1316.3 - Statistical Update Queensland (Newsletter), Mar 2004
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 08/03/2004
Page tools: Print Page Print All | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Editorial
2002 General Social Survey - A Snapshot of the Well-being of Australia In December 2003, the ABS released the first results from its inaugural General Social Survey, conducted in March-June 2002. Topics covered by the survey include: social attachment, crime and safety, personal stressors, financial stress and income, health and disability, use of information technology and transport. The survey enables analysis and comparison of the interrelationship of social circumstances and outcomes across population groups. The first results of the survey were released in General Social Survey, Summary Results, 2002 (cat. no. 4159.0) on 18 December 2003. General Social Survey state and territory tables were released as Excel datacubes on 21 January 2004 and General Social Survey: Users' Guide (cat. no. 4159.0.55.002) was released 19 February 2004. Some findings of interest from the Queensland datacube (cat. no. 4159.3.55.001) include: For further information contact Jenny Harber on 02 6252 5508 or jenny.harber@abs.gov.au.
Printed ABS Publications Can Now Be Purchased On-line The ABS ‘e-Commerce’ facilities have been upgraded to allow clients to order and pay for printed publications on-line. For selected publications, users will be given an additional option to order printed publications, in addition to the normal download (electronic) option. Users will then pay for their orders via the CommWeb banking web site, and the order will be fulfilled by the ABS print contractor in a similar manner to that currently in place for 2001 Census maps. Most current issue publications will be available for purchase on-line. Later on, printed copies of previous editions of some publications will become available on-line. The following types of publication, however, will not be made available for purchase in printed form via e-Commerce: For further information contact Keith Gilligan on 02 6252 6521 or keith.gilligan@abs.gov.au. Regional Small Business Statistics for 2000–01 The experimental small business statistics published in Experimental Estimates, Regional Small Business Statistics, Australia (cat. no. 5675.0) are part of an ABS strategy to respond to our clients’ need for regional data. Queensland Government clients, may find this data very useful to inform policy decision making which affects regional communities. Economic data may also add an important element to modelling economic behaviour in regional Queensland Small businesses are defined as those businesses whose total income or expenses were between $10,000 and $5m in the financial year. The data exclude tax-exempt businesses and some government activity. The statistics in this publication cover 76% of all businesses, but only 24% of business income. Although the data provide information on the economic activity of small business at a regional level, they cannot provide a complete measure of economic activity in a region. Despite these limitations the data are valuable for several reasons. They offer a viable regional time series and the health of large business is often reflected in that of small business. Further, the performance of small business is generally a good indicator of the prevailing economic conditions in a particular region. The data have been compiled from confidentialised files provided by the Australian Taxation Office of completed tax returns for companies, partnerships and trusts, and individuals declaring business income. Some Findings About Queensland In most statistical divisions, Queensland’s small businesses fared well from 1995–96 to 2000–01 with three statistical divisions out of 11 demonstrating average annual income growth of over 4% a year. As the map below shows, Far North, Northern, Mackay and Wide Bay–Burnett SDs had an average annual growth of less than 1.5%. Experimental Estimates: Average Annual Growth in Total Income of Small Businesses by Statistical Division, 1995-96 to 2000-01 Source: ATO files for companies, partnerships and trusts and individuals for 1995–96 to 2000–01 The agricultural industry contributed strongly to income growth in the south-west corner of Queensland. In the South West and Central West SDs, over 55% of small business income was sourced from agriculture, forestry and fishing in 2000–01 and the increase in cattle prices had a substantial effect on the business income in the area in recent years. The average profit of small businesses in the agricultural industry in Central West SD was $72,082 in 2000–01 and $25,387 in the South West SD, both increasing substantially from 1995–96 to 2000–01. The agricultural industry contributed strongly to income growth in the south-west corner of Queensland. In the South West and Central West SDs, over 55% of small business income was sourced from agriculture, forestry and fishing in 2000–01 and the increase in cattle prices had a substantial effect on the business income in the area in recent years. The average profit of small businesses in the agricultural industry in Central West SD was $72,082 in 2000–01 and $25,387 in the South West SD, both increasing substantially from 1995–96 to 2000–01. Growth, Income and Expenses, Agriculture, Queensland, 1995–96 to 2000–01 Income Growth by Industry, Moreton Statistical Division, 1995–96 to 2000–01 The other area to demonstrate strong growth in small business was Moreton SD. All industry divisions grew by at least 8% over the period 1995–96 to 2000–01. Small business in the service industries were generally the fastest growing, with Property and business services and Finance and insurance having the largest absolute increases in income. In the slower growing seaboard SDs, where the Agriculture, forestry and fishing industry represented between 17% and 25% of small business