Introduction

As in other organisations, public accountants may face work-related ethical issues and may also observe violations of workplace behaviour. Based on a survey of practicing audit staff in the US, Taylor et al. (2012) find the most common wrongdoings to be audit-related issues such as collecting insufficient audit evidence, inadequate documentation and review, truncating of small samples, false or premature sign-offs and failure to do adequate research. Prior studies have also identified wrongdoings in audit organisations to include under-reporting of time (Ponemon 1992; Akers and Eaton 2003) and shredding of working papers (Kaplan and Whitecotton 2001).

It has been argued that whistle-blowing, which refers to “the disclosure by organisation members (former or current) of illegal, immoral or illegitimate practices under the control of their employers, to persons or organisations that may be able to effect action” (Near and Miceli 1985, p. 4), is a particularly important mechanism in improving quality control in audit organisations (Taylor and Curtis 2010). Previous studies in the area establish that the intention to perform an act is influenced by various individual-level characteristics (Ajzen 1991; Graham 1986). Ajzen’s (1991) theory of planned behaviour proposes that an individual’s attitude towards the behaviour, subjective norms and perceived behavioural control are predictors of his/her behavioural intentions. Graham’s (1986) model of principled organisational dissent proposes personal cost of reporting, personal responsibility for reporting and the seriousness of the irregularity, as key variables influencing the reporting of the act. In the case of whistle-blowing, it is also contingent upon contextual or situational variables such as organisational climate and supervisor/co-worker support (Mesmer-Magnus and Viswesvaran 2005; Miceli et al. 2008; Kaptein 2011).

However, the literature has shown a general reluctance on the part of public accountants to report wrongdoing committed by colleagues (Finn and Lampe 1992; Kaplan 1995; Kaplan and Whitecotton 2001). This has been attributed to their belief that their organisations do not welcome reports of wrongdoing that challenges the organisational hierarchy (Mesmer-Magnus and Viswesvaran 2005), and their fear of retaliation through nullification, isolation, defamation, or expulsion (Dworkin and Baucus 1998). Berry (2004) and Kaptein (2011) argue that to stimulate employees to report wrongdoing, it is important to consider aspects of organisational culture. In fact, the latter highlights how different dimensions of organisational ethical culture affect different choices of actions or inactions in reporting wrongdoings in the organisation.

While organisational culture is important in influencing intended behaviour, it is equally important to consider individual-level characteristics. There is a dearth of research to date that looks at the moderating effects of individual-level characteristics and organisational culture, specifically perceived organisational support (POS), on public accountants’ intentions to report wrongdoing. We propose that the main drivers in stimulating the intention to act are the individual-level characteristics with POS acting as the catalyst. Therefore, in this paper, we examine the moderating effects of POS and individual-level characteristics on public accountants’ intentions to whistle-blow internally and externally.

The current study empirically tests part of a conceptual model advanced by Alleyne et al. (2013). Their model proposed relationships between five independent variables (attitudes toward whistle-blowing, perceived behavioural control, independence commitment, personal responsibility for reporting and personal cost of reporting), three moderating variables (POS, team norms and perceived moral intensity) and two dependent variables (internal whistle-blowing intentions and external whistle-blowing intentions). In this paper, we only test the moderating impact of POS to reduce the complexity in their model.

Since the majority of research on public accountants’ whistle-blowing intentions has been conducted in developed countries, particularly in the US and Canada (Miceli et al. 2009), and there are limited research in the context of emerging economies (e.g. Sims and Keenan 1999 in Jamaica; de Maria 2005 in Africa), we further contribute to the literature by surveying whistle-blowing intentions of public accountants in Barbados. In a small emerging economy with a population of about 300,000, the reputation of the audit firm and the integrity of the employees are vital for the firm’s survival (Alleyne et al. 2006). Therefore, it is in the best interest of audit organisations to establish an organisational culture that promotes ethical conduct and impedes unethical behaviour in the workplace (Kaptein 1998, 2011). More importantly, we argue that employees’ perceptions of the levels of organisational support may influence their choice of response.

Whistle-blowing in audit firms may be done internally through an ethics partner, line manager and confidential hotline, and/or externally to the profession, regulator or media, though authorisation varies across countries. For example, in the US, the government has implemented legislations (Sarbanes and Oxley Act 2002; Dodd-Frank Wall Street Reform and Consumer Protection Act 2010) to encourage and protect ‘effective’ whistle-blowers (Miceli and Near 2002). Generally, the accounting profession encourages its membership to whistle-blow internally, before considering external reporting (ACCA 2012). There is no legal protection for whistle-blowers in Barbados, with the two main mechanisms in place being the anti-money laundering legislation and a crime stoppers hotline for reporting wrongdoing.

The remainder of the paper is organised as follows. The next section presents the development of the hypotheses, followed by the research method employed. This is then followed by the discussion of the results and ends with the concluding remarks.

Theory and Hypotheses Development

We adopted the conceptual model advanced by Alleyne et al. (2013) and in the following paragraphs, the theoretical development of the five individual antecedents and the moderating variable (POS), as well as the specific hypotheses for this study are outlined. Figure 1 presents the conceptual model to be tested in this study.

