• Summary
• Downloads
• Explanatory Notes
• Related Information
• Past Releases

SUMMARY COMMENTARY


1. INTRODUCTION

This paper is intended as an indicative guide for users undertaking comparisons of selected health risk factors across the National Health Surveys (NHS) of 1989-90, 1995 and 2001. To enable users to check against their own analysis, crude rates are presented from each survey for alcohol consumption, body mass index (BMI), exercise level, and smoking status. In some instances, there has been a change over time in the wording of questions or associated inclusions, exclusions and prompts. To assist in understanding whether these factors may influence the level of response, questions and other notes on the questionnaire are included. A provisional assessment of the reliability of the NHS time series is given for each selected risk factor from both the trend in risk factor prevalence and the questionnaire differences. This information may help inform decisions on the suitability of using data for particular risk factors for analysis over time.

Throughout this paper, comments are made about the changes between surveys and their expected impact on the comparability of data. These are general comments based on ABS experience in survey development, results of testing, and a preliminary examination of results.

2. BACKGROUND

The 2001 NHS collected information about the health status of Australians, their use of health services and facilities, and health-related aspects of their lifestyle. Previous surveys in the series were conducted in 1989-90 and 1995. Surveys conducted by the ABS in 1977-78 and 1983, while not part of the NHS series, also collected information similar to that obtained in the National Health Surveys. The NHS is now expected to be collected every three years.

Changes to the 2001 NHS design were made to better meet the needs of key users, while maximising the reliability and validity of the data within cost and operational constraints.

Health risk factors have been identified as a key area of interest among clients undertaking across time analysis. This publication will provide users with a resource to assist in making NHS inter-survey comparability for health risk factors across the 1989-90, 1995 and 2001 surveys.

3. COMPARISON OF SURVEY DESIGN AND OPERATION FOR THE 1989-90, 1995 AND 2001 NHS

Understanding the comparability of data from the 2001 NHS with data from the previous National Health Surveys is important for the use of those data when interpreting changes in health characteristics over time. While the 2001 NHS is deliberately the same or similar in many ways to the 1995 NHS (and in part to the 1989-90 NHS), there are important differences in survey methodology and content, definitions, classification, etc. These differences will affect the degree to which data are directly comparable between the surveys, and hence the interpretation of apparent changes over the period from 1989-90 to 2001. This paper presents crude rates for selected health risk factors, and describes the methodological approach used in each survey to 'capture' the data. Issues of coding and classification will also be addressed.

The following table shows general characteristics of the three surveys, and where differences have the potential to affect comparability between surveys. It also provides some general comments as to the likely effects of these differences.

1989/90 NHS	1995 NHS	2001 NHS(G)
Personal interview with adult respondents; Personal interview with parents consent for children aged 15-17 years (otherwise a proxy interview); Proxy interview for children less than 15 years	Personal interview with adult respondents; Personal interview with parents consent for children aged 15-17 years (otherwise a proxy interview); Proxy interview for children less than 15 years	Personal interview with adult respondents; Personal interview with parents consent for children aged 15-17 years (otherwise a proxy interview); Proxy interview for children less than 15 years.
Self-completion women's health questionnaire for adult female respondents	Self-completion women's health questionnaire for adult female respondents	Self-completion women's health questionnaire for adult female respondents
	Self-completion general health & well-being questionnaire for adult respondents (SF 36)

TABLE 2: Relative standard errors (%)
Size of estimate: Australia	1995 NHS	2001 NHS(G)
10,000	13.2	30.6
100,000	4.4	9.7
1,000,000	1.3	2.4

As a result of the sample design, and through the weighting process, survey estimates at the State and part of State level by sex and by broad age group will be the same or very similar to the benchmark populations. Because the characteristics of the sample are not identical to those of the benchmark population (see table 3), some records will receive higher or lower weights than others. As a result, the Relative Standard Error (RSE) on estimates for groups which are affected by this may also be slightly higher or lower than the average RSE shown in the table above. As this will vary between surveys, it is a factor to consider in comparing the 1989/90, 1995 and 2001 data, but its impact on comparability is expected to be small.