income, depressed sugar prices and poor recent seasons contributed to relatively slow growth. In these SDs the increase in income was generally exceeded by the increase in expenses, contributing to slower growth and reduced profitability from 1995-96 to 2000-01. Average profit to small business in this industry fell by over 70% in Mackay and Northern SDs to $7,626 and $10,285, respectively, in 2000-01. For further information about these statistics please contact Mark Chalmers on 07 3222 6307 or mark.chalmers@abs.gov.au. The Queensland Coast Attracts Retirees The ABS released Census of Population and Housing: Australia in Profile - A Regional Analysis (cat. no. 2032.0) on 16 January 2004. The analysis used results from the 2001 Census of Population and Housing to describe the major differences in the social and economic characteristics of people living in different parts of Australia. It covers population growth and distribution, cultural diversity, living arrangements, education, employment and unemployment, income and living standards and housing. Each chapter includes a comprehensive indicator table with data for each region covered in the publication. A case study on North West Queensland Statistical Division is also featured which draws together a wide range of census data relating to the population living and working in the area. Key findings of the report relating to Queensland include: Some items of interest from the case study on North West Queensland (Chapter 8) include: For further information contact Chris Mason on 02 6252 6214 or chris.mason@abs.gov.au. What’s Happening in Local Government Finance? The ABS Local Government Statistics Unit (LGSU) has been developing new forms in conjunction with most state/territory agencies to lessen the burden on councils reporting annual financial data. A full suite of these forms has been developed for use by Grants Commissions or Departments of Local Government in all jurisdictions in gathering financial information for the 2002–03 year. The LGSU also completed a post enumeration survey of the Local Government Finance Statistics quarterly estimates survey. This involved a series of interviews with councils currently in the survey to look at issues such as the quality of data provided to the ABS and the Electronic Data Reporting process and instrument. A number of recommendations from councils have been implemented and have been incorporated into the form relating to the September quarter 2003 which is currently in the field. In 2003–04, the LGSU will be conducting a methodological review of this survey. In particular, the ABS will be looking at possible improvements in the way the sample of councils in the survey is selected. Any outcomes from that review will be implemented in the September 2004 quarter. The LGSU made some minor changes to the sample for the September 2003 collection by rotating some councils in/out of the survey. The LGSU is also looking at the quality of the local government frame and, in particular, at the quality of frames currently being used in the ABS to select councils and other local government units for inclusion in various surveys. This project will be completed by mid 2004. Previously there has been a number of sources within the ABS for making such selections, and the aim of this project is to amalgamate, rationalise and centralise them. This should result in an improvement in data quality and also make for more representative selections of local government units in ABS surveys. In particular, this project will be looking at the treatment of Indigenous councils and public non-financial corporations. The consultation phase of the development of the Local Government Purpose Classification has been completed. The response received from the consultation was excellent, with great interest shown from a range of users within the sector. The ABS has incorporated feedback received into the classification and is currently embarking on a formal internal ABS approval process. It is expected that the classification will be approved and available for use in April 2004. For further information contact Dean Bloom on 07 3222 6404 or dean.bloom@abs.gov.au. Queensland State Supplementary Survey, 2004 Each year, ABS Queensland calls for submissions from state government agencies on topics for the Queensland State Supplementary Survey. This survey is conducted as a supplement to the Australia-wide monthly population survey using a multistage area sample. Information is obtained from approximately 5,000 private dwellings in urban and rural areas of Queensland. The information collected can relate to individuals in the household, the household itself or dwelling characteristics. This year six submissions were received, an increase from recent years. These submissions were assessed by the ABS according to their suitability for the methodology of the Queensland State Supplementary Survey and recommendations made to the Queensland State Statistical Consultative Committee for their selection of the preferred topic. The topic selected for the 2004 Queensland State Supplementary Survey was submitted by the Department of Housing. Proposed data items include characteristics of a dwelling, reasons for choosing a dwelling, intention to change dwellings and reasons for changing. The topic will be developed further over the next few months and the survey will go into the field in October. Data are due to be published in April 2005. For further information contact Robert Boyle on 07 3222 6213 or robert.boyle@abs.gov.au. Children and Youth Statistics In 2003, the ABS established the National Children and Youth Statistics Unit (NCYSU) in response to the need for a statistical evidence base to support community and government policy relating to children and youth. The objectives of the NCYSU include: maintaining dialogue with key stakeholders about emerging issues, data gaps and needs; developing statistical products relating to children