Fig. 1
figure 1

The relationships between individual-level antecedents and the moderating effects of POS on internal and external whistle-blowing intentions

Full size image

Theory of Planned Behaviour

Ajzen’s (1991) theory of planned behaviour (TPB) has been widely used in studying a variety of behavioural intentions and behaviours, including whistle-blowing intentions (Park and Blenkinsopp 2009). Behavioural intention can be defined as the individual’s assessment of the likelihood of choosing a given behavioural alternative (Ajzen 1991). Ajzen (1991) argues that there is a strong positive relationship between behavioural intention and actual behaviour. Prior studies have used whistle-blowing intentions as a proxy for actual whistle-blowing due to the inherent difficulty of gaining access to actual whistle-blowers (e.g. Kaplan and Whitecotton 2001; Chiu 2003; Curtis 2006). In our study, whistle-blowing intention, the dependent variable, is the likelihood of the auditor actually engaging in the act of whistle-blowing. For the purposes of this study, we utilise two dimensions of TPB, i.e. attitude towards the behaviour and perceived behavioural control, which we discuss in more detail below.

Attitude Towards Whistle-Blowing Intentions

TPB suggests the existence of a relationship between a given behavioural action and the individual’s attitude towards the behaviour (assessment of the extent of approval or disapproval and the behavioural consequences) (Ajzen, 1991). Accordingly, a positive attitude will influence the intention to perform the act, while a negative attitude will constrain intention towards performing the act (e.g. Fishbein and Ajzen, 1975; Turrisi and Jaccard, 1992; Ajzen, 1991). Prior research has found attitudes to significantly influence ethical intentions, including intentions to whistle-blow (Flannery and May 2000; Bobek and Hatfield 2003; Carpenter and Reimers 2005; Buchan 2005; Fang 2006; Park and Blenkinsopp 2009). Within the auditing context, while Gibson and Frakes (1997) find no support for attitude influencing public accountants to report unethical action or wrongdoing, a later study by Buchan (2005) notes a supportive relationship. Hence, our first hypothesis is stated as follows:

H 1

Attitude towards whistle-blowing is positively associated with both internal and external whistle-blowing intentions.

Perceived Behavioural Control (PBC) and Whistle-Blowing Intentions

PBC refers to the individual’s perception of the level of ease or difficulty it would take to perform a specific behaviour, and is contingent on the presence or absence of opportunities and obstacles (Ajzen 1991). When the public accountant expects that (s)he can successfully overcome obstacles and perform a particular behaviour, the public accountant is more likely to act ethically (such as reporting) when faced with a dilemma (Beu et al. 2003; Alleyne et al. 2013). On the other hand, if the public accountant perceives that there is some difficulty in reporting which may be due to significant obstacles (such as unethical and punitive senior management and no trusted channels), s(he) is unlikely to report. Thus, the greater the individual’s PBC (i.e. ability to control the behaviour and overcome obstacles), the more likely (s)he will perform the behaviour in question (Ajzen 1991). Prior research has found support for PBC (self-efficacy)Footnote 1 in predicting ethical intentions (e.g. Flannery and May 2000). More specifically, Park and Blenkinsopp (2009) find PBC to be significantly and positively related to internal whistle-blowing intentions in the case of the police force in South Korea. For the purpose of this study, we argue that prior to taking action to report wrongdoing at the workplace, the public accountant is likely to assess how much control (s)he has over the situation and the possibility of succeeding in triggering a positive outcome (Bandura 1997). Thus, if the public accountant feels (s)he has greater control in overcoming obstacles such as unethical organisational pressures, then (s)he may blow the whistle. Stated as a hypothesis:

H 2

Perceived behavioural control is positively associated with both internal and external whistle-blowing intentions.

Gendron et al.’s (2006) Notion of Independence Commitment

Gendron et al. (2006, p. 170) defined independence commitment as “the extent to which the individual accountant considers auditor independence as a key attribute of the profession, and believes that regulatory standards of auditor independence (issued by the profession and/or external regulatory agencies) should be rigorously binding and enforced in the public accounting domain”. As professionals, public accountants have a commitment to act with integrity, and independence commitment is deemed a key and visible expression of integrity. Thus, to act and to be seen as being independent is a salient feature of the audit profession and an important part of its code of professional conduct. The public accountant who is imbued with independence values should be able to withstand pressures within and outside of the workplace. Hence, an important feature of professional and ethical conduct among public accountants is to have a high level of ethical commitment to independence.

Independence Commitment and Whistle-Blowing Intentions

Public accountants who possess a strong commitment to ethical values within the profession (in the form of being independent) are more likely to have a strong sense of moral obligation to act in the interest of the profession by reporting any questionable acts (Hall et al. 2005). A public accountant should have the ability to uphold the ethical values of the accounting profession in a manner that s(he) can resist pressures from colleagues and bosses when deciding to report any misconduct. In other words, a high level of independence commitment will lead public accountants to take action to report unethical behaviour, and hence, the next hypothesis is as follows:

H 3

Independence commitment is positively associated with both internal and external whistle-blowing intentions.

