As set out in Table 1.2, eight surveys were used in the project to collect data on individual experiences and institutional practices. Figure 1.2 outlines the relationships between the resulting data sets. While not the only basis for the findings presented in the later chapters, this large body of empirical research has been instrumental in developing a new overview of how public sector whistleblowing is managed.

Squares indicate surveys of agency-wide systems and procedures.

Circles indicate surveys of individuals.

Arrows indicate surveys designed to give consistent/comparable data.

First, the agency survey sought data on the extent, content and operation of whistleblowing procedures in agencies, as well as participation in further stages of the research. With the support of the partner organisations, this survey was distributed to almost all agencies of the four participating governments (793 agencies in total), resulting in 304 returns. The participating agencies represented a wide cross-section of government organisations in each jurisdiction, from small to large, covering a wide range of functions and portfolios. They included many of the most major departments and statutory authorities in each jurisdiction, government-owned corporations, the military and local governments of varying sizes.

Similar data on practices and procedures were sought from specialist integrity agencies in the jurisdiction—including partner organisations to the project—through a corresponding integrity agency survey.

Of the 304 agencies that answered the agency survey, 175 also supplied their written whistleblowing-related procedures, which were analysed in a separate procedures assessment, comparing their comprehensiveness and completeness using a 24-item rating scale. The research team developed this scale on the basis of existing literature and the Australian Standard, as discussed in Chapter 10.

The largest single research activity was the employee survey, in which 136 agencies initially agreed to participate, with 118 agencies ultimately doing so. This was a confidential, anonymous survey of a random sample of staff from each agency. While much of the survey design was original, it also drew on instruments used in previous research in Australia and internationally (including US Merit Systems Protection Board 1981, 1993, 2003; Zipparo 1999a, 1999b; Keenan 2002; Near et al. 2004). Target sample sizes were varied according to agency size, ranging from 1–2 per cent of all staff in very large agencies (more than 5000 employees) to 50–100 per cent of all staff in very small agencies (150 employees or less). From July to October 2006, printed surveys were distributed by agencies to approximately 23 000 public employees, with responses returned directly to Griffith University via a reply-paid envelope.

A total of 7663 public officials responded to the employee survey, providing the main data analysed throughout this book. It is the largest data set on whistleblowing ever gathered in Australia, and probably the largest per capita in the world. At times, these data are analysed in conjunction with the results of the agency survey (all 118 employee survey agencies also completed this), as well as the results of the procedures assessment (in which 102 of the employee survey agencies also participated).

In addition, 87 of the agencies volunteered to participate in further research as ‘case study agencies’, with 15 agencies selected by the research team for this role in May 2006. In this report, the further data drawn from these agencies will be reported in aggregate, with differences between organisations to be analysed in later publications. It is significant, however, for the data reported here, that this smaller group of 15 agencies is just as diverse as the larger group of 118 agencies from which the employee survey data set is drawn, with many of the results confirming that as a group, the experiences of the case study agencies remain in many ways typical of those of the total group. In all, 2116 responses to the employee survey were received from the case study agencies, meaning that while these agencies represented only 13 per cent of the larger group, their respondents accounted for 28 per cent of the total employee survey data set. Accordingly, interim results from the three further case study agency surveys are also reported here, not only for their interest in their own right, but because these are likely to be typical of organisational experience more generally.

The purpose of these three surveys was to gather targeted, in-depth information about the whistleblowing arrangements and management practices of these agencies and the experience of particular target groups of staff. In each case, respondents were also invited to volunteer for a confidential interview. Each survey was designed for consistency of approach with the employee survey and with each other.