TABLE 3: Comparison of weighted estimates to benchmark populations
	1989/90 NHS		1995 NHS		2001 NHS
Sex / age (yrs)	% of adults in sample	% of adults in pop'n	% of adults in sample	% of adults in pop'n	% of adults in sample	% of adults in pop'n(a)
Males	48.8	49.4	48.4	49.2	45.6	48.9
Females	51.2	50.6	51.6	50.8	54.4	51.1
18-24	14.3	15.6	13.9	14.2	9.1	12.5
25-34	23.4	22.5	21.9	21.2	19.4	19.8
35-44	21.1	20.6	21.8	20.5	22.7	20.6
45-54	14.3	14.4	17.0	16.7	17.9	18.4
55-64	11.9	11.7	11.0	11.3	12.6	12.7
65-74	9.7	9.7	9.3	10.1	10.1	9.1
75 and over	5.2	5.5	5.1	6.0	8.2	6.9
Persons	100.0	100.0	100.0	100.0	100.0	100.0
(a) Benchmark population as at 30 June 2001.

Household type	All household members enumerated	All children enumerated (a) but one or more adults not enumerated	All adults enumerated but one or more children not enumerated	One or more adults and one or more children not enumerated	Total
Single person	4 326	na	na	na	4 326
Couple only	6 (b)	4 561	na	na	4 567
Single adult with child(ren)	651	592	378	26	1 647
Couple with child(ren)	683 (b)	3 198	421	1 421	5 723
Other	65	1 455	19	116	1 655
Total - number	5 731	9 806	818	1 563	17 918
Total - per cent	32.0	54.7	4.6	8.7	100.0

Enumeration period

The 2001 NHS was enumerated over about a ten month period, compared with a 12 month period for the 1995 survey; the 2001 survey was not enumerated in December or January, nor during a 6 week period mid-winter (coinciding with the conduct of the 2001 Census of Population and Housing and the Post Enumeration Survey) (ABS 2003a).

4. OTHER ISSUES AFFECTING COMPARABILITY

In addition to survey design and operation, other factors may influence the way respondents answer questions and therefore, the comparability of surveys over time. Two such factors include context effects and social desirability.

Context effects are the influences on response by accompanying or associated questions. For example, in the 2001 NHS, before the questions on alcohol, respondents were asked certain questions about their nutrition, whereas for 1995 this was not the case. While the affect of this difference is expected to be limited, this may have provided an opportunity for respondents to think about consumption of food and drink, thus resulting in a higher alcohol consumption rate in 2001 when compared to 1995.

A further factor which may affect comparability is social desirability. The reported prevalence of risk factors is complex and dynamic, and directly a function of respondent knowledge and attitudes. This in turn may be influenced by the availability of health services and health information such as public education and awareness. Social desirability has been found to affect how respondents answer questions on some health risk factors. For example, for self-reported data, estimates of height are generally found to be overestimates and for weight, underestimates (ABS 1998). In addition, due to the implementation of promotional campaigns, drinking heavily and smoking may be increasingly less socially desirable and respondents may be under-reporting these activities. Similarly, campaigns to encourage people to exercise more have been introduced in recent years and this may influence respondents to over-report their exercise level. As there is currently no objective way to measure these differences, it is not possible to determine the degree to which social desirability has affected responses to alcohol consumption, smoking status or exercise level.

5. SELECTED HEALTH RISK FACTORS

The criteria used to determine the usefulness of a particular comparison over time are based on both the conceptual consistency between surveys and the plausibility of the trend produced. These assessments should be considered provisional as it is not possible to quantify the effect of changing the questionnaire, independently from the actual change in prevalence of risk factors among respondents over time. Each user is encouraged to take into consideration their own data needs and the purpose for which comparisons are being made before assessing whether the comparisons over time would be legitimate. For each selected risk factor, assessments are made for three inter survey comparisons (1989-90 to 1995, 1995 to 2001 and 1989-90 to 2001) and categorised with the following notations:

• Acceptable: No significant changes between surveys;
• Acceptable with limitations: Difficult to quantify the impact of questionnaire changes. Use with caution;
• Not acceptable: Significant inter survey differences or unexplainable major change in prevalence preclude the reliable use of data for comparison across time.