Graham’s (1986) Principled Organisational Dissent

Graham refers to principled organisational dissent (POD) as “a protest and/or effort to change the organisational status quo because of a conscientious objection to current policy” (1986, p. 1). Hence, when an employee observes circumstances in the workplace that somehow violates the values or principles that (s)he believes in, then POD may be manifested. In other words, the behaviour displayed is embodied by positive intentions based on personal principles. Dissenting allows for corrective feedback to monitor unethical and immoral behaviour, ineffectual organisational practices and policies, poor and unfavourable decision making and insensitivity to employees’ workplace needs and desires, hence signalling employee dissatisfaction or organisational decline. Two types of dissent that will be considered in this study are perceived personal responsibility for reporting and perceived personal cost of reporting.

Personal Responsibility for Reporting (PRR) and Whistle-Blowing Intentions

Drawing on the work of Graham’s (1986) POD, Schultz et al. (1993) propose that PRR influences an individual’s intention to act. Prior research has argued that responsibility for reporting an unethical act is related to personal ethical values, as well as feelings of socio-professional responsibility (Curtis 2006). In the audit profession, reporting the wrongdoing to those who can rectify is considered part of the public accountant’s responsibility (ISA 240 2010). The codes of professional conduct (e.g. AICPA and ICAEW) have stipulated that their memberships have collective responsibilities to the public, clients and colleagues, beyond the legal and regulatory responsibilities. Kaplan and Whitecotton (2001) and Schultz et al. (1993) find significant association between strong feelings of personal responsibility and the likelihood of blowing the whistle on questionable acts. Hence, the next hypothesis is as follows:

H 4

Personal responsibility for reporting is positively associated with both internal and external whistle-blowing intentions.

Personal Cost of Reporting (PCR) and Whistle-Blowing Intentions

Another important antecedent is the public accountant’s assessment of the PCR. Drawing on Graham’s (1986) POD work, Schultz et al. (1993) propose this variable, by arguing that it is the individual’s perception of the risk of retaliation from the members in the organisation that could affect one’s willingness to report wrongdoing. Curtis (2006) argues that the personal cost of reporting to the individual may come in the form of refusal of pay increases, unfair performance reviews, lack of peer support (e.g. ostracism), transfers to undesirable posts or jobs and possible firing. Further research has found support for a negative relationship between perceived PCR and reporting intentions (Arnold and Ponemon 1991; Schultz et al. 1993; Kaplan 1995; Kaplan and Whitecotton 2001). Thus, low PCR may lead to reporting, while high PCR may lead to no reporting. Hence, our next hypothesis is:

H 5

Personal cost of reporting is negatively associated with both internal and external whistle-blowing intentions.

Eisenberger et al. (1986) Organisational Support Theory

Organisational support theory (OST) “supposes that to determine the organization’s readiness to reward increased work effort and to meet socio-emotional needs, employees develop global beliefs concerning the extent to which the organization values their contributions and cares about their well-being” (Rhoades and Eisenberger 2002, p. 698). According to Levison (1965), actions taken by agents of the organisation are often viewed as indications of the organisation’s intent rather than attributed solely to the agents’ personal motives. Such personification of the organisation may be attributed to the organisation’s legal, moral and financial responsibility as well as organisational policies, norms and culture. On the basis of the organisation’s personification, employees view their favourable or unfavourable treatment as an indication that the organisation favours or disfavours them. In this study, we focus on one aspect of OST, i.e. perceived organisational support.

Perceived Organisational Support (POS)

Eisenberger et al. (1990, p. 51) describe POS as employees’ perception of “the extent to which the organisation values their contributions and cares about their well-being”. According to organisational support theory (OST), of which POS is based upon, employees will increase their efforts in assisting the organisation to attain its goals if they perceive that the organisation is willing and capable of giving a high level of support (Aselage and Eisenberger 2003). In this regard, organisational support may stimulate prosocial behaviour (i.e. concern for the well-being of the organisation without promise of reward) that may encourage public accountants to uphold organisational rules and ethics (Adebayo 2005). Thus, OST suggests that when employees feel valued by the organisation, higher levels of POS are achieved, which also has positive implications for organisational commitment and acting in the best interest of the organisation (Eisenberger et al. 1986, 1990, 2001; Rhoades and Eisenberger 2002).

In the case of audit firms, it is in the interest of the organisation to stimulate among its employees identification with the organisational ethics. An organisational culture of care with the right set of core values, which promote integrity, honesty, public interest and provision of support commitment, signals a high expectation of ethics among staff in the workplace. Consequently, this fosters ‘rationalized loyalty’ and encourages staff to report any observed unethical behaviour (Gouldner 1960; Vandekerckhove and Commers 2004). Hence, the intentions of public accountants to behave ethically are influenced by their perceptions of the firm’s expressed values (e.g. ethical policies) and the socio-emotional resources such as respect, care and expected degree of support that they will receive from the organisation. Therefore, if the public accountant strongly believes that the unethical behaviour warrants reporting and the level of POS is high, the public accountant will more likely choose to report the wrongdoing using internal channels such as his/her direct line manager, the ethics committee or the ethics hotlines set up by the audit firm. In contrast, if the public accountant strongly believes that the unethical behaviour warrants reporting but the level of POS is low, then (s)he will opt for external channels such as regulators or the media.

