Fairly and Justly? Are Employment Tribunals Able to Even Out Whistleblowing Power Imbalances?

In Britain, Employment Tribunals (ET) adjudicate on whistleblowing legislation. They do so with the overriding aim to adjudicate cases fairly and justly, by hearing parties on an equal footing. This paper presents research questioning this rule-of-law assumption vis-a-vis power imbalances that relate to whistleblowing. Using multinomial logistic regression analysis, we analyse all cases at ET in England and Wales between 2015 and 2018, that included a whistleblowing claim and that went to preliminary hearing or beyond. We find that several variables have an effect on the relative representational strength (RRS) at ET, but not on the outcome of the whistleblowing claim. However, whistleblowing claims brought in combination of discrimination claims (41%) have lower RRS and less favourable outcomes for the whistleblowing claim.


Introduction
Extant research suggests whistleblowing is a process marked by power imbalances. The power imbalance most often starts when a worker raises a concern about wrongdoing to someone higher in the organisational hierarchy (Vandekerckhove & Phillips, 2019). Further, a whistleblower's lack of power is inherent in the most common definition of whistleblowing, namely 'a disclosure made by an organisational member about illegal, illegitimate or unethical practices that fall under the responsibility of the organisation, to someone who can effect action' (Near & Miceli, 1985, p. 4). As Kenny et al. (2020) point out, whistleblowing to 'someone who can effect action' implies that the whistleblower themself cannot stop or correct the malpractice but instead calls upon someone else to step in.
Whistleblower power attributes further determine whether change will occur-i.e. wrongdoing is stopped (Near & Miceli, 1995) and whether the whistleblower experiences retaliation (Mesmer-Magnus & Viswesvaran, 2005, Skiveness & Trygstad, 2010, Vandekerckhove & Phillips, 2019. Those who experience retaliation can make use of whistleblower protection legislation to seek redress. In that sense, the judicial system is the whistleblower's final means to seek justice. This justice is based on the assumption that adjudicating institutions, through their processes, can compensate for power imbalances that led to the injustice brought to them to adjudicate on. In other words, if power imbalances lead to whistleblower detriment, then whistleblowers can use the legislation to have their case heard in a context where all are equal. Such is the assumption of the rule-of-law. Our research addresses issues of power and justice in the context of work, and is as such pertinent to the field of business ethics. More specifically, our research contributes to current discussions on institutional reform to implement whistleblowing legislation (Abazi, 2019(Abazi, , 2020Loyens & Vandekerckhove, 2018;Vandekerckhove, 2006). This paper questions whether the rule-of-law assumption that an adjudicating institution can hear parties on an equal footing despite power imbalances, holds for whistleblowers in Britain. The paper is based on an analysis of judgments from Employment Tribunals (ET) in England and Wales on whistleblowing cases over a period of 4 years to test whether British ET, as an adjudicating institution, succeed in evening out power imbalances relevant to whistleblowing. More precisely, we 1 3 distinguish power imbalances relevant to whistleblowing prior to proceedings at ET, and power imbalances at ET. The former concerns factors that make it more likely that whistleblowing with one's employer leads to detriment; the latter concerns the relative representational strength (RRS) of claimant and defendant at ET. We make assumptions and develop hypotheses around how these factors might affect the outcome of the whistleblowing claim, i.e. how ET adjudicate the whistleblowing claim.
The paper is structured as follows. The next section discusses in more detail how the whistleblowing literature looks to notions and theories of power. The section after that explicates our assumptions about whistleblowing cases at ET and develops the hypotheses. Then, in the methods section, we briefly set out the British whistleblowing legislation and the route to adjudication at ET. We also describe how we collected and coded the data, and we provide the preliminary tests for our multinomial logistic regression analysis (MLRA), which we use to test the hypotheses. We then present the results of the MLRA and come to our findings for the assumptions and hypotheses. In the section after that, we discuss our findings, clarifying our contribution to current scholarly debates and deriving implications for institutional reform.

