General Social Survey: User Guide, Australia

The 2014 GSS collected information about:

• demographic characteristics
• housing and mobility
• education (includes parental education)
• employment
• transport and mobility
• subjective wellbeing and general life satisfaction measures
• health and disability
• difficulty accessing service providers
• family and community involvement
• social networks and participation
• experiences of homelessness
• voluntary work
• crime and feelings of safety
• sports attendance and participation
• attendance at selected cultural and leisure venues
• information technology
• financial stress, resilience and exclusion
• income
• housing
• assets and liabilities
• discrimination
• visa status
• sexual orientation (collected for people aged 18 years and over).
   

All households were asked questions relating to the topics listed above. Most information was collected from the selected person aged 15 years or over. However, some information relating to the household e.g. financial and housing items may have been collected from a household spokesperson if the selected person nominated a more appropriate person to report on behalf of other members of the household.

The content of the 2014 GSS was finalised after extensive consultation with major users regarding data needs and priorities. The GSS Reference Group, comprising members from various commonwealth and state/territory government departments and agencies, universities and social research organisations, was established to advise on definitions, concepts, analysis needs and data uses.

While much of the GSS content remains unchanged between each survey cycle, a number of new topics and items were included for 2014. Cognitive interviews were conducted for these new topics to ensure concepts were understood by respondents. Cognitive interviews are semi-structured interviews in which the interviewer asks the respondent about their interpretation of questions and formulation of answers. Two rounds of cognitive interviews were conducted in March and April 2013. The next phase of survey development involved field testing the survey questionnaire and procedures. This was done with a 'dress rehearsal', conducted in South Australia from late September 2013 to early October 2013 and covering about 260 fully responding households.

The final enumeration of the survey was conducted from March to June 2014.

The GSS is designed to collect information for a core set of topics in each cycle, to allow analysis of changes over time, and a cyclical component to collect additional information on emerging topics of social concern. Approximately 80% of the content of the 2010 GSS was repeated in the 2014 GSS. A detailed voluntary work module similar to what formed part of the 2006 iteration, was included as part of the cyclical component for 2014. This allows for more direct comparison with the extended 2006 volunteering data.

The following sections detail some of the new or changed topics in the 2014 GSS. Refer also to the content comparison table between the 2014 and 2010 cycles in the Comparability with 2010 GSS section below. Details about topics not changed since the 2010 GSS can be found in the General Social Survey: User Guide, 2010 (cat. no. 4159.0.55.002), and a comprehensive list of survey definitions are available in the Glossary in the General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

For the first time, a new question was included in the 2014 GSS asking respondents about their sexual orientation. The inclusion of this question enables further investigation of the relative outcomes of people with different sexual orientations and allows a better understanding of whether they are more vulnerable to social disadvantage across a range of factors. This question was asked only of those people aged 18 years and over

A module on experiences of discrimination was included in the 2014 GSS to measure the relationship between discrimination and social exclusion. The types of discrimination prompted in the questions were: skin colour; nationality, race or ethnic group; language; dress or appearance; gender; age; disability or health issue; marital status; family status; sexual orientation; occupation; religious beliefs; political position; and other.

Data items available on discrimination include:

• whether experienced discrimination or been treated unfairly
• how often experienced discrimination or been treated unfairly
• places experienced discrimination or been treated unfairly
• reasons for most recent incident of discrimination.
   

In addition to level of education completed and barriers to education, questions were included to identify the highest level of parent/guardian educational attainment. The primary interest in these data is to measure the influence of inter-generational educational attainment on entrenched disadvantage. Two separate questions were asked about male and female parent/guardians of the respondent. The first relates to the highest year of school completed, and the second relates to the level of highest non-school qualification completed

Questions relating to time stress were included in the 2014 GSS as they provide an indication of how people feel about their work/life balance, and how it affects the amount of quality time they have to spend with family and friends.

Data items available on time stress include:

• frequency of feeling rushed or pressed for time
• main reason rushed or pressed for time
• how often has enough quality time with family and friends.
   

