|Page tools: Print Page Print All|
TECHNICAL NOTE 1
8 The SE of an estimate may be obtained from the tables below.
RELATIVE STANDARD ERRORS
9 The SE can also be expressed as a percentage of the estimate and this is known as the relative standard error (RSE). In general, the size of the SE increases as the size of the estimate increases. Conversely, the RSE decreases as the size of the estimate increases. Very small estimates are thus subject to such high RSEs that their value for most practical purposes is unreliable.
10 The RSE is determined by dividing the SE of an estimate SE(x) by the estimate x and expressing it as a percentage. That is:
11 Proportions and percentages formed from the ratio of two estimates are also subject to sampling error. The size of the error depends on the accuracy of both the numerator and the denominator. A formula to approximate the RSE of a proportion is given below. This formula is only valid when x is a subset of y.
12 In the tables in this publication, only estimates with RSEs of 25% or less are considered reliable for most purposes. Estimates with RSEs greater than 25% but less than or equal to 50% are preceded by an asterisk (eg *3.4) to indicate they are subject to high SEs and should be used with caution. Estimates with RSEs of greater than 50% are preceded by a double asterisk (eg **0.3) and are considered too unreliable for general use. They should only be used to aggregate with other estimates to provide derived estimates with RSEs of 25% or less.
13 The following tables provide standard errors and relative standard errors for estimates of persons. Standard errors for estimates of hours are available on request.
TECHNICAL NOTE 2 REPROCESSING OF 1995 DATA
1 The 1995 Survey of Voluntary Work was conducted as a supplementary to the Monthly Population Survey (MPS). Trained interviewers visited each randomly selected household and interviewed a responsible adult member of that household. This person was asked to answer questions on their own behalf and for all other household members who satisfied the scope and coverage inclusion criteria. The lead-in to the supplementary topic on voluntary work required the respondent to identify any household member who had done any unpaid voluntary work in the previous 12 months. A personal interview with all persons so identified was then arranged.
2 The 2000 Survey of Voluntary Work was conducted as part of the Population Survey Monitor (PSM). Like the MPS, the PSM relies on trained interviewers visiting randomly selected households. However, in the PSM, one member of the household is randomly selected for personal interview.
3 As the first results of the 2000 Survey of Voluntary Work became available there appeared to be a very large and unexpected increase in the volunteer rate, from 19% of the population aged 15 years and over in 1995 to around 30% of the population aged 18 years and over in 2000. Excluding the 15-17 year olds from the 1995 data had little effect, increasing the volunteer rate to 19.4%. The discrepancy triggered an investigation of the effect of the changed methodology.
4 Despite the difference in the volunteer rate, results of both surveys in terms of the characteristics of volunteers and non-volunteers were similar. This suggested the possibility of some form of systematic undercount of volunteers in 1995, or an overcount in 2000. Based on results from the International Social Science Survey/Australia (IsssA), which recorded volunteer rates of 27% in 1995 and 33% in 1999, the former proposition was deemed the more likely.
5 The search for a possible source of undercount in the 1995 survey focussed on the methodology, which relied on one person in the household to identify all members of the household who undertook voluntary work. Where people had self-identified as volunteers, the volunteer rate was 27% while for those who been identified by another household member, the volunteer rate was 5.5%; a ratio of about 5:1. This large discrepancy suggests that in, general, people are unaware of the volunteer activities of other household members. The ratio varied between different age groups and relationships, being lowest for the 35-44 years age group and for parents with dependants, and highest for the 18-24 years age group and relatives other than partners or children. This variation accords with the premise that people are more likely to know about the activities of others if they have a close personal relationship.
6 A decision to reprocess the 1995 survey, using only those records where a full personal interview was conducted, was taken in the belief that this would make the results of the two surveys more closely comparable. However, the resulting sample is no longer strictly random. It is not possible to say with certainty that the sample is unbiased but the reprocessed data are likely to contain less bias than the original.
7 Using only records where a full personal interview was obtained reduced the sample size from 54,500 to 34,300, with a corresponding increase in standard errors. Standard errors for the reprocessed estimates presented in this publication are available on request.
8 The sample records were reweighted by adjusting the initial selection weights to take into account the fact that only responses from people who provided a full personal interview were being used. A calibration process was then used to weight the data up to known population totals at the following levels: state, part of state (metropolitan/ex-metropolitan), sex, ten-year age groups and labour force status, along with some cross-classifications of these categories. The calibration process ensured that the re-weighted data would be representative of the population for the variables that were used in the calibration. Any differences in volunteer characteristics between people who identified themselves as volunteers and those identified by another household member will thus be adjusted for if such differences were due to differences in the calibration variables. However, there may be differences which were not accounted for by the calibration. The reprocessed estimates should therefore be used with some caution.
COMPARISON OF 1995 ORIGINAL AND REPROCESSED ESTIMATES
9 Overall, reprocessing resulted in an increase in the number of volunteers aged 15 years and over, from 2.6 million to 3.3 million, in the volunteer rate from 19.0% to 23.9%, and in the number of hours from 434 million to 521 million.
These documents will be presented in a new window.