As mentioned earlier, POS by itself may not stimulate intention to report wrongdoing, but together with individual-level characteristics, it enhances the likelihood to report and also the channel to opt.Footnote 2 Hence, we anticipate the intention to whistle-blow either internally or externally to be influenced by the interactive effects of POS and the individual-level antecedents. The following sections discuss in further detail our hypotheses on the relationships between POS and individual-level characteristics on the intention to whistle-blow either internally or externally.

Moderating Effect of POS and Individual-Level Antecedents on Channels of Whistle-Blowing

Attitudes & POS

A public accountant imbued with a positive attitude towards reporting unethical behaviour at the workplace may consider the organisational support available before deciding which reporting channel to use. If there is a high level of POS, the public accountant may be more confident in utilising internal channels in upholding the ethical values of the organisation but if there is a low level of POS, the public accountant imbued with a positive attitude may opt for external channels if s(he) believes strongly that the wrongdoing should be reported. Hence, the interactive effect of positive attitudes and high levels of POS may cause the public accountant to whistle-blow internally, while the interactive effect of positive attitudes and low levels of POS may cause him/her to consider external whistle-blowing options. Hence, our first set of hypotheses is as follows:

H 1a

The relationship between attitudes toward whistle-blowing and internal whistle-blowing intention is significantly positive when POS is high.

H 1b

The relationship between attitudes toward whistle-blowing and external whistle-blowing intention is significantly positive when POS is low.

Perceived Behavioural Control & POS

When deciding whether to report wrongdoing at the workplace, the public accountant will consider how much control (s)he has over the situation in producing a positive outcome. If the public accountant feels (s)he has greater control in overcoming obstacles (such as unethical organisational pressures) and a high level of POS exists, then (s)he may select internal channels of whistle-blowing. On the contrary, if the public accountant feels that (s)he has greater control in overcoming obstacles, but there is a low level of POS, then (s)he may opt to blow the whistle externally. Hence, the next set of hypotheses is stated as follows:

H 2a

The relationship between perceived behavioural control and internal whistle-blowing intention is significantly positive when POS is high.

H 2b

The relationship between perceived behavioural control and external whistle-blowing intention is significantly positive when POS is low.

Independence Commitment & POS

Independence commitment is a core value for public accountants as they are expected to take action to report unethical behaviour. A public accountant with a strong sense of independence commitment is more likely to choose internal channels to report if there is a high level of POS. On the other hand, public accountants with a strong sense of independence commitment are more likely to choose to report externally when the level of POS is low. This leads us to the following set of hypotheses:

H 3a

The relationship between independence commitment and internal whistle-blowing intention is significantly positive when POS is high.

H 3b

The relationship between independence commitment and external whistle-blowing intention is significantly positive when POS is low.

Personal Responsibility for Reporting & POS

The reporting of a violation or wrongdoing committed by a colleagueFootnote 3 within the audit firm is a morally derived feeling of responsibility, as non-disclosure may result in losses for the organisation, the public and the profession. Reporting internally may help protect the firm’s reputation and the profession’s credibility, while reporting externally may increase loss of reputation and revenue for the audit firm and motivate members of the public to become suspicious of public accountants. We further argue that strong feelings of personal responsibility of reporting, augmented by high levels of POS, are more likely to drive public accountants to blow the whistle internally. On the contrary, a public accountant with strong feelings of personal responsibility of reporting will choose to report externally when the level of POS is low. Thus, our next hypotheses are as follows:

H 4a

The relationship between personal responsibility for reporting and internal whistle-blowing intention is significantly positive when POS is high.

H 4b

The relationship between personal responsibility for reporting and external whistle-blowing intention is significantly positive when POS is low.

Personal Cost of Reporting & POS

Research has shown that individuals with high PCR are less likely to blow the whistle (Kaplan and Whitecotton 2001; Curtis 2006). However, Curtis (2006, p. 194) argues that although high personal cost may hinder reporting of wrongdoing, this may be moderated by “belief in the existence of support and protection for dissidents, such as the belief that professionalism demands reporting behaviour”. Hence, a public accountant with high PCR in an organisation that has high POS may most likely choose the internal channel to report wrongdoing as s(he) feels that it is the right thing to do and trusts that the organisation will be supportive of her/his action in protecting the reputation of the organisation by seeking to resolve the problem internally rather than going externally. In contrast, if the support and protection system is perceived by the individual to be low, s(he) is more likely to blow the whistle externally due to the lack of trust in the organisation in supporting the individual’s action. Therefore, the next set of hypotheses is as follows:

H 5a

The relationship between personal cost of reporting and internal whistle-blowing intention is significantly positive when POS is high.

H 5b

The relationship between personal cost of reporting and external whistle-blowing intention is significantly positive when POS is low.

Table 1 presents a summary of our hypotheses, theoretical underpinnings and the expected relationships on the moderating effects of POS and various individual-level factors, on intentions to whistle-blow either internally or externally.