Respondents without a job and who were not permanently unable to work were asked two questions to determine if they had experienced barriers preventing them from having a job.

The disability module as well as formal and informal caring responsibilities were retained in 2014 with some changes to data items. The 2010 data items 'Delays to medical consultation' and 'Purchasing medication' were changed to 'Barriers to accessing health care' and 'Health care services for which people experienced barriers'

Long term conditions (LTC) aims to identify a broad range of conditions that have been diagnosed by a doctor or nurse. Mental health conditions such as depression, feeling depressed, behavioural and emotional disorders, and feeling anxious were included.

The LTC module allows comparison with results from the Australian Health Survey.

The table below summarises differences in content between the 2014 and 2010 GSS. The 2014 GSS was designed to provide, where possible, information comparable with the 2010 GSS. However, users should always consider the comparability notes in the table below, and any other issues noted within this User Guide (such as the possible impact of survey procedures, as discussed in Data collection) before making any comparisons between output from the 2010 and 2014 surveys.

A comparison of key estimates from the 2014 GSS with those from other ABS surveys is presented in Data comparability with other ABS sources (Appendix) on the Methodology page of the General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

In General Social Survey: Summary Results, 2014 (cat. no. 4159.0) most of the data tables provide information on both the estimated number of persons and the percentage of persons. Percentages in the tables are used to express proportions. For a brief summary on how to use and calculate percentages, see Statistical Language! 2008 (cat. no. 1332.0.55.002).

In General Social Survey: Summary Results, 2014 (cat. no. 4159.0) depending on the purpose of the table, the percentages in the tables will use different data as the denominator in the percentage calculation:

• in most tables, the 'all persons' total is used as the denominator - this percentage is used to describe what proportion of a population has a particular characteristic
• in other tables a sub population is used as the denominator - this percentage is used to describe what proportion of the sub population has a particular characteristic.
   

To determine which number has been used as the denominator in any percentage calculation, refer to the 'total' row/column containing 100% in the table. Examples of some of the different population/sub populations used as the denominator in General Social Survey: Summary Results, 2014 (cat. no. 4159.0) include (but are not limited to):

• Table 2 (row 17) - the total number of persons in each of the selected population groups is used as the denominator (100%) to calculate the proportion of people ranking their overall life satisfaction at different levels.
• Table 7 (row 16) - the total number of persons who experienced a barrier to accessing healthcare when they needed it in the last 12 months is used as the denominator (100%) to calculate the proportion of people who could not access different types of healthcare on the most recent occasion.
• Table 21 (column G) - the total number of volunteering involvements for different types of organisations is used as the denominator (100%) to calculate the proportion of involvements for different levels of annual hours volunteered.
   

When using the percentages shown in General Social Survey: Summary Results, 2014 (cat. no. 4159.0), it is important to pay attention to which population or sub population has been used as the denominator in any table. The population or sub population that is the denominator will impact how the data should be interpreted and described. For examples of how the data should be described, refer to the detailed summary of findings in General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

Due to the multi-dimensional nature of the GSS, data pertaining to different reference periods were collected. The main reference period was 'during the last 12 months', but other periods included 'during the last four weeks' and 'lifetime'. When interpreting the results presented in General Social Survey: Summary Results, 2014 (cat. no. 4159.0), reference should be made to the title of each table, data item labels and footnotes to determine the reference period

Users should also take into consideration the quality of the estimates when interpreting data from the GSS. Estimates with a relative standard error (RSE) of less than 25% are considered sufficiently reliable for most purposes, and only estimates of such precision are referred to in General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

To determine if there is a real difference between two estimates, statistical significance testing has been conducted for estimates and proportions presented in General Social Survey: Summary Results, 2014 (cat. no. 4159.0). Only differences where this test confirmed a statistically significant difference, with a 95% level of confidence, are described in the commentary (unless otherwise noted). Otherwise, despite what might appear to be differences between estimates, it cannot be stated with confidence that the differences are real and not the result of sampling error.

