| Information about the New Zealand Gaming Survey |
Availability
| Valid From: .. | |
| To: .. | Ongoing |
| Frequency: .. | One off |
Design
Purpose: The purpose of the survey was to provide information that would assist the development of gaming and related mental health policy and contribute to the development of robust frameworks for future studies into the prevalence and wider economic and social impacts of problem gambling. The purpose was also to provide a sound baseline to enable assessments of future changes in the national prevalence of problem gambling and gambling participation generally within the adult population.
General Information ..Target Population
The target population was defined as the New Zealand non-institutionalised, usually resident population aged 18 years and over.
Exclusions from the target population were:
· Long term residents of retirement homes, hospitals and psychiatric institutions
· Inmates of penal institutions
· Overseas diplomats and non-New Zealand staff members of foreign diplomats
· Overseas visitors who expect to be usually resident in New Zealand for less than 12 months
Population covered
The survey population was defined as the New Zealand on-shore, non-institutionalised, usually resident population aged 18 years and over, living in households that had a non-mobile, listed phone for private use at the time of sample selection.
Statistical Unit
Individual
Selection Unit
Individual
Collection Unit
Individual
Under Coverage
The survey population under-covered the target population where:
· People did not live in households with a telephone for private use
· People had an unlisted telephone number
· People only had access to a mobile telephone
· People lived on offshore islands (excluding Waiheke)
· Members of armed forces lived in barracks
Of these, the largest sources of under-coverage were the first two.
The 1996 Census indicated that 4 percent of all adults aged 18 years and over lived in dwellings without access to a telephone. However, this percentage was higher for certain ethnic groups: 14 percent for Maori adults and 16 percent for Pacific Island adults. This level of under-coverage in-groups of particular interest should be considered when making inferences from the sample to the total population.
Telecom figures show that approximately 13% of all residential telephone numbers are unlisted (i.e. are not published). These telephone numbers were therefore unavailable for selection.
There are no figures available for the other three sources of under-coverage.
Population and Sample Size
A total of 9300 telephone numbers were selected for Phase 1. The achieved sample size was 6452 adults.
Output Variables
Output variables are pieces of individual information that can be extracted from the survey/output data. Often output variables can be cross tabulated with other output variables, for example sales by industry classification. The list below contains all current and past output variables which have been released for this survey/output.
Changes in Output Variables over time
Guide to Interpreting Data
Summary of Changes to Survey/Output ..The Gaming Survey was a one-off survey and not subject to change.
Usage and Limitations of the Data ..Limitations:
The main limitations are:
- the high sampling errors associated with small estimates - this makes many of the smaller estimates unreliable or unusable.
- the exclusion of residents of households without telephones and people with unlisted telephone numbers.
- lower response rates for some groups, including young adults and Pacific Islands people.
Related Data Sources ..Household Economic Survey
Annual Enterprise Survey
Sampling Errors ..The estimates derived from this survey were based on a sample of households and individuals within households. Somewhat different figures might have been obtained if a complete census had been taken using the same questionnaire, interviewers, supervisors, processing methods, etc. as those actually used. The variability of a survey estimate due to the random nature of the sample selection is measured by its sampling error.
Given a certain sample size, the level of sample error for any given variable depends on the number of sampled individuals in the category of interest and the design effect for that variable. The design effect measures the performance of a complex sample design. It is the ratio of the variance under the complex sample design to the variance under simple random sampling.
Sampling within the household added extra complexity to the design. Because of this, sample errors for key outputs were estimated by the jack-knife method, rather than using analytic formulae. This accounted for all aspects of the sample design and estimation process, including the stratification, two stages of selection, and weighting adjustments.
The jack-knife is a computer intensive method of producing sampling errors. It is based on a methodology of comparing repeated estimates. The advantage of this method is its flexibility: the same method can be used for a variety of simple estimates, including totals, means, proportions, rates, and even complex statistics such as chi-square statistics and regression coefficients with no modifications. The disadvantage is that it is quite demanding on computer processing power, time and disk space, and furthermore it cannot easily deal with order statistics, e.g. medians and quartiles.
The basic idea behind the jack-knife (for a stratified cluster sample) is as follows:
1. Form random groups of units (households) in each stratum
2. Drop a group of units from a stratum
3. Re-run estimation without the dropped group
4. Store the estimate
5. Replace the group in the stratum, and repeat for all the groups in that stratum
6. Repeat for all strata
7. Compare the estimates - take the average squared difference.
Sample Error Guidelines
Estimates with large sample errors could be unreliable. For this reason, it is important to take into account the sample error measure when assessing the reliability of an estimate.
Tables from the survey with a relative sample error of 50 percent or more should be treated with extreme caution. Cells with a relative sample error between 30 and 50 percent need to be treated with caution.
These standards should be applied to any further analysis that is undertaken.
Confidence Intervals for proportions
Much of the output of Phase 1 will involve the reporting of proportions, e.g. statements of the kind "the proportion of males who are probable pathological gamblers". Using jack-knife replication for variance estimation (refer section 10.2) means reliable estimators of the variance of any particular proportion. Statistical inference can be derived, however, it requires the construction of a confidence interval (or equivalently provide a significance test).
Current best practice is to construct an estimate of a confidence interval using the Normal approximation, i.e.,
.
, as:
,
.
, the logit of the proportion, 
, were calculated as
,
was a variance estimate for 
. The 95 percent confidence limits for 
were obtained by applying the inverse logit transformation to the limits given above.