Alcohol

Comparability:
1989/90 and 1995: Acceptable with limitations
1995 and 2001: Acceptable with limitations
1989-90 to 2001: Acceptable with limitations

1989/90	1995	2001
Q 428 The next few questions are about alcoholic drinks. In the last seven days have you had any drinks at all that contain alcohol, including home made wine and beer?	Q 219 The next few questions are about alcoholic drinks. How long ago did you last have an alcoholic drink? 1 week or less More than 1 week to less that 2 weeks 2 weeks to less than a month 1 month to less than 3 months 3 months to less than 12 months 12 months or more Never Don't remember	Q 311 The next few questions are about alcoholic drinks. Some people may drink more or less than others, depending on their lifestyle and individual choices. How long ago did you last have an alcoholic drink? 1 week or less More than 1 week to less that 2 weeks 2 weeks to less than 1 month 1 month to less than 3 months 3 months to less than 12 months 12 months or more Never Don't remember
Q 429 How long ago did you last have an alcoholic drink? More than one week to less than 1 month ago 1 month to less than 3 months ago 3 months to less than 12 months ago 12 months or more ago Never	Q 222 On which days in the last seven did you have drinks that contained alcohol?	Q 312 On which days in the last 7, did you have drinks that contained alcohol?
Q 432 Did you have any drinks that contained alcohol on (specify each day of week starting from yesterday)?	Q 224 What did you have to drink on (specify day) (up to the last 3 days alcohol was consumed)? (Interviewer: prompt for quantity if not given)	Q 314B What did you have to drink on (specify day)? (Interviewer: prompt for quantity and brand type if not given)
Q 433 Interviewer: for each day on which respondent drank, ask both questions below. What kind of drinks did you have on (specify day)? How much (specify drink) did you have on (specify day)?	Q 225 Is the amount you drank last week more, less or about the same compared to most weeks? (more/less/about the same)	Q 318 Is the amount you drank last week more, about the same, or less, compared to most weeks? (more/about the same/less)
Q 459 Is the amount you drank last week more, less or about the same as you would drink most weeks? (more/less/about the same)

1989/90		1995		2001
Males Did not consume%	26.5	Males Did not consume %	34.2	Males Did not consume %	28.6
Low risk %	58.8	Low risk %	55.3	Low risk %	58.1
Risky %	7.7	Risky %	5.5	Risky %	6.9
High risk %	7.1	High risk %	5.0	High risk %	6.4
Total %	100.0	Total %	100.0	Total %	100.0
Females Did not consume%	48.2	Females Did not consume%	54.6	Females Did not consume%	47.7
Low risk %	44.4	Low risk %	39.3	Low risk %	43.9
Risky %	5.8	Risky %	4.8	Risky %	6.6
High risk %	1.6	High risk %	1.3	High risk %	1.9
Total %	100.0	Total %	100.0	Total %	100.0

1989/90		1995		2001
Underweight	<20	Underweight	<20	Underweight	< 18.5
Acceptable Weight	20-25	Acceptable Weight	20-25	Normal Range	18.5 - <20
Overweight	26-30	Overweight	26-30		20 - <25
Obese	>30	Obese	>30	Overweight	25 - <30
				Obese	>= 30

1989/90		1995		2001
Males Underweight %	2.6	Males Underweight %	2.0	Males Underweight %	1.8
Normal range (18.5-<20) %	5.4	Normal range (18.5-<20) %	3.5	Normal range (18.5-<20) %	2.9
Normal range (20-<25) %	47.6	Normal range (20-<25) %	41.0	Normal range (20-<25) %	36.8
Overweight %	33.7	Overweight %	35.7	Overweight %	38.1
Obese %	7.8	Obese %	10.5	Obese %	14.2
Not Stated %	2.9	Not Stated %	7.3	Not Stated %	6.1
Total %	100.0	Total %	100.0	Total %	100.0
Females Underweight %	7.2	Females Underweight %	5.2	Females Underweight %	4.8
Normal range (18.5-<20) %	11.5	Normal range (18.5-<20) %	9.1	Normal range (18.5-<20) %	8.6
Normal range (20-<25) %	48.3	Normal range (20-<25) %	43.0	Normal range (20-<25) %	39.7
Overweight %	19.7	Overweight %	20.4	Overweight %	22.1
Obese %	8.6	Obese %	10.3	Obese %	14.6
Not Stated %	4.7	Not Stated %	12.1	Not Stated %	10.2
Total %	100.0	Total %	100.0	Total %	100.0