The ABS advises users to consider RSEs in any data analysis, and conduct significance testing on data to determine whether there is a statistically significant difference between estimates. For further information about how to calculate RSEs and conduct significance testing, refer to the Data quality (technical note) section in the Methodology of General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

The Methodology page of the General Social Survey: Summary Results, 2014 (cat. no. 4159.0) contain information about the survey's scope, coverage and sample design.

ABS Interviewers conducted personal interviews at selected dwellings during the period March to June 2014. Much of the detail obtained from the GSS was provided by one person aged 15 years or over, randomly selected from each participating household. The random selection of this person was made once basic information had been obtained about all household members. Some financial and housing items collected in the GSS required the selected person to answer on behalf of other members of the household. In some cases, particularly where household information was not known by the selected person, a spokesperson for the household was nominated to provide household information

Selected households were sent pre-approach material by mail to inform them of their selection in the survey. The pre-approach material consisted of a guide (sent 15 days before the survey), a registration letter and leaflet (sent 11 days before the survey) and a reminder letter (sent five days before the survey). The registration letter contained log on credentials for households to register their contact details online so an Interviewer could arrange a convenient time for an interview. Households unable to register online were asked to phone the ABS. An Interviewer then contacted the selected households to arrange an interview. Where respondents did not respond to the pre-approach material, either online or via phone, an Interviewer approached selected dwellings to conduct an interview or arrange a suitable time for interview.

At the first face-to-face interview contact with the household by an Interviewer, general characteristics of the household were obtained from a responsible adult member of the household (any responsible adult - ARA) using the survey instrument's Household Form. This information included basic demographic characteristics of all usual residents of the dwelling (e.g. age and sex) and the relationships between household members (e.g. spouse, son/daughter, not related).

From the information provided by the ARA regarding household composition, the survey instrument identified those people in scope and coverage of the survey and randomly selected one person aged 15 years or over to be included in the survey. A personal interview was conducted with the randomly selected person. If the randomly selected person lived with one or both parents, a parent may have been selected to be the household spokesperson and provide some information about the household (e.g. household financial information). If the dwelling contained no usual residents aged 15 years or more, the dwelling was not enumerated.

Children aged 15-17 years were interviewed with the consent of a parent or responsible adult. If they refused to allow the child to be interviewed personally, the interview would not proceed. The household spokesperson still completed the household spokesperson modules. In cases where the randomly selected person was aged 15-17 years old but did not live with a parent or guardian, Interviewers proceeded with the interview without obtaining parental consent.

In some instances selected adult respondents were unable to answer for themselves because of age, illness, intellectual disability or difficulty with the English language. In these cases, a proxy interview was conducted such that a person responsible for the respondent was interviewed on their behalf, provided the Interviewer was assured this was acceptable to the selected person. Where there were language difficulties, another person in the household may have acted as an interpreter if this was suggested by the respondent. If not, arrangements were made, where possible, for the interview to be conducted either by an ABS Interviewer fluent in the respondent’s own language or with an ABS interpreter. If the person could answer for themselves by using sign language etc. the questions were interpreted by another member of the household, only if the interpreter could assure the Interviewer this was acceptable to the intended respondent.

In order to obtain a personal interview with appropriate respondents, Interviewers made appointments to call back as necessary to the household. In some cases, appointments for callbacks were made by telephone; however, all interviews were conducted face-to-face. Interviews may have been conducted in private or in the presence of other household members according to the wishes of the respondent.

In cases where a respondent initially refused to participate in the survey, a follow-up letter was sent and a second visit was made to the respondent to explain the aims and importance of the survey, and to answer any particular concerns the respondent may have had about the interview. People excluded from the survey through non-contact or refusal were not replaced in the sample.

On average, the interview took 48 minutes per fully responding household.

Interviewers were primarily recruited from a pool of trained ABS Interviewers having previous experience with ABS household surveys. Those Interviewers who agreed to work on this survey underwent two days of classroom training and were provided with detailed written instructions aimed at emphasising the survey concepts, definitions and procedures in order to ensure a standard approach was employed by all Interviewers.