The internal witness survey was a long, written questionnaire designed to elicit more extensive information from whistleblowers. These were recruited between November 2006 and August 2007 through invitations for individual staff to contact the research team (in confidence), distributed in three ways: by internal advertisement within each agency; by direct contact from agency management to known whistleblowers; and by a number of partner organisations (integrity agencies) to known whistleblowers relevant to the case study agencies. This recruitment strategy produced 455 expressions of interest to participate from these 15 agencies—a far lower number of whistleblowers than are known to exist within these agencies based on the results of the employee survey. This response itself tends to confirm that recruitment strategies that rely heavily on employees self-identifying as a ‘whistleblower’ are only ever likely to capture quite a small, even if important, proportion of the total whistleblower population. The data set was further limited by seeking information only in relation to reports of wrongdoing made between July 2002 and June 2004, to ensure that respondents’ experiences were relatively current and more likely to have been dealt with under current procedures, as well as reflective of a similar two-year period to that studied in the employee survey. Accordingly, only 242 individuals ultimately returned surveys, of whom only 114 fit fully within the requested parameters, but who nevertheless provided detailed, quality information.

The case-handler and manager surveys were slightly shorter questionnaires designed to elicit more extensive, comparable information from these two groups within the case study agencies. Case-handlers were defined as including: internal investigation, audit and ethics staff; human resource management staff; internal and external (for example, contracted) employee welfare and assistance staff; and union staff. Surveys were typically distributed to all, or a large proportion of, the identifiable case-handlers in each case study agency and to a random selection of managers (typically 5 per cent of the total population of managers or 150 individuals, whichever provided the larger figure). Data collection occurred from April to December 2007. This report includes analysis of the 828 responses received from both groups (13 per cent of all surveys distributed): 315 from case-handlers (many of whom also identified as managers) and 513 from managers (who did not necessarily also identify as case-handlers). In addition, a corresponding integrity case-handler survey was distributed to relevant case-handling staff from specialist integrity agencies in each jurisdiction (including partner organisations).

All data were collected in accordance with ethical approvals issued by the Griffith University Human Research Ethics Committee. All instruments were piloted in the final stages of design and refined in light of pilot results. Copies of the instruments are available on the project web site: www.griffith.edu.au/whistleblowing

As with all data sets, there are several limitations in the way in which it should be interpreted, as noted in relevant sections of the report. While care was taken to recruit a large sample size across the four jurisdictions, the study did not provide coverage across the full Australian public sector population, therefore its generalisability was constrained. Although a large number of agencies participated, clearly a great many also elected not to do so. As a result, it is arguable that the results most likely provide a best-case scenario of the ways in which whistleblowing is currently handled in the public sectors concerned, since agencies with poorer systems might have been more likely not to participate. Whether this is the case is obviously unknown; it is known that some agencies with significant integrity challenges did choose to participate, in order to find ways to rectify these problems. Some of the results are also sufficiently concerning to suggest they might well be realistic. Nevertheless, the sample of agencies was not total.

Within agencies, the extent to which the survey samples and respondent groups are representative of the demography of individual agencies has not been established. While participating agencies were requested to draw random, representative samples, whether or not they really did so was not within the researchers’ control. By sampling current employees, the major employee survey data set does not include former employees, such as those who might have observed and reported wrongdoing but have since left the organisation. This is a particular issue for Chapter 5, which includes analysis of the incidence of reprisals reported by whistleblowers. An effort is made there to estimate the alternative range of results that might have been expected had former employees been included in the sample.

The study relies on the self-reported perceptions of individual survey respondents, which are inevitably subject to errors in recall and specificity.

The bulk of survey results presented are quantitative, supplemented in several places by additional qualitative accounts provided by survey respondents as free text. It can be argued that quantitative research can only ever be of limited value in understanding the true nature and true lessons of any particular whistleblowing incident, given the complexity of all such incidents. In particular, it has been argued that only detailed case studies of more prominent cases will provide meaningful insights into the challenge of achieving better outcomes. Qualitative data can indeed play a useful role, and interviews with whistleblowers, case-handlers and managers from the case study agencies are providing the basis for further analysis in the second report. Here, however, quantitative research methods are used to paint a larger picture across thousands of individual reporting incidents, in order to help shift attention from whistleblowers as individuals to the performance of organisations in response to whistleblowing as a process. Whatever their limitations, the data reported in this book provide a new and different basis for understanding how whistleblowing is being managed in the Australian public sector.