and youth and providing statistical leadership regarding children and youth statistics. The NCYSU is guided by an advisory group comprising ABS and non-ABS representatives. In November 2003, the NCYSU launched the Children and Youth Theme Page on the ABS web site. The theme page provides information on key ABS data sources, and relevant non-ABS sources. The Theme Page can be accessed through the ABS web site by selecting ‘Themes’ from the navigator bar, then clicking on ‘Children and Youth’ under the ‘People’ sub-heading. The NCYSU newsletter, which is available on the theme page, details recent developments in children and youth statistics. Some recent and forthcoming releases on children and youth: Other recent and forthcoming publications with data on children and youth: For further information contact Carrington Shepherd on 08 9360 5255 or carrington.shepherd@abs.gov.au. Census Papers - Helping You Understand the Data Nine new working papers in the 2001 Census Papers program have recently been released on the ABS web site. These papers provide background information on data collection and procedural issues, and discuss data quality issues that have potential to affect the interpretation of census results. Topics now available include: 2001 Census Papers, and working papers from previous censuses are available free of charge on the ABS web site. For further information contact Kirsty Laughlin on 3222 6111 or kirsty.laughlin@abs.gov.au. Analytic Program Examines Census Data The Australian Census Analytical Program is producing a series of releases each of which focuses on a different aspect of census data. The Australian Census Analytical Program is run in conjunction with senior researchers from universities in Australia including RMIT University, Swinburne University, Australian National University and University of Canberra. Topics covered by the program include Counting the Homeless, Caring Labour in Australia’s Community Services, Indigenous Australians in the Contemporary Labour Market, etc. The first of the series, Australian Census Analytic Program: Counting the Homeless (cat. no. 2050.0) was released in November 2003. On census night in 2001 the homeless population in Australia was 99,900, compared with 105,304 homeless people on census night in 1996. The report found that absolute homelessness, such as sleeping out and improvised shelter, accounted for only 14% of homelessness in Australia. Most homeless people were sheltered somewhere at night, about half staying temporarily with friends, acquaintances and relatives, but as a group homeless people were highly transient. Australian Census Analytic Program: Australia Online: How Australians are Using Computers and the Internet (cat. no. 2056.0) was released on 12 January 2004. The study found that small country town usage was well below the national average for home computers (32% compared with 42% nationally) and the Internet (25% compared with 37% nationally). In contrast, people living out of town in rural areas (the rural balance), enjoyed higher home computer usage (41%) only marginally less that the national average. Internet usage in these areas was 32%. The study highlighted the need for access to the Internet in places other than the home or at work. Australians not in the work force, Indigenous Australians, children and those in disadvantaged households often accessed the Internet in ‘other’ places like public libraries and schools. Australian Census Analytic Program: Indigenous Australians in the Contemporary Labour Market (cat. no. 2052.0) was released on 20 January 2004. The study found that work undertaken by Indigenous Australians was more likely to be concentrated in the public sector and low skilled occupations. Indigenous Australians remained three times more likely to be unemployed, and less likely to be either working or looking for work than other Australians. The study found that poor education levels were the major cause of the employment differences between Indigenous and other Australians. Australian Census Analytical Program: The Micro-Dynamics of Change in Australian Agriculture: 1976–2001 (cat. no. 2055.0) was released on 9 February 2004. The purpose of this study was to provide a better understanding of the changes within rural Australia, particularly in the farm sector and the implications for resource management. While rural adjustment trends are well documented in other countries such as Canada and the United Sates, this is not so in Australia. The project explores methods of combining socio-economic data of persons and households from the Population Census with agricultural activity data from farms as collected by the Agricultural Census to help provide some answers. Other publications yet to be released include: Australian Census Analytical Program: Caring Labour in Australia’s Community Services (cat. no. 2051.0) Australian Census Analytical Program: Australia’s Most Recent Immigrants (cat. no. 2053.0) Australian Census Analytical Program: Australians’ Ancestry: 2001 (cat. no. 2054.0) For further information contact Michael Beahan on 02 6252 7007 or michael.beahan@abs.gov.au. Outcomes from the Review of Statistical Geography of South East Queensland After considerable consultation, the ABS has determined a new statistical geography for Statistical Divisions (SDs) for South East Queensland, which is detailed in the maps below. Changes from this review will come into effect with the Australian Standard Geographic Classification (ASGC) Edition 2006, in time for use in the next census. Map 1 shows the existing geography of South East Queensland. MAP 1 QUEENSLAND, ASGC 2003 STATISTICAL DIVISIONS The changes to SDs are described below: Moreton SD As the population of South East Queensland has grown, particularly in the Gold Coast and Sunshine Coast regions, the areas in the Moreton SD have developed distinct regional identities. A better statistical measurement of these regions can be