Non-sampling errors Errors that are not related to sampling may occur at almost every phase of a survey operation. Interviewers may misunderstand instructions, respondents may make errors in answering questions, the answers may be incorrectly entered into Blaise and errors may be introduced in the processing and tabulation of the data. These are all examples of non-sampling errors.
Errors occurring systematically will contribute to biases in the survey estimates. Considerable time and effort was made to reduce non-sampling errors in the survey. Quality assurance measures were implemented at each step of the data collection and processing cycle to monitor the quality of the data. These measures included:
· the use of highly skilled interviewers
· extensive training of interviewers with respect to the survey procedures and questionnaire
· monitoring of returned questionnaires to detect problems of questionnaire design or misunderstanding of instructions
· procedures to ensure that data capture errors were minimised
· coding and edit quality checks to verify the processing logic.
A major source of non-sampling error in surveys is the effect of non-response on the survey results. The extent of non-response varies from partial non-response (failure to answer just one or some questions) to total non response. Total non-response may occur because the interviewer was either unable to contact the selected household, the household refused to participate in the survey, or the selected respondent refused to participate in the survey. Total non-response was handled by adjusting the weight of households who responded to the survey to compensate for those who did not respond.
Caveats on Release ..Data should be suppressed where the relative sample errors are greater than 50 percent, and flagged where the relative sample errors are between 30 and 50 percent.
Customised Output
Catalogue & Reference Numbers
Other Comments
Classification(s) used
Glossary of Terms
Continuous gaming activity
Continuous gambling is characterised by the opportunity for a continuous repeated cycle of placing a stake, playing, determination and ability to collect and reinvest winnings, such as occurs with gaming machines, casino betting and horse betting.
Current problem gambler
A current problem gambler is a person who has met the South Oaks Revised Gambling Screen criteria for problem gambling in the past six months.
Current probable pathological gambler
A current probable pathological gambler is a person who has met the South Oaks Revised Gambling Screen criteria for probable pathological gambling in the past six months.
Employed
The employed are all persons in the working-age population who during the last week:
a) worked for one hour or more for pay, profit, or payment in kind in a job, business, farm or professional practice or,
b) worked without pay for one hour or more in work which contributed directly to the operation of a farm, business or professional practice operated by a relative or,
c) had a job or business they were temporarily absent from.
Ethnicity
Ethnicity is the ethnic group or groups reported by respondents in the ethnic group question.
Ethnic group
A social group whose members have the following four characteristics:
a) share a sense of common origins
b) claim a common and distinctive history and destiny
c) possess one or more dimensions of collective cultural individuality
d) feel a sense of unique collective solidarity
Infrequent gambler
A person who has gambled in their lifetime, but who gambles less than once a week.
Lifetime problem gambler
A lifetime problem gambler is a person who has at some time in their lives met the South Oaks Revised Gambling Screen criteria for problem gambling.
Lifetime probable pathological gambler
A lifetime probable pathological gambler is a person who has at some time met the South Oaks Revised Gambling Screen criteria for probable pathological gambling.
Never gambled
A person who has not gambled in their life.
Non-continuous gaming activity
A non-continuous gaming activity is an activity where there may be many hours or days between placing a stake or buying a ticket and the determination of the win or loss. Gambling on lotteries is an example of a non-continuous activity.
Non-gambler
A non-gambler is someone who has not gambled in the past six months.
Not in the labour force
Persons in the working-age population who are neither employed nor unemployed are “not in the labour force”.
Occupation
An occupation is a set of jobs, which involve the performance of a common set of tasks.
Past six month gambler
A past six-month gambler is someone who has gambled in the past six months.
Prioritised ethnic group
Respondents were asked to specify which ethnic group(s) they belong to, and were given the option of specifying as many as they wished. The priority system allocates people to one group only, on the following basis:
· “European” includes those who specified a European group as their sole ethnic group
· “NZ Maori” includes all those who specified NZ Maori either alone or in combination with other groups, “Pacific Island” includes all people who specified a Pacific Island ethnic group except those who also specified NZ Maori
· “Asian” includes all those who specified an Asian ethnic group except those who also specified NZ Maori or Pacific Island ethnic group
· “Other” includes all those who specified another ethnic group except those who also specified NZ Maori or a Pacific Islands or Asian ethnic group.
Probable pathological gambling
Probable pathological gambling is defined as those who have scored 5 or more on the South Oaks Revised Gambling Screen.
Problem gambling
Problem gambling is defined as those people who have scored 3 or 4 on the South Oaks Revised Gambling Screen.
Regular continuous gambler
A regular continuous gambler is a person who participates in a continuous gaming activity, once a week or more.
Regular non-continuous gambler
A regular non-continuous gambler is a person who participates in a non-continuous activity once a week or more and who is not a regular continuous gambler.
South Oaks Revised Gambling Screen
The South Oaks Revised Gambling screen (SOGS-R) was the instrument used to measure the extent of problem gambling in the survey population. SOGS-R is an international psychometric measure of whether someone is defined as a problem or probable pathological gambler.
Unemployed
The unemployed are all persons in the working-age population who are, or during the reference week were, without work, available for work and have actively sought work in the past four weeks.
Contact Details
Liability
Statistics New Zealand gives no warranty that the information or data supplied contains no errors. However, all care and diligence has been used in processing, analysing and extracting the information. Statistics New Zealand shall not be liable for any loss or damage suffered by the customer consequent upon the use directly, or indirectly, of the information supplied in this product.