1989-90		1995		2001
Low level	<1500	Sedentary	<100	Sedentary	<100 (including no exercise)
Moderate level	1500-<=3250	Low level	100 <1600	low level	100 <1600
High level	>3250	Moderate level	1600 - 3200 or >3200 and <2 hours vigorous exercise.	Moderate level	1600 - 3200 or >3200 and <2 hours vigorous exercise.
		High level	>3200 and 2 hours or more of vigorous exercise	High level	>3200 and 2 hours or more of vigorous exercise

1989-90		1995		2001
Males Sedentary %	34.6	Males Sedentary %	33.1	Males Sedentary %	30.0
Low %	29.2	Low %	31.1	Low %	33.6
Moderate %	25.5	Moderate %	26.5	Moderate %	26.4
High %	10.7	High %	9.3	High %	10.0
Total %	100.0	Total %	100.0	Total %	100.0
Females Sedentary %	36.0	Females Sedentary %	34.4	Females Sedentary %	32.0
Low %	37.3	Low %	39.0	Low %	41.3
Moderate %	22.4	Moderate %	22.2	Moderate %	22.5
High %	4.4	High %	4.4	High %	4.3
Total %	100.0	Total %	100.0	Total %	100.0

1989-90	1995	2001
	Due to change in scores for exercise intensity and scores for exercise level cut-offs, the categorised levels of exercise between 1989/90 and 1995 are not directly comparable.	Scores for exercise intensity changed between 1995 and 2001.

1989/90	1995	2001
Q 401. I would now like to ask you some questions about smoking. Do you currently smoke?	Q 215. I would now like to ask you some questions about smoking. Do you currently smoke?	Q 220. I would now like to ask you some questions about smoking. Do you currently smoke?
Q 402. Do you smoke - cigarettes? cigars? or a pipe?	Q 216. Do you smoke regularly, that is, at least once a day?	Q 221. Do you smoke regularly, that is, at least once a day?
Q 403. How old were you when you started smoking regularly?	Q 217. Have you ever smoked regularly (that is, at least once a day)?	Q 222. Have you ever smoked regularly (that is, at least once a day)?
Q 405. Do you usually smoke packet cigarettes or roll-your-own cigarettes?		Q 224. Does anyone else in this household smoke regularly, that is at least once a day?
Q 406. How many cigarettes do you usually smoke a day?		Q 225. How many other people in this household smoke regularly?
Q 407. Have you ever tried to give up smoking cigarettes?
Q 408. What is the tar content of your cigarettes?
Q 409. What is the nicotine content of your cigarettes?
Q 413. Have you ever smoked regularly?
Q 414. Did you smoke - cigarettes? cigars? a pipe?
Q 415. How old were you when you started smoking regularly?
Q 417. Have you ever smoked cigarettes regularly?
Q 418. Did you usually smoke packet cigarettes or roll your own cigarettes?
Q 419. How many cigarettes did you usually smoke a day?
Q 420. How old were you when you last gave up smoking cigarettes?
Q 421. Why did you give up smoking cigarettes? advice of family/friend/doctor (Why did ____ advise you to give up? Health (Interviewer ask: what particular aspect of health do you mean?) smoking harmful to health /cancer /lung disease/heart disease cough/sore throat reduces fitness/restricts activity other health reasons eg. operations, pregnancy Offensive to others Expense Lost interest/didn't feel like it anymore Other reasons
Q 423. Which was the main reason for giving up smoking cigarettes (Enter code from 421)