achieved through the creation of three SDs out of the current Moreton SD, as outlined below: Brisbane SD The Brisbane SD should contain the anticipated development of the city for the next 15–20 years. The Brisbane SD will be extended to include the whole of the Caboolture and Ipswich LGAs and the northern urban section of the Beaudesert LGA. There will be some further minor refinement of the SD boundaries which intersect Beaudesert LGA. The northern sections of the Gold Coast will be excluded from the Brisbane SD and the relevant definitions in the ASGC will be revised to accommodate this change. MAP 2 QUEENSLAND, REVIEW OF STATISTICAL GEOGRAPHY—ASGC 2006 No changes were made to Darling Downs SD, Wide Bay–Burnett SD or Fitzroy SD as they are considered to be representative of their regions. Should you wish to discuss the changes or for further information contact Maria Shpakoff on 07 3222 6321 or maria.shpakoff@abs.gov.au. Chi square test in bivariate analysis. What does it mean? Chi square (2) (pronounced kai square) is a term that pops up in statistics but just what does it mean and what is a 2 test - what does it test? Chi square tests can be used to determine whether a relationship between variables identified in a sample study (e.g. males and females have different beverage preferences) can be generalised to the population or was due to chance alone. The 2 test is typically based on the assumption that there is no relationship between the two variables in the total population. This is called the null hypothesis. The test determines whether there is a statistically significant relationship or not within the data by comparing the actual sample results with the set of values that would be expected if there were no relationship between the variables. If a relationship does exist the 2 test will enable us to reject the null hypothesis. Consider the following table of beverage preferences of a sample of 50 males and 50 females randomly chosen from the population. (These are the observed or O values.)
For statistical propriety, we first need to determine how certain of our results we need to be. For this analysis an arbitrary probability of error of 5% will do, as not much of consequence would happen if we were wrong in our conclusions in extrapolating the results of the survey to the whole population. To determine whether there is any relationship evident between beverage preference and sex in the data we need to compare these data with the figures that would have been obtained if there were no relationship in the data (the null hypothesis), calculated from the row and column totals. To do this we take the product of the row and column totals in the table above for each cell and divide by the total for all cells. So for tea drinking males we multiply the top row total (50) by the first column total (16) and divide by 100 to get an E value of 8. Similarly for all other cells. This results in a table like this with 50% of the column total for both males and females in each cell. These are our expected or E values for the null hypothesis.
In plain language, the formula for chi squared is : 2 = the sum of all the values of (O–E)2/E where O = observed number and E = expected number. The values of O–E are:
Our (O–E)2/E values are:
Adding all the (O–E)2/E values gives 2 = 9.622. Now the question arises how do we interpret this result? We need to refer to a table of values of the 2 distribution. But which row of the 2 table do we use to interpret this result? To do this we need to find the number of degrees of freedom in our original table. We might think that there are 10 as there are 10 cells in the table (not counting the row and column totals), but this is not the case. In general, the number of degrees of freedom of a table is the number of rows in the table of observed frequencies minus one, multiplied by the number of columns, minus one, (r–1)(c–1), where neither (r–1) or (c–1) = 0. (In that case it is the number of cells – 1.) Ours is a 5 x 2 table so the number of degrees of freedom it has is 4 x 1 = 4. This table shows the first five lines of the 2 table. p value
A check along the df = 4 line to the probability of error 0.05 column shows that the value of 2 we obtained (9.622) is larger than the value tabulated for p = 0.05 (9.49), so we can reject the null hypothesis (the idea that there is no evidence of beverage preference attributable to sex) and assume there is some relationship between beverage preference and sex with only a 5% chance of being wrong. This means that only 5% of similar samples of a population in which there was no relationship between sex and beverage preference can be expected to give a value of 2 as high as this by chance. In this case the hypothesis we are testing is that there is no relation between beverage preference and sex (the null hypothesis). We are able to reject that. Note, however that 2 cannot prove a hypothesis is correct. Looking at the original data we might conclude that males have a preference for beverages beginning with letters in the first half of the alphabet, but that would not prove any causal relationship! So what can 2 do for you? A 2 test can help you sort the chaff from the wheat — it can be used to reject wrong hypotheses but cannot, alas, be used to prove right ones. In fact, two different hypotheses applied to the same data can both give probabilities large enough to prevent their rejection. In that case the test is telling you to collect more data! For further information contact Caitlin James on 07 3222 6305 or caitlin.james@abs.gov.au.
Statistical Update will be electronic only in future. Please send your email address as soon as possible. If you wish to subscribe to Statistical Update and receive it free of charge to your computer or change your subscription in any way, please contact Arthur Poulter on 07 3222 6084 or arthur.poulter@abs.gov.au. Document Selection These documents will be presented in a new window.
|