Table 1 Summary of theoretical underpinning, variables, hypotheses & expected relationship with whistle-blowing intentions
Full size table

Research Methods

Sample and Data Collection

The respondents in our survey comprised public accountants working in Barbados. The sample was drawn from a list of membership provided by the ICABFootnote 4 and the sample size was determined based on Green’s (1991) suggestion for regression analysis requiring a minimum of 114.Footnote 5 We contacted the audit partners of the four large accounting firms (with 295 audit staff), 14 small and medium size firms (with 89 audit staff), as well as 116 sole practitioners, to gain access to their staff.Footnote 6 We disseminated a total of 500 packages across the audit firms. Each respondent was given a package containing a cover letter, the general instructions letter and survey questionnaire, an informed consent form and a self-addressed envelope for returning the questionnaire.

A total of 268 completed questionnaires were received of which 42 were unusable due to missing data and outliers, resulting in a final usable sample of 226 (45.2 %) responses for analysis.Footnote 7 Of the 226 completed responses, 79 % were from the large firms (60.7 % response rate) and 21 % from medium and smaller firms (53 % response rate). Statistical tests showed that there were no significant differences between responses of public accountants from the large, medium and small firms. The possibility of non-response bias was also checked by comparing early and late respondents using Armstrong and Overton’s (1977) approach, and t-tests indicated no statistically significant differences on any of the independent and dependent variables.

Table 2 presents a summary of the demographic profile of respondents. As can be seen, out of 226 respondents, 54.9 % were female, and the mean age was 30.8  (standard deviation (SD) = 9.7) years. In terms of position, 74.3 % of the sample comprised audit seniors and junior audit staff. As for qualifications, 78.3 % held Bachelor’s degrees, while 74.3 % of the sample had professional qualifications, with 50 % of the sample completing the ACCA’s professional qualification.

Table 2 Summary of respondents’ profile
Full size table

Development of the Research Instrument

The research instrument for this study consists of two parts.Footnote 8 The first part requests respondents’ demographic information, such as gender, age, highest academic qualification, professional working experience, tenure and organisational position. The second part presents the scenario and the related questions. Given the difficulties in gaining access to subjects and observing actual ethical or unethical behaviours, the scenario approach, commonly used in the areas of accounting and marketing research (Silver and Valentine 2000; Sweeney and Roberts 1997), is adopted. This involves a description of a hypothetical situation and the subject is required to respond by placing himself or herself as an actor in the situation (Ferris et al. 1997). Since we are interested in understanding public accountants’ whistle-blowing intentions, this approach is helpful as it “helps to standardize the social stimulus across respondents and at the same time makes the decision-making situation more real” (Alexander and Becker 1978, p. 103). The scenario used in this study (see Appendix) is adapted from the scenario used in the study by Rau and Weber (2004) that highlights a violation of the principle related to “Integrity” in the IFAC’s Code of Ethics. The scenario describes an illegal act, which involves the issue of observing the shredding of audit documentation by a senior member in the audit firm.Footnote 9

Respondents are required to answer a series of questions related to the scenario. We adapt scales by Park and Blenkinsopp (2009) to measure the dependent variables, internal and external whistle-blowing intentions. Each whistle-blowing intention construct (internal and external) is measured using four statements (see Table 3). Respondents are asked whether they would report the shredding act within (to 4 specific internal targets) or outside (to 4 specific external targets) the organisation. These statements are measured on a 7-point Likert scale, ranging from 1 (not at all) to 7 (very much). Items are averaged to form a composite score. Higher scores indicate higher intentions to whistle-blow.

Table 3 Descriptive statistics for the dependent variables—whistle-blowing (reporting) intentions
Full size table

Using Park and Blenkinsopp’s (2009) scale, Attitude is measured by asking how true the respondent think with regard to the following five salient consequences of a public accountant reporting the wrongdoing or questionable acts of the audit organisation in the scenario [i.e. (1) preventing harm to the audit organisation, (2) controlling unethical behaviour, (3) enhancing public interest, (4) performing duty as an audit employee and (5) achieving moral satisfaction] (see Table 4A). Respondents are also required to evaluate the importance of each of the five consequences by responding to the question, “If you reported the shredding action, how important do you think the following consequences would be to you?”, based on a seven-point Likert-type scale. The scale measuring consequences of reporting ranged from 1 (not very true) to 7 (very true), and the scale measuring the importance attached ranged from 1 (not very important) to 7 (very important). The overall attitude score is derived by multiplying each salient consequence by its related importance and then averaging them. Higher scores indicate more positive Attitudes.

Table 4 Scale items and means for Attitudes toward whistle-blowing & PBC
Full size table

Using Park and Blenkinsopp’s (2009) scale, PBC is measured using eight items; four related to control factors and the remaining four related to perceived power. The control factors contain four statements that tap into the perceived difficulties that the individual is likely to encounter in the process of reporting, such as (a) the audit organisation hindering or ignoring the reporting, (b) difficulties likely to be faced when reporting, (c) reporting would not make a difference and (d) retaliation by the audit organisation (see Table 4B). The control factors are rated on a seven-point Likert-type scale ranging from 1 (not likely) to 7 (very likely). The perceived power items (based on an evaluation of each control factor) are rated on a scale ranging from 1 (not very important) to 7 (very important). Thus, PBC is derived by multiplying each perceived control factor by its related perceived power. The resulting product scores are combined to form the composite average score for PBC. Similar to Park and Blenkinsopp (2009, p. 551), the survey asks respondents to determine the difficulties in whistle-blowing, with lower responses indicating higher PBC, as respondents find it easier to gauge difficulty rather than ease of reporting. However, for data analysis, responses are recoded so that higher scores indicate higher PBC.