1989/90		1995		2001
Males Smoker %	32.1	Males Smoker %	29.2	Males Smoker %	27.6
Ex-smoker %	28.8	Ex-smoker %	31.1	Ex-smoker %	30.3
Never smoked %	39.1	Never smoked %	39.6	Never smoked %	42.0
Total %	100.0	Total %	100.0	Total %	100.0
Females Smoker %	24.7	Females Smoker %	22.0	Females Smoker %	21.1
Ex-smoker %	17.8	Ex-smoker %	21.5	Ex-smoker %	22.4
Never smoked %	57.4	Never smoked %	56.5	Never smoked %	56.4
Total %	100.0	Total %	100.0	Total %	100.0

1989/90	1995	2001
	Data collected in 1995 were not as detailed as data collected in 1989/90 and 2001.	Collection method different to 1995 NHS and may include persons aged less than 18 years old due to the question "How many other people in the household smoke regularly?"

ABS 2003b. Occasional paper: long-term health conditions: a guide to time series comparability from the National Health Survey, Australia, cat. no. 4816.0.55.001 last viewed 14 May 2004.

ABS 1998. How Australian's Measure Up, cat. no. 4359.0.

ABS 1996, National Health Survey, Australia, 1995 Users' Guide, cat. no. 4363.0

ABS 1992, National Health Survey, Australia, 1989-90 Users' Guide, cat. no. 4363.0

7. ATTACHMENT A

RELIABILITY OF ESTIMATES

Measuring sample variability

Since the estimates produced from data from the National Health Surveys are based on information obtained from a sub-sample of usual residents of a sample of dwellings, they are subject to sampling variability. That is, they may differ from those that would have been produced if all usual residents of all dwellings had been included in the survey. One measure of the likely difference is given by the standard error (SE), which indicates the extent to which an estimate might have varied by chance because only a sample of dwellings was included. There are about two chances in three that a sample estimate will differ by less than one SE from the number that would have been obtained if all dwellings had been included, and about 19 chances in 20 that the difference will be less than two SEs. Another measure of the likely difference is the relative standard error (RSE), which is obtained by expressing the SE as a percentage of the estimate.

CALCULATION OF STANDARD ERRORS

Tables of SEs and RSEs for estimates of numbers of persons appears at the end of this appendix. These figures will not give a precise measure of the SE for a particular estimate but will provide an indication of its magnitude.

An example of the calculation and the use of SEs in relation to estimates of persons is as follows. Consider the estimate for Australia in the 2001 National Health Survey of women aged 18 years and over who reported being current smokers (1,530,200). Since this estimate (1,530,200) is between 1,000,000 and 2,000,000 in the SE table for 2001, the SE will be between 24,000 and 30,000 and can be approximated by linear interpolation (see below) as 27,200 (rounded to the nearest 100). Therefore, there are about two chances in three that the value that would have been produced if all dwellings had been included in the survey will fall in the range 1,503,000 to 1,557,400 and about 19 chances in 20 that the value will fall within the range 1,475,800 to 1,584,600.

As can be seen from the SE tables at the end of this appendix, the smaller the estimate the higher the RSE. Very small estimates are thus subject to such high RSEs as to detract seriously from their value for most reasonable uses. Only estimates with RSEs of less than 25% are considered sufficiently reliable for most purposes. Estimates with larger RSEs should be used with caution and estimates with RSEs greater than 50% are considered too unreliable for general use.

How to interpolate between the estimates for SE tables

To interpolate between the estimates for SE tables the following formula is used:

The estimate 1,496,300 lies between:

x = estimate corresponding to lower RSE%

y = actual estimate

z = estimate corresponding to upper RSE%

= 0.9 + (1,496,300 - 2,000,000) / (1,000,000 - 2,000,000) x (1.3 - 0.9)

= 0.9 + (503,700 / 1,000,000) x 0.4

= 1.1

This formula can also be used to determine whether the confidence intervals of two proportions from two different National Health Surveys overlap. A confidence interval is a statement to the effect that the true value for a population lies within a given range of values at a specified level of confidence.

For example, in order to determine whether the crude smoking rate for women in 2001 was likely to be different from that in 1995, the following calculations can be made.

First, determine an approximation of RSEs for the two proportions. This will be done by neglecting the RSE of the denominator as suggested above.

Crude smoking rate in 2001 is:

1,530,300 / 7,238,300 x 100 = 21.1%.