Each Interviewer was supervised in the field in the early stages of the survey, and periodically thereafter to ensure consistent standards of interviewing procedures were maintained. In addition, regular communication between field staff and survey managers was maintained throughout the survey via database systems set up for the survey.

Interviewers were allocated a number of dwellings (a workload) at which to conduct interviews. The size of the workload was dependent upon the geographical area. Interviewers living close to their workload area in urban areas usually had larger workloads. Overall, workloads averaged 23 dwellings, enumerated over a two week period.

The questionnaire was designed to be administered using standard ABS procedures for conducting population interview surveys with regard to the particular aims of the survey and of the individual topics within it, and to the methodological issues associated with those topics. Other factors considered in designing the questionnaire included the length of individual questions, the use of easily understood words and concepts, the number of subjects and overall length of the questionnaire, and the sensitivity of topics. Where appropriate, standard questions from previous ABS surveys were included.

Interviews were conducted using a Computer Assisted Interviewing (CAI) questionnaire. CAI involves the use of a notebook computer to record, store, manipulate and transmit the data collected during interviews. The CAI questionnaire for the 2014 GSS was based on the 2010 version and modified to incorporate new and changed survey content.

The questionnaire employed a number of different approaches to recording information at the interview:

• questions where responses were classified by Interviewers to one or more predetermined response categories. This approach was used for recording answers to more straightforward questions, where logically a limited range of responses was expected, or where the focus of interest was on a particular type or group of responses (which were listed in the questionnaire, with the remainder being grouped together under ‘other’);
• questions asked in the form of a running prompt i.e. predetermined response categories read out to the respondent one at a time until the respondent indicated agreement to one or more of the categories (as appropriate to the topic) or until all the predetermined categories were exhausted;
• questions asked in association with prompt cards i.e. where printed lists of possible answers were handed to the respondent who was asked to select the most relevant response(s). By listing a set of possible responses (either in the form of a prompt card or a running prompt question) the prompt served to clarify the question or to present various alternatives, to refresh the respondent’s memory and, at the same time, assist the respondent select an appropriate response;
• to ensure consistency of approach, Interviewers were instructed to ask the interview questions as shown in the questionnaire. In certain areas of the questionnaire, Interviewers were asked to use indirect and neutral prompts, at their discretion, where the response given was, for example, inappropriate to the question asked or lacked sufficient detail necessary for classification and coding.
   

A copy of the 2014 GSS questionnaire is available in the Data downloads section of the General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

As with all ABS surveys, the questionnaire was tested using experienced ABS Interviewers and applying the procedures and methods planned for the final survey. As described in Survey content, the testing program for the GSS included cognitive interviews and a Dress Rehearsal. The broad aims of the testing program were: to test the suitability of new and modified survey content to ascertain respondent reactions and identify any sensitivities associated with the survey content; and to test operational and procedural aspects of the survey such as interview time, the suitability of overall survey procedures, documentation and interviewer training. As a result of the testing program, the survey instrument was progressively improved and the methodology and survey procedures refined

Ideally, interviews would be conducted with all people selected in the sample. However, in practice, some level of non-response is inevitable. Non-response is classified as people who refuse to cooperate, cannot be contacted or are contacted but cannot be interviewed. It is important that response be maximised in order to reduce sampling variability and minimise bias. Sampling variability is increased when the sample size decreases. Consequently, bias can arise if the people who fail to respond to the survey have different characteristics from those who did respond.

The ABS sought the willing cooperation of selected households. Measures taken to encourage respondent cooperation and maximise response included:

• Information provided to selected households in the 2014 GSS, initially by a guide, registration letter and leaflet, explaining that their dwelling had been selected for the survey, the purposes of the survey, its official nature and the confidentiality of the information collected.
• Stressing the importance of participation in the survey by selected households, by explaining that each household selected represented a number of others similar in size, composition, location, occupation, lifestyle and health. Further explanation that the cooperation of those selected was important to ensure all households/people were properly represented in the survey and properly reflected in survey results.
• Stressing the importance of the survey itself, which measures the wellbeing of Australians and therefore helps plan and provide support to those groups in need.
• Stressing the confidentiality of all information collected. The confidentiality of data is guaranteed by the Census and Statistics Act 1905. Under provisions of this Act, the ABS is prevented from releasing any identifiable information about individuals or households to any person, organisation or government authority.
   