Table 5A presents the descriptive statistics for Independence commitment which is measured using four items adapted from Gendron et al. (2006). A sample item for this scale is “I believe that independence is one of the main foundations of the accounting and auditing profession”. Respondents are asked to reflect on their current organisation and in the context of the scenario, assess their level of Independence commitment. The responses for each individual item are measured using a 7-point Likert-type scale to rate each statement from 1 (completely disagree) to 7 (completely agree) and averaged to form an overall score, with higher scores indicating higher levels of Independence Commitment.

Table 5 Descriptive statistics for independence commitment, PCR, PRR & POS
Full size table

Table 5B presents the descriptive statistics for PRR and PCR. Both items are adapted from Schultz et al.’s (1993) single-item scales. The former asks respondents to assess their personal responsibility (i.e. duty or obligation) in reporting the shredding action, while the latter asks respondents to assess their personal costs (i.e. trouble, risk, discomfort) as a public accountant in reporting the shredding action. Each item is measured using a 7-point Likert scale, ranging from 1 (very low) to 7 (very high). On these two items, higher series indicate higher PRR and higher PCR.

Table 5C presents the descriptive statistics for POS which is measured using Eisenberger et al.’s (1997) 8-item short version of Eisenberger et al.’s (1990) scale. Based on comments arising from the pilot stage, an additional question is introduced to test the perceived level of ethical support that is valued by the organisation. Thus, a revised 9-item scale is used. Respondents are asked to reflect on their current organisation and in the context of the scenario, indicating the extent of their agreement on each of the nine items on a 7-point Likert-type scale, ranging from 1 (completely disagree) to 7 (completely agree). Scores across the nine items are averaged to form one composite POS score with higher scores indicating higher POS.

Given that this research uses self-reported data and respondents may tend to bias such reports by providing socially desirable responses, we test for social desirability bias (SDB), which may exist in research on ethical and whistle-blowing behaviour (Paulhus 1984; Shafer 2008). Similar to prior research which measures and controls for SDB, we utilise Paulhus’ (1989) 20-item impression management (IM) scale (e.g. Shafer 2008). The IM scale ranges from a low score of 0 to a high score of 20, where high scores are reflective of those respondents who are prone to exaggerate responses. Thus, higher scores indicate a greater propensity for participants to bias their responses in a socially desirable fashion.

The research instrument is piloted and refined through three stages: (1) interviewing four audit practitioners, (2) questioning four university students of auditing and (3) gaining feedback from three academics with practical auditing experience. To check for the presence of common method variance (Podsakoff et al. 2003), we use Harman’s (1976) single factor test to explore the variables of interest. Using principal component factor analysis, all factors loaded consistently within the specific constructs of interest. These findings indicate that common method variance may not have been an influential factor on the results of the study. Furthermore, all multi-item scales have Cronbach’s alphas above 0.81, thus indicating high internal reliabilities (Cronbach 1987).

Method of Analysis

Since we are interested in the interactive effects of POS and the five individual-level independent predictor variables [(POS* IV i=5 ), i.e. Attitude, PBC, Independence Commitment, PRR and PCR] on the dependent variables [IWB = internal whistle-blowing and EWB = external whistle-blowing], we use hierarchical multiple regression analysis as it allows the ordering of the variables to be entered into the analysis (Cohen and Cohen 1983; Baron and Kenny 1986). Consistent with prior research in the ethics literature, hierarchical multiple regression analysis is used to test for the moderating effects (e.g. Flannery and May 2000; Adebayo 2005; Chiu 2003; Barnett and Vaicys 2000).Footnote 10 We zero-centered the predictor and POS variables before running the regression analysis (Aiken and West 1991). We also control for the possible influences of demographic and other variables (SDB, gender and tenure), given their theoretical and empirical influences from the literature (Flannery and May 2000; Mesmer-Magnus and Viswesvaran 2005).

The test incorporates SDB, gender and tenure as control variables in step 1; the five independent variables (IV i=5 ) in step 2 (main effects); POS in step 3 (the moderator); and the interactions (POS* IV i ) in step 4 (the moderating effect). Thus, the regression equation is as follows:

$$WB = Controls + IV_{i = 5} + POS + (POS*IV_{i} ).$$

ModGraph-1 programme by Jose (2008) was used to plot the interaction graphs, based on information from SPSS output. To test whether the simple slopes of POS (i.e. low, medium and high levels of POS) differ significantly from zero, ModGraph-1 is used to calculate the simple slopes (betas, standard errors, t-values and p-values).

Results

Descriptive Analysis

Table 6 presents the means, standard deviations and Pearson correlation coefficients for the variables in the study. The results in the correlation matrix, and the fact that the variance inflation factors (VIF) do not exceed 2 in the regression models, indicate no major concern with multicollinearity.Footnote 11 The mean for the dependent variables indicates a higher preference by respondents for IWB (4.96) rather than EWB (3.45).

