionnaire)

Manner of collection (face-to-face interviews, self-administered questionnaire, telephone, other?)
About three-quarters of household interviews for the 2001 Disability Survey were conducted using Computer Assisted Telephone Interviewing (CATI). Where respondents could not be interviewed by phone, such as in the case of a profoundly deaf person or in the case of a person living in a household with no access to a phone, the interview was conducted in person (face-to-face). Interviews of respondents living in residential facilities (institutions) were conducted in person by a trained interviewer.
Type of respondent (who replied? The disabled person or proxy for the person?)
Self and proxy responses were accepted for respondents living in residential facilities. Proxy response (from a knowledgeable adult) was the main method of collection for children. For adults living in private households, proxy response was only accepted where the selected respondent was not capable of answering for him or herself.
Format of questions (how many questions? pre-coded standardised?)
For respondents living in private households, the survey questions were split into two questionnaires. The screening questionnaire established whether the respondent had any disabilities and the nature, duration, cause and underlying condition. The content questionnaire was only asked of respondents who were established as having a disability in the screening questionnaire. The content questionnaire covered topics such as, services and assistance, employment, education, transport accommodation, special equipment and economic characteristics. Separate screening and content questionnaires were developed for adults and children. Only one questionnaire was administered to respondents living in residential facilities. It incorporated the screening questions administered to adults living in households, and some questions on equipment and assistance used.
List the languages used in the survey or census
The questionnaires were available in two languages: English and Samoan. In addition, interviewers who could conduct interviews in Maori and NZ sign language interpreters were available if required. Most interviews were conducted in English.

8. Measurement error

Which special data collection procedures were undertaken to reduce measurement error (e.g. pilot testing of questions and questionnaire? training of interviewers? detailed coding instructions? follow-up of non response by specialist staff?)
Centralised Computer Assisted Interviewing (CATI) enabled CATI supervisors to monitor interviewers calls and to review the work of individual interviewers. The CATI interviewers made up to six calls to contact someone in the household or to contact the respondent. If someone at the household was contacted, the best time to call the respondent was established. Pen and paper interviewers (PAPI) made up to three attempts to talk with someone at the respondent's address and a further three attempts to complete an interview with the respondent. Both CATI and PAPI interviewers attempted to contact selected survey respondents who had moved from their usual residence on Census night. The use of an electronic questionnaire (EQ) for the majority of household interviews ensured that interviewers could not skip a question without putting an answer in, and ensured the routing was correct. The EQ also ensured that answers made sense, with built-in edits. The questionnaire used for the Residential Facilities was simpler than that used for households, and extensive use was made of family and staff proxies for respondents unable to answer for themselves due to memory loss or other conditions.

9. Highlights of the strengths and limitations of survey or census

Strengths: (e.g. a very detailed survey; sound sample design; use of international standard concept and classifications, etc.)
The major strengths of the survey are that it provides the only comprehensive and reliable statistical information on the nature and extent of disability in New Zealand and the barriers faced by people with disabilities in their day-to-day lives. The survey content was developed with major input from the disability sector, and is therefore relevant and appropriate to their needs. The questions were cognitively tested and the survey instrument and field procedures tested in a pilot survey. Data from the 2001 Survey is comparable with that from the 1996/97 Survey. Oversampling of the Maori and Pacific populations in the 2001 Disability Survey increased the reliability of estimates for these populations.

Limitations: (e.g. high relative standard error for small population groups/geographic areas; respondent overload due to length of questionnaire; lack of consistency or survey items across surveys, etc.)
A major limitation of the survey is that it underestimates intellectual and psychiatric disability, a consequence of the fact that the survey excludes people living in some types of residential facilities, such as prisons and acute psychiatric wards. A further limitation is inability to produce estimates for small areas and small population groups. Because the survey is ad hoc, there is no guarantee that it will be repeated in the future.

10. List published sources on disability statistics

Disability Counts, Statistics New Zealand, March 1998

Disability in New Zealand: Overview of the 1996/97 Surveys, Health Funding Authority and Ministry of Health.

Disability Counts 2001, Statistics New Zealand, May 2002

Disability Surveys Snapshots:

Key facts

Maori

Pacific Peoples

People in Residential Care

Children

Sensory Disabilities

Physical Disabilities

Intellectual Disabilities

Psychiatric and Psychological Disabilities

(The snapshots are published on the Statistics New Zealand's website: www.stats.govt.nz)

See also: New Zealand Disability Survey contributed paper

Return to top