Through call-backs and follow-up at selected dwellings, every effort was made to contact the occupants of each selected dwelling and to conduct the survey in those dwellings. Interviewers made several call-backs before a dwelling was classified as ‘non-contact’. Call-backs occurred at different times during the day and days of the week to increase the chance of contact. If any person selected for the survey was absent from the dwelling when the Interviewer called, arrangements were made to return and interview at a later date. Interviewers made return visits as necessary in order to complete the questionnaire for the selected person in scope of the survey. In some cases, the selected adult within a dwelling could not be contacted or interviewed, and these were classified as non-contacts.

As described above, respondents who refused to participate were usually followed-up by letter and a subsequent visit. There were instances in which respondents were willing to answer some, but not all, of the questions asked, or did not know an answer to a particular question. The survey instrument was programmed to accept 'don't know' responses as well as refusals on sensitive topics, such as income. Respondents who refused or did not know an answer to only these sections of the questionnaire were classified as 'adequate complete'. Approximately 2,920 respondents (18.1% of eligible dwellings) did not provide one or more required answers but were deemed to have responded adequately.

The initial sample selection of approximately 18,574 private dwellings was reduced to approximately 16,145 dwellings due to the loss of households that had no residents in scope for the survey and where dwellings proved to be vacant, under construction or derelict. Of the eligible dwellings remaining, 80.1% responded fully (or adequately), which yielded a total sample from the survey of 12,932 dwellings. The following table shows the number of fully responding households achieved for each state and territory, and the corresponding response rate achieved in the 2014 GSS

Consistent with the aim of exploring the relative outcomes of people vulnerable to socio-economic disadvantage, the 2014 GSS sampling methodology targeted sample from low socio-economic areas, although the targeting was less extreme than it was for the 2010 GSS. For the 2010 GSS sample design the 2006 Census was used to target areas with higher concentrations of households experiencing multiple disadvantage, using additional benchmarks in the weighting process to compensate for the over sampling. In 2014, the GSS sample design utilised the Master Sample's layer of income-based socio-economic stratification to over-sample the low socio-economic strata relative to other strata. Selection areas were stratified by State, Part of State (Capital City/Balance of State) and then by the Socio-economic level, defined as follows:

• Low socio-economic stratum: for each State by Part of State group, the 20% of Base Frame Units (BFUs) with the highest proportion of low income households (weekly total Equivalised Household Income < $400) were put into the low socio-economic stratum. All of these BFUs had a proportion of low income households above some threshold. The threshold was different for each State by Part of State group, but was typically a proportion of between 0.25 and 0.4. BFUs on the Household Sample Frame were classified using Census 2011 weekly total Equivalised Household Income.
• High socio-economic stratum: a high income household was defined as having weekly total Equivalised Household Income > $1250. For each State by Part of State group, the 20% of BFUs (excluding those already assigned in step 1) with the highest proportion of high income households were put into the high socio-economic stratum. All of these BFUs had a proportion of high income households above some threshold. The threshold was different for each state by part of state, but was typically between 0.3 and 0.45 for metropolitan and between 0.2 and 0.3 for ex-metropolitan.
• Medium socio-economic stratum: the remaining 60% of BFUs were assigned to the medium socio-economic stratum. These BFUs do not contain a high proportion of either high or low income households.
   

In the 2014 GSS, people from the lowest strata had a higher probability of being selected in the sample.