Table 6 Descriptive statistics and correlation matrix
Full size table

Hierarchical Regression Analysis

Table 7 presents the results of the hierarchical regression analysis used to test the hypotheses in this study. As can be seen in Table 7 (Step 2), Attitude is positively and significantly associated with only internal whistle-blowing (β = .180, p < .05), but not external whistle-blowing, thus our H1 is only partially supported.Footnote 12 The variable, PBC, is found to be positively and significantly associated with both channels of whistle-blowing (IWB—β = .179, p < .05 and EWB—β = .137, p < .05), thus fully supporting H2. IC and PRR are both found to be significantly and positively related to internal whistle-blowing (IC—β = .233, p < .01 and PRR—β = .265, p < .01), but not external whistle-blowing, thus partially supporting H3 and H4. PCR is found to be significantly and negatively related to only internal whistle-blowing intentions (β = −.169, p < .05), hence, partially supporting H5.

Table 7 Results of hierarchical regression analysis
Full size table

Interactive Effects

Table 7 (Step 4) and Modgraph plots (Figs. 2, 3, 4, 5, 6, 7, 8, 9) show the results of the interactive effects.Footnote 13 In terms of the interactive effects of Attitude and POS, results indicate the relationship between attitude and internal whistle-blowing intentions is positively intensified when POS is high (simple slope β = .691, t = 5.88, p < .01), and similarly, the relationship between attitude and external whistle-blowing intentions is positively intensified when POS is low (simple slope β = .310, t = 3.34, p < .01) as displayed in Figs. 2 and 3, respectively. Thus, our hypotheses H1a and H1b are fully supported. The results suggest that those with strong attitudes to report wrongdoing tend to prefer internal channels when there is higher organisational support and protection, but will opt for external channels when organisational support is deemed to be low.

Fig. 2
figure 2

Plot of interaction effect of ATT & POS on IWB

Full size image
Fig. 3
figure 3

Plot of interaction effect of ATT & POS on EWB

Full size image
Fig. 4
figure 4

Plot of interaction effect of PBC & POS on IWB

Full size image
Fig. 5
figure 5

Plot of interaction effect of PBC & POS on EWB

Full size image
Fig. 6
figure 6

Plot of interaction effect of IC & POS on IWB

Full size image
Fig. 7
figure 7

Plot of interaction effect of PRR & POS on IWB

Full size image
Fig. 8
figure 8

Plot of interaction effect of PCR & POS on IWB

Full size image
Fig. 9
figure 9

Plot of interaction effect of PCR & POS on EWB

Full size image

Similarly, the results in Table 7 (Step 4) show the significant interactive effects of PBC and POS, thus indicating that the relationship between PBC and internal whistle-blowing intentions is positively intensified when POS is high (simple slope β = .491, t = 5.02, p < .01), while the relationship between PBC and external whistle-blowing intentions is positively intensified when POS is low (simple slope β = .341, t = 3.94, p < .01). Thus, hypotheses H2a and H2b are fully supported. The results suggest that public accountants may be consequentialists, i.e. they will evaluate the consequences of their actions when making decisions on the choice of channel in reporting wrongdoing. The presence of effective organisational support and protection should give public accountants the opportunity to anticipate the expected outcomes, and hence, they are more likely to choose internal channels. On the other hand, when there is low organisational support and protection, public accountants would choose external channels. Figures 4 and 5 illustrate the interactions between PBC and whistle-blowing intentions moderated by levels of POS.

The results in Table 7 (Step 4) show the significant interactive effects of independence commitment and POS, thus indicating that the relationship between independence commitment and internal whistle-blowing intentions is positively intensified when POS is high (simple slope β = .607, t = 5.29, p < .01), thus supporting H3a. The relationship between independence commitment and external whistle-blowing intentions, moderated by low POS, is found to be insignificant, thus rejecting H3b. The results suggest that public accountants with commitment to independence will most likely choose the internal channel to report wrongdoing if there is high organisational support and protection. This is an important finding and may explain why wrongdoing within the audit profession seems not to be highly publicised.

As for the interactive effects of PRR and POS, results indicate PRR and internal whistle-blowing intentions to be positively intensified when POS is high (simple slope β = .643, t = 6.03, p < .01), hence supporting H4a. However, POS does not significantly influence the relationship between PRR and external whistle-blowing intentions, thus rejecting H4b. The results suggest that those with high levels of PRR will rationally choose internal channels especially when there is high organisational support. Figures 6 and 7 show the interactions between the two independent variables viz. Independence Commitment and PRR, on internal whistle-blowing intentions moderated by high levels of POS.

In the case of the interactive term PCR and POS, the result in Fig. 8 indicates the relationship between personal cost of reporting and internal whistle-blowing intentions to be positively intensified when POS is high (simple slope β = .172, t = .207, p < .05), thus supporting our H5a. Note that the simple slope in Fig. 8 also indicates the relationship between personal cost of reporting and internal whistle-blowing intentions to be negatively intensified when POS is low (simple slope β = −.418, t = −4.95, p < .01), thus further supporting our earlier H5. This is an important finding as it suggests that Barbadian public accountants in general would not blow the whistle when there is high PCR (H5), but in the presence of high POS, there is strong likelihood for them to whistle-blow internally.