RSE of 1,530,300 = 1.8%

1,496,300/6,799,200 x 100 = 22.0%

RSE of 1,496,300 = 1.1

(where p¹ = the proportion)

lower 95% confidence limit = 21.1 - (21.1 x 0.018 x 1.96)

= 21.1 - 0.7

= 20.4%

= 21.1 + 0.7

= 21.8%

For the crude smoking rate for women in 1995...

lower 95% confidence limit = 22.0 - (22.0 x 0.011 x 1.96)

= 22.0 - 0.5

= 21.5%

= 22.0 + 0.5

= 22.5

Now, by considering the two confidence intervals it can be seen that the intervals overlap. Therefore, we cannot be confident that there is a true difference between the two proportions and we cannot conclude that the true smoking rate in 2001 was higher than in 1995. Using this method, in order for there to be a difference between two estimates, the lower confidence limit of one proportion would need to be higher than the higher confidence limit of the other proportion. That is, the confidence intervals of the two proportions would not overlap.

Published estimates may also be used to calculate the difference between two survey estimates (numbers or percentages). Such an estimate is subject to sampling error. The sampling error of the difference between the two estimates depends on their SEs and the relationship (correlation) between them. An approximate SE of the difference between two estimates (p1-p2) may be calculated by the following formula:

The imprecision due to sampling variability, which is measured by the SE, should not be confused with inaccuracies that may occur because of imperfections in reporting by respondents, recording by interviewers, and errors made in coding and processing data. Inaccuracies of this kind are referred to as non-sampling error, and they may occur in any enumeration, whether it be a full count or a sample. Every effort is made to reduce non-sampling error to a minimum by careful design of questionnaires, intensive training and supervision of interviewers, and efficient operating procedures.

STANDARD ERRORS FOR PERSON ESTIMATES, NATIONAL HEALTH SURVEY 2001
	STANDARD ERROR							AUSTRALIA
	NSW	Vic.	Qld	SA	WA	Tas.	ACT	Standard error	Relative standard error
Size of estimate	no.	no.	no.	no.	no.	no.	no.	no.	%
500	520	488	499	404	438	342	268	468	93.7
1,000	848	782	777	647	686	526	397	750	75.0
1,500	1,113	1,019	997	839	880	666	492	978	65.2
2,000	1,342	1,222	1,184	1,002	1,046	780	570	1,174	58.7
2,500	1,548	1,403	1,350	1,145	1,190	880	635	1,350	54.0
3,000	1,734	1,566	1,500	1,272	1,320	969	693	1,512	50.4
3,500	1,904	1,718	1,638	1,390	1,439	1,047	742	1,659	47.4
4,000	2,064	1,860	1,764	1,496	1,548	1,120	788	1,800	45.0
4,500	2,219	1,989	1,881	1,598	1,652	1,184	832	1,930	42.9
5,000	2,360	2,115	1,995	1,690	1,745	1,245	870	2,055	41.1
6,000	2,622	2,346	2,202	1,866	1,920	1,362	942	2,286	38.1
8,000	3,088	2,752	2,568	2,160	2,232	1,552	1,056	2,696	33.7
10,000	3,500	3,100	2,880	2,420	2,490	1,710	1,160	3,060	30.6
20,000	5,040	4,440	2,060	3,340	3,460	2,260	1,480	4,440	22.2
30,000	6,180	5,400	4,920	3,960	4,140	2,610	1,680	5,490	18.3
40,000	7,080	6,160	5,600	4,440	4,680	2,880	1,840	6,320	15.8
50,000	7,850	6,800	6,200	4,850	5,100	3,100	1,950	7,050	14.1
100,000	10,600	9,100	8,300	6,200	6,600	3,800	2,300	9,700	9.7
200,000	13,800	12,000	10,800	7,600	8,400	4,400	2,600	13,200	6.6
300,000	16,200	13,800	12,600	8,400	9,600	4,800	3,000	15,600	5.2
400,000	17,600	15,200	14,000	8,800	10,400	5,200	2,800	17,600	4.4
500,000	19,000	16,500	15,000	9,500	11,000	. .	. .	19,000	3.8
1,000,000	23,000	20,000	19,000	11,000	13,000	. .	. .	24,000	2.4
2,000,000	28,000	24,000	22,000	. .	. .	. .	. .	30,000	1.5
5,000,000	35,000	. .	. .	. .	. .	. .	. .	40,000	0.8
10,000,000	. .	. .	. .	. .	. .	. .	. .	50,000	0.5
20,000,000	. .	. .	. .	. .	. .	. .	. .	60,000	0.3
. . not applicable