The sample sizes differed between the 2014 and 2010 GSS. As noted above, the number of fully or adequately responding households achieved in the survey was 12,932 in 2014, compared with 15,028 for the 2010 cycle. The 2014 GSS had a smaller initial sample size (18,574 possible dwellings) compared with the 2010 initial sample size (19,576 possible dwellings). The achieved response rate was also lower in 2014 (80.1%) than in 2010 (87.6%), with a higher rate of survey non-response from eligible households in 2014 compared with 2010. These differences in sample size for 2014 and 2010 should be considered when comparing results. The following table shows the number of fully responding households achieved for each state and territory, and the corresponding response rate achieved in the 2010 GSS.

Computer based systems were used to process the data from the survey. Internal system edits were applied in the CAI instrument to ensure the completeness and consistency of the questionnaire and responses during the interview. The interviewer could not proceed from one section of the interview to the next until responses had been properly completed.

Workloads were electronically loaded on receipt in the ABS office in each state or territory. Checks were made to ensure interviewer workloads were fully accounted for and that questionnaires for each household and respondent were completed. Problems with the questionnaire identified by interviewers were resolved by office staff, where possible, using other information contained in the questionnaire, or by referring to the comments provided by interviewers.

Computer assisted coding was performed on responses to questions on country of birth, language, family relationships, educational qualifications and occupation. Geography data was also coded. The classifications used to code data can be found in the Methodology of the General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

Information collected by the survey, other than names and addresses, was stored on a computer output file in the form of data items. In some cases, items were formed from answers to individual questions, while in other cases data items were derived from answers to several questions.

During processing of the data, checks were performed on records to ensure that specific values lay within valid ranges and that relationships between items were within limits deemed acceptable for the purposes of this survey. These checks were also designed to detect errors that may have occurred during processing and to identify instances which, although not necessarily an error, were sufficiently unusual or close to agreed limits to warrant further examination.

Throughout processing, frequency counts and tables containing cross-classifications of selected data items were produced for checking purposes. The purpose was to identify any problems in the input data that had not previously been identified, as well as errors in derivations or other inconsistencies between related items. In the final stages of processing, additional output editing and data confrontation was undertaken to ensure GSS estimates conformed to known or expected patterns, and were broadly consistent with data from the previous GSS or from other ABS data sources, allowing for methodological and other factors that might impact comparability.

The procedures and checks outlined above were designed primarily to minimise errors occurring during processing. In some cases it was possible to correct errors or inconsistencies in the data which were originally recorded in the interview, through reference to other data in the record; in other cases, this was not possible and some errors and inconsistencies remain on the data file.

A four level data file was produced as outlined below. Some of these levels have a hierarchical relationship

1. Household
   2. Person
      3. Voluntary work
      4. Access to services

Broadly, each level provides the following:

• Household level - information about the household size and structure and household income details
• Person level - all demographic and socio-economic characteristics of the survey respondents, and most of the information they provided
• Voluntary work level - information about the characteristics of each episode of volunteering the survey respondent described
• Access to services level - information about the types of services that were difficult to access and the reasons why they were described as difficult.
   

A hierarchical file is an efficient means of storing and retrieving information which describes one to many, or many to many, relationships. For example, a person may have had more than one service they had difficulty accessing. In this circumstance, different record levels are used to store the details related to these incidents.

Most data from the GSS is available at the person level and describes personal characteristics, or characteristics of the household to which the person belongs.

Information about weighting, benchmarking and estimation can be found in the Methodology of the General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

Information about sampling error, standard error and significance testing can be found in the Technical Note in the Methodology of the General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

Errors made in giving and recording information during an interview can occur regardless of whether the estimates are derived from a sample or from a complete enumeration. Inaccuracies of this kind are referred to as non-sampling errors.

The major sources of non-sampling error are:

• errors related to the survey scope
• response errors such as incorrect interpretation or wording of questions, interviewer bias, etc.
• bias due to non-response as characteristics of non-responding persons may differ from responding persons
• errors in processing such as mistakes in the recording or coding of the data obtained.
   

These sources of error are discussed in turn below.