However, the hypothesised relationship between PCR and external whistle-blowing intentions (see Fig. 9), when POS is low, is found to be insignificant (simple slope β = −.078, t = −1.11, p > .05), hence H5b is rejected. Interestingly, the simple slope in Fig. 9 indicates the relationship between personal cost of reporting and external whistle-blowing intentions to be positively intensified when POS is high (simple slope β = .217, t = 2.83, p < .01), a result similar to H5a. This is another important finding as it suggests that Barbadian public accountants with high PCR are also willing to blow the whistle externally if there is high POS. However, based on the results of the coefficients, there is a higher tendency to whistle-blow externally (simple slope β = .217, p < .01) than internally (simple slope β = .172, p < .05) in the presence of high POS. A possible explanation for the difference in the choice of channel may be driven by the individual’s perception that the external channel is more trustworthy and less riskier. However, this issue needs further probing to better understand the reason.

Concluding Remarks

The results of our findings highlight the importance of POS within audit organisations in relation to stimulating appropriate behaviour. Public accountants in Barbados demonstrate more willingness to engage in reporting wrongdoing internally if they believe there is full support from their organisation’s ethical mechanisms. Such support may be manifested in the culture of the organisation through not only encouraging ethical behaviour but also showing that the wrongdoing will not be tolerated and that no form of retaliation will be taken against the reporters of wrongdoing. Audit organisations can thus implement appropriate institutional arrangements (regulative, normative and culture-cognitive isomorphisms) that will influence the acceptable ethical behaviour and norms (DiMaggio and Powell 1983). This finding is consistent with Blau’s (1964) social exchange theory and Gouldner’s (1960) norm of reciprocity, which suggest that if employees feel that the organisation will support them, they are more likely to show loyalty and commitment by doing what is necessary to protect it from reputational damage.

Another important finding is that the mean scores in this study tended to be around the middle of the scale of 1 to 7 for internal reporting, but on the lower side for external reporting, thus suggesting a general reluctance by Barbadian public accountants to whistle-blow externally. This may be attributed to the influence of their culture of being family-oriented, hierarchical and one of seeking security, which may be a direct result of the remnants of African heritage, colonial past and religious values. Barbadians can be considered as more oriented to “being” rather than “doing”, which suggests a need to be a part of one’s personal and professional ties. In addition, Barbadians accept unequal distribution of power as appropriate.Footnote 14 Similarly, Stoddart (1995) argues that institutions such as religious and educational bodies encourage obedience which reinforces the beliefs and practices required to maintain the status quo in society. However, in the case of high PCR, Barbadians have a higher tendency to whistle-blow externally than internally in the presence of high POS as they feel more secure with the external channel.

A number of policy implications can be drawn from the findings. Since Barbadian public accountants prefer internal reporting, befitting the country’s culture, audit organisations should enhance their internal reporting channels by implementing organisational support such as anonymous hotlines, having an open-door policy, availability of ethics and risk managers for reporting the wrongdoing, and setting clearly defined procedures within the organisation. The creation of a supportive ethical culture within the organisation as well as purposive ethics training may encourage public accountants to blow the whistle internally. An investment in organisational support for internal reporting is necessary, as it responds to the moral needs of Barbadian public accountants. High levels of perceived organisational support and organisational commitment should encourage the public accountants to do all that is possible in avoiding harm to the organisation, profession and society. However, having perceived organisational support alone may be insufficient, as long as personal cost remains high and the code of ethics stays silent on enforcing and protecting the public accountants’ rights in reporting wrongdoing. The profession stands to benefit from supporting ethical behaviour rather than enabling wrongdoing, which may be costly to the audit organisation in the long term.

Audit organisations need to ensure that the recruitment process selects individuals who are ethically sensitive, have the capacity to stand up for their beliefs and be able to report questionable acts committed by colleagues. For example, the recruitment process can include the use of personality and integrity testing, which should assist in improving the quality of potential entrants to the profession. In addition, human resource departments in audit organisations can use training programmes designed to sensitise staff on the steps required to resolve ethical conflicts, and raise the level of moral development, enhance professional attitudes and increase personal responsibility for reporting.

Audit organisations need to implement measures and appropriate interventions that help improve attitudes toward whistle-blowing, the level of control over reporting and personal responsibility in an effort to reduce fear of reprisals. Senior management in audit organisations must ensure that the right values (e.g. morality, openness and ethics) are entrenched within the organisation, so that audit staff can feel that there is a personal responsibility to report. There should be procedures and policies to prevent retaliation, possibly through an independent committee outside of the organisation for staff to make their reports. These interventions can assist in creating an ethical environment, which can only lead to the enhancement of the organisation’s image and reputation.

There are several limitations to this research. This study focused primarily on public accountants to determine their willingness to report on senior colleagues. Although every attempt was made to make the sample representative of the population, there is still potential for sampling error. This study also sought to measure whistle-blowing intentions and not actual behaviours, and did not examine the influence of national culture on whistle-blowing intentions, given that some cultures, such as in the US, sensationalise the phenomenon. Future research may empirically consider looking at the impact of groups on the whistle-blowing process in actual field settings since most of the research on whistle-blowing has been conducted using questionnaires, experiments and student samples. In addition, rich information can be obtained from using qualitative methodologies such as focus groups.