STANDARD ERRORS FOR PERSON ESTIMATES (USING CORE WEIGHTS), NATIONAL HEALTH SURVEY 1995(a)
	STANDARD ERROR								AUSTRALIA
	NSW	Vic.	Qld	SA	WA	Tas.	NT	ACT	Standard error	Relative standard error
Size of estimate	no.	no.	no.	no.	no.	no.	no.	no.	no.	%
150	. .	. .	. .	. .	. .	. .	. .	. .	. .	. .
200	. .	. .	. .	. .	. .	. .	100	. .	. .	. .
250	. .	. .	. .	. .	. .	. .	120	130	. .	. .
300	. .	. .	. .	. .	. .	. .	130	140	. .	. .
400	. .	. .	. .	. .	. .	. .	160	160	200	50.2
500	. .	. .	. .	. .	. .	. .	190	170	230	46.1
600	. .	. .	. .	. .	. .	. .	210	180	260	43.0
700	. .	. .	. .	. .	. .	380	230	190	280	40.5
800	. .	. .	. .	. .	. .	400	250	200	310	38.4
900	. .	. .	. .	460	. .	410	270	210	330	36.7
1 000	. .	. .	. .	470	. .	420	290	220	350	35.1
1 100	. .	. .	. .	490	. .	440	310	230	370	33.8
1 200	. .	. .	. .	500	. .	450	320	240	390	32.7
1 300	. .	. .	. .	520	. .	460	340	240	410	31.6
1 400	. .	. .	. .	530	730	470	350	250	430	30.7
1 500	. .	. .	. .	540	750	480	370	260	450	29.8
1 600	. .	. .	. .	550	760	490	380	260	460	29.1
1 700	. .	880	. .	560	780	500	400	270	480	28.3
1 800	. .	890	. .	570	800	510	410	270	500	27.7
1 900	. .	910	. .	580	810	510	420	280	510	27.1
2 000	. .	930	. .	590	830	520	440	280	530	26.5
2 100	. .	950	. .	600	840	530	450	290	550	26.0
2 200	. .	960	. .	610	850	540	460	290	560	25.5
2 300	. .	980	. .	620	870	550	470	300	570	25.0
2 400	. .	990	1,230	630	880	550	490	300	590	24.6
2 500	. .	1,010	1,250	640	890	560	500	310	600	24.1
3 000	. .	1,080	1,320	680	950	590	550	330	670	22.3
3 500	1,790	1,140	1,400	720	1,010	630	610	350	730	20.9
4 000	1,880	1,200	1,460	750	1,060	650	650	370	790	19.8
4 500	1,960	1,250	1,520	780	1,100	680	700	380	850	18.8
5 000	2,040	1,300	1,570	810	1,140	700	740	400	900	17.9
6 000	2,190	1,390	1,680	870	1,220	750	820	430	990	16.6
8 000	2,440	1,540	1,850	960	1,360	830	960	470	1,170	14.6
10 000	2,650	1,670	2,000	1,040	1,480	900	1,090	510	1,320	13.2
20 000	3,440	2,160	2,560	1,350	1,920	1,170	1,570	660	1,920	9.6
30 000	4,010	2,520	2,970	1,580	2,240	1,370	1,930	770	2,380	7.9
40 000	4,480	2,800	3,300	1,770	2,510	1,530	2,220	860	2,770	6.9
50 000	4,880	3,050	3,580	1,940	2,740	1,680	2,480	930	3,100	6.2
100 000	6,350	3,960	4,650	2,580	3,620	2,260	3,460	1,210	4,400	4.4
200 000	8,290	5,150	6,080	3,460	4,810	3,080	4,740	1,570	6,190	3.1
300 000	9,700	6,020	7,130	4,120	5,700	3,720	5,660	1,820	7,510	2.5
400 000	10,840	6,720	8,000	4,680	6,440	4,270	6,400	2,030	8,600	2.1
500 000	11,820	7,320	8,750	5,170	7,080	4,750	7,020	2,210	9,540	1.9
1 000 000	15,480	9,570	11,600	7,100	9,570	6,720	9,290	2,860	13,070	1.3
2 000 000	20,310	12,530	15,490	9,850	13,020	9,640	12,110	3,710	17,720	0.9
5 000 000	29,140	17,950	22,940	15,420	19,780	15,920	16,790	5,250	26,070	0.5
10 000 000	38,350	23,620	31,090	21,890	27,360	23,670	21,110	6,820	34,480	0.3
20 000 000	50,540	31,130	42,420	31,400	38,110	35,750	26,150	8,870	45,130	0.2
. . not applicable (a) The full sample of respondents were sequenced through the 1995 smoking, exercise, and body mass index topics and the core weight is used when calculating the estimates. These standard errors should be used for these topics.