Whistleblowing and Power
The most commonly used definition of whistleblowing in scholarship (see Brown et al., 2014) is that of Near and Miceli (1985, p. 4) who define whistleblowing as a disclosure about wrongdoing 'to persons or organisations that may be able to effect action'. In a later paper, these same authors explain how whistleblowing is a process best explained through theories of power, more precisely: the use of power by one or more social actors over others in order to change the behavior of some or all members of the organisation (Near & Miceli, 1995, p. 686) Near and Miceli (1995) develop a model for predicting effective whistleblowing, i.e. whistleblowing that 'results in wrongdoing cessation' (p. 680). Their model comprises of five factors, namely characteristics of whistleblowers, recipients, wrongdoers, wrongdoing and organisation. Affecting the factors are sets of individual and situational variables. Near and Miceli (1995) draw on power theories to link these variables. The individual variables include a whistleblower's expertise and experience (resource dependence theory, Pfeffer & Salancik, 1978), the extent to which a whistleblower is seen as credible, confident and competent (minority influence theory, e.g. Greenberger et al., 1987), and a whistleblower's referent power, status or position in the organisation (individual power bases, French & Raven, 1959). Two of the situational variables the model uses are first, that of the organisation's power in its environmentthe extent to which an organisation will feel threatened by its environment-which Near and Miceli (1995) base on resource dependence theory. The second situational variable we deem relevant for our purposes here, revolves around the wrongdoing that is being reported. Near and Miceli (1995) argue that whistleblowers who report wrongdoing for which there is clear evidence and which is unambiguously illegal or illegitimate, will be harder for the organisation to neglect or retaliate against.
It is important to note that Near and Miceli's (1995) model of effective whistleblowing relates to internal whistleblowing. Our research, however, relates to whistleblower detriment. A whistleblower who has been effective in bringing about change and is not retaliated against-i.e. who has been credible and powerful-will not need to call upon the law to seek justice through an ET. The implication is that we develop our hypotheses based on a number of assumptions about the transition of 'whistleblowing at work' to 'whistleblowing in ET'. We first explicate these assumptions before developing our hypotheses. Our assumptions about our sample are based on the inverted propositions of the Near and Miceli model. Thus, where Near and Miceli (1995) have the proposition that the more powerful a whistleblower is, the more likely it is that their whistleblowing will be effective, the assumptions for our research are that in our sample of whistleblowers at ET, the whistleblowers with attributions of 'less power' will be overrepresented. Therefore, our general assumption is that: A: whistleblowers that Lack Power Will be Overrepresented at ET When it comes to operationalise power of a whistleblower, extant literature has used tenure and gender as two independent variables. Empirical studies have shown mixed results (e.g. Near & Miceli, 1996;Sims & Keenan, 1998;Skiveness & Trygstad, 2017). However, a meta-study of predictors and correlates of whistleblowing behaviour and of retaliation against whistleblowers (Mesmer-Magnus & Viswesvaran, 2005) draw on resource dependency theory and individual power base theory to assert that tenure can be seen as an operationalisation of the power variable. Mesmer-Magnus and Viswesvaran (2005) hypothesize a positive correlation between tenure and successful internal whistleblowing. The implication for whistleblowers going to ET is that we would expect those with a lower tenure to be overrepresented. The caveat is however, that Mesmer-Magnus and Viswesvaran (2005) also expect tenure to correlate positively with actual whistleblowing behaviour. This means that workers with a higher tenure are more likely to blow the whistle in the first place. That might of course even out the expected effect of low power on the need to seek justice at ET. Nevertheless, we assume that: A1: whistleblowers at ET Will Have Low Tenure There is also extant whistleblowing research that operationalises whistleblower power as gender. Gao and Brink (2017) note that studies of whistleblowing intentions in an accounting context do not find significant correlations of work experience or gender on whistleblowing intentions. However, Gao and Brink (2017) note that gender can interact with other variables that can compensate a lower power status. For example, Gao and Brink (2017) find that gender interacts with type of whistleblowing channel, suggesting that women report wrongdoing more if they can do so via an anonymous channel. Researching whistleblowing in a higher education context, Horbach et al., (2020) find that the less powerful organisation members are 'younger employees, people with temporary work contracts, women, or people lower in the organisation's hierarchy' (pp. 1598-1599, emphasis added). It leads Horbach et al. (2020) to hypothesize based on power base theory (French & Raven, 1959), that women are less likely to blow the whistle, but if they do blow the whistle they will be less successful than men. Hence, with a similar reasoning and caveat as with tenure, we assume that: A2: female Whistleblowers Will be Overrepresented at ET Another proposition we will rely on from the Near and Miceli (1995) model, is that whistleblowers are more likely to be successful in getting their disclosure addressed and not being retaliated against in an organisation with less power over their environment. The reasoning is that weak organisations fear the regulatory or competitive environment and hence will prefer to listen to the whistleblower rather than see the whistleblower raise the issue outside of the organisation. The implication is that for our research on whistleblowing cases at ET, we expect an overrepresentation of employers that do not have to feel threatened by their environments. Thomas (2020) considers how whistleblowing is affected not only by the pre-existing distribution of power inside an organisation but also outside of the organisation. Hyde (2016) discusses this for whistleblowing in a healthcare context and argues that there can be wider institutional interests to silence whistleblowers. Hyde (2016) asserts that the power of the wider institutional setting is often overlooked in research. We concur with that assertion. In the absence of extant whistleblowing literature operationalising organisational power vis-à-vis its environment, we speculate here with regard to sector. In general, a public sector organisation is less threatened by its environment for its survival. Unlike private sector organisations, there are no competitors to whom market share can be lost. Public sector organisations are of course regulated, as are private sector organisations, but it is far less likely that an imposed fine would cripple a public sector organisation. Hence, public sector organisations have less to fear from their environments. We, therefore, expect public sector whistleblowers to be less successful in their whistleblowing disclosure than their counterparts in the private sector. In addition, following Hyde's (2016) assertion, we can expect the public sector to have wider institutional interests to silence a whistleblower, for example in an attempt to avoid tarnishing the reputation of the public sector in a context of increasing privatisation of public services. We thus assume that: A3: whistleblowers from the Public Sector Will be Overrepresented at ET Another proposition from the Near and Miceli (1995) model revolves around the alleged wrongdoing that whistleblowers report. The proposition is that internal whistleblowing is more likely to be successful when whistleblowers report activity that is clearly illegal and unambiguous. Thus, for our research on whistleblowing cases at ET, we expect the more ambiguous wrongdoing to be overrepresented at ET. When a claimant files a claim at ET, they can combine different claims in their case, namely any claim related to legislation that falls within the remit of ET. These include claims that the employer breached procedures of handling an internal dispute fairly, claims that the employer breached wage regulation, claims related to discrimination, and claims related to non-standard forms of employment. Thus, a whistleblower at ET will claim that they