Some dwellings may have been inadvertently included or excluded because, for example, the distinction between whether they were private or non-private dwellings may have been unclear. All efforts were made to overcome such situations by constant updating of lists both before and during the survey. Furthermore, some people may have been inadvertently included or excluded because of difficulties in applying the scope rules concerning who was identified as usual residents, and concerning the treatment of some overseas visitors

In this survey, response errors may have arisen from three main sources: deficiencies in questionnaire design and methodology, deficiencies in interviewing technique and inaccurate reporting by the respondent.

Errors may be caused by misleading or ambiguous questions, inadequate or inconsistent definitions of terminology used, or by poor overall survey design (e.g. context effects where responses to a question are directly influenced by the preceding questions). In order to overcome problems of this kind, individual questions and the overall questionnaire were thoroughly tested in cognitive interviews and a dress rehearsal, before being finalised for use in the survey. As a result of both forms of testing, modifications were made to question design, wording, ordering and associated prompt cards, and some changes were made to survey procedures. In considering modifications, it was sometimes necessary to balance a better response to a particular item/topic against increased interview time or effects on other parts of the survey. The result is that in some instances it was necessary to adopt a workable/acceptable approach rather than an optimum approach. Although changes would have had the effect of minimising response errors due to questionnaire design and content issues, some will inevitably have occurred in the final survey enumeration.

Response errors may also have occurred due to the long nature of the interview, resulting in interviewer and/or respondent fatigue (i.e. loss of concentration). While efforts were made to minimise errors arising from deliberate misreporting or non-reporting by respondents (including emphasising the importance of the data and checking consistency within the survey instrument), some instances will have still occurred.

Recall error may also have led to response errors. Information recorded in this survey is essentially 'as reported' by respondents, and hence may differ from information available from other sources or collected using different methodologies. Responses may be affected by imperfect recall or individual interpretation of survey questions. Reference periods used in relation to each topic were selected to suit the nature of the information being sought; in particular to strike the right balance between minimising recall errors while ensuring the period was meaningful, representative (from both respondent and data use perspectives) and would yield sufficient observations in the survey to support reliable estimates. It is possible the reference periods did not suit every person for every topic, and that difficulty with recall may have led to inaccurate reporting in some instances.

A further source of response error is lack of uniformity in interviewing standards. Methods employed to achieve and maintain uniform interviewing practises included training and re-training programs, and regular supervision and checking of interviewers' work. These programs aimed to ensure a high level of response accuracy was achieved. An advantage of the CAI technology used in conducting interviews for this survey is that it potentially reduced non-sampling error by enabling edits to be applied as the data was being collected. The interviewer was alerted immediately if information entered into the computer was either outside the permitted range for a question, or contradictory to information previously recorded during the interview. These edits allowed the interviewer to query respondents and resolve issues during the interview. CAI sequencing of questions was also automated such that respondents were asked only relevant questions and only in the appropriate sequence, eliminating interviewer sequencing errors.

Some respondents may have provided responses they felt were expected, rather than those that accurately reflected their own situation. Every effort has been made to minimise such bias through the development and use of culturally appropriate survey methodology. Respondent perception of the personal characteristics of the interviewer can also be a source of error as the age, sex, appearance and manner of the interviewer, may influence the answers obtained.

One of the main sources of non-sampling error is non-response by persons selected in the survey. Non-response can affect the reliability of results and introduce bias. The magnitude of any bias depends upon the level of non-response and the extent of the difference between the characteristics of those people who responded to the survey and those who did not, as well as the extent to which non-response adjustments can be made during estimation through the use of benchmarks.

To reduce the level and impact of non-response, the following methods were adopted in this survey:

• face-to-face interviews with respondents
• the use of interviewers who could speak languages other than English (where necessary)
• follow-up of respondents if there was initially no response
• ensuring the weighted file is representative of the population by aligning the estimates with population benchmarks.
   

Of the dwellings selected in the 2014 GSS, 19.9% did not respond fully or adequately. As the non-response to the GSS was relatively low, the impact of non-response bias is considered to be negligible.