	STANDARD ERROR								AUSTRALIA
	NSW	Vic.	Qld.	SA	WA	Tas.	NT	ACT	Standard error	Relative standard error
Size of Estimate	no.	no.	no.	no.	no.	no.	no.	no.	no.	%
100	1,970	550	1,310	400	560	390	60	160	80	80.0
200	2,040	660	1,350	460	640	420	100	200	140	70.0
300	2,100	740	1,400	500	690	450	130	220	190	63.3
500	2,210	860	1,480	560	780	490	200	260	280	56.0
700	2,310	950	1,550	610	860	520	250	290	350	50.0
1,000	2,430	1,060	1,640	670	940	570	320	320	450	45.0
1,500	2,600	1,210	1,760	750	1,060	630	410	370	590	39.3
2,000	2,740	1,330	1,860	820	1,160	680	500	420	710	35.5
2,500	2,850	1,450	1,950	900	1,250	750	550	450	800	32.0
3,000	3,000	1,500	2,050	950	1,350	800	650	500	900	30.0
3,500	3,100	1,600	2,100	1,000	1,400	800	700	500	1,000	28.6
4,000	3,200	1,700	2,150	1,050	1,450	850	750	550	1,100	27.5
5,000	3,350	1,800	2,300	1,100	1,600	900	850	600	1,250	25.0
7,000	3,650	2,050	2,500	1,250	1,800	1,050	1,050	700	1,550	22.1
10,000	4,000	2,350	2,800	1,450	2,050	1,200	1,250	800	1,900	19.0
15,000	4,550	2,700	3,150	1,700	2,400	1,450	1,600	950	2,400	16.0
20,000	4,950	3,050	3,450	1,900	2,750	1,650	1,850	1,100	2,850	14.3
30,000	5,650	3,550	4,000	2,250	3,250	2,000	2,250	1,350	3,550	11.8
40,000	6,250	3,950	4,450	2,600	3,700	2,300	2,550	1,550	4,150	10.4
50,000	6,800	4,350	4,850	2,850	4,100	2,550	2,850	1,700	4,650	9.3
100,000	8,900	5,800	6,450	4,000	5,750	3,750	3,800	2,450	6,600	6.6
150,000	10,550	6,900	7,750	4,950	7,050	4,750	4,500	3,050	8,000	5.3
200,000	12,000	7,800	8,850	5,750	8,200	5,700	5,000	3,600	9,200	4.6
300,000	14,450	9,300	10,800	7,150	10,250	7,350	5,750	4,550	11,050	3.7
500,000	18,500	11,750	14,000	9,600	13,650	10,400	6,850	6,100	13,850	2.8
1,000,000	26,500	16,250	20,450	14,500	20,550	17,050	8,400	9,350	18,450	1.8
2,000,000	38,950	22,650	30,750	22,500	31,700	29,000	10,050	14,600	24,100	1.2
5,000,000	67,500	35,850	54,950	41,650	58,100	61,750	12,250	27,000	33,400	0.7
10,000,000	105,600	51,400	88,100	68,250	94,300	113,750	13,750	44,050	41,750	0.4