Errors may also occur during data processing, between the initial collection of the data and final compilation of statistics. These may be due to a failure of computer editing programs to detect errors in the data, or during the manipulation of raw data to produce the final survey data files; for example, in the course of deriving new data items from raw survey data or during the estimation procedures or weighting of the data file.

To minimise the likelihood of these errors, a number of quality assurance processes were employed, including:

• computer editing - edits were devised to ensure that logical sequences were followed in the questionnaires, that necessary items were present and that specific values lay within certain ranges. These edits were designed to detect reporting and recording errors, incorrect relationships between data items or missing data items.
• data file checks - at various stages during processing (such as after computer editing or after derivation of new data items) frequency counts and/or tabulations were obtained from the data file showing the distribution of persons for different characteristics. These were used as checks on the content of the data file, to identify unusual values which may have significantly affected estimates and illogical relationships not previously identified. Further checks were conducted to ensure consistency between related data items, and in the relevant populations.
• where possible, checks of the data were also undertaken to ensure consistency of the survey outputs against results of the previous GSS and data available from other sources.
   

Care has been taken to ensure the results of this survey are as accurate as possible. All interviews were conducted by trained ABS officers. Extensive reference material was developed for use in the field enumeration and intensive training was provided to interviewers in both classroom and on-the-job environments. However, there remain other factors which may have affected the reliability of results, and for which no specific adjustments can be made.

For a number of GSS data items, some respondents were unwilling or unable to provide the required information. No imputation was undertaken for this missing information. However, labour force status was imputed for about 150 people. For most of these people, other responses they had provided during the survey were used to derive labour force status with a high degree of accuracy. For the remainder, unaffected data collected for other people in the survey was used to apply a ratio distribution to derive labour force status. These changes had minimal impact on data quality. Where responses for a particular data item were missing for a person or household they were recorded in a 'not known' or 'not stated' category for that data item. These 'not known' or 'not stated' categories are not shown in the publication tables. However, the person or household within these categories has been included in table totals. The Methodology of the General Social Survey: Summary Results, 2014 (cat. no. 4159.0) contain a table showing the number and proportion of missing values for key GSS data items.

The publication, General Social Survey: Summary Results, 2014 (cat. no. 4159.0), presents summary results from the survey. The publication was released in June 2015 and is available free of charge on the ABS website. Data tables are available in spreadsheet format via the Data downloads section, and are predominantly national level data tables, although some tables show state and territory results. Estimates, RSEs of estimates, proportions and RSEs of proportions are presented for most publication tables.

Two tables (Tables 3 and 15) in the summary publication contain data for each state and territory. They are available in spreadsheet format, with estimates, RSEs of estimates, proportions and RSEs of proportions presented for each table. The three voluntary work tables in the summary publication (Tables 19, 20 and 21) have also been produced for each state and territory, although the same level of detail is not available for several of the data items. These tables (22, 23 and 24) are available from the Data downloads section of General Social Survey: Summary Results, Australia, 2014 (cat. no. 4159.0).

Supporting material is available to assist data users in analysing the data from the survey. A representation of the computer assisted interview questionnaire and prompt cards used in the GSS can be downloaded from the Data downloads section of the publication General Social Survey: Summary Results, 2014 (cat. no. 4159.0).

For users who wish to undertake more detailed analysis of the survey data, microdata from the 2014 GSS has been released in the form of a TableBuilder and Expanded Confidentialised Unit Record File (CURF). TableBuilder is accessible via the ABS website, using a secure log-on portal. The Expanded CURF is available via the Remote Access Data Laboratory, which is a secure Internet based data query service, and the ABS Data Laboratory. Information regarding the Expanded CURF is included in Microdata: General Social Survey, Australia (cat. no. 4159.0.30.004).

Special tabulations of GSS data are available on request on a fee-for-service basis. Subject to confidentiality and sampling variability constraints, tabulations can be produced from the survey incorporating data items, populations and geographic areas selected to meet individual requirements. Contact the National Information and Referral Service on 1300 135 070 or email client.services@abs.gov.au for further information. The ABS Privacy Policy outlines how the ABS will handle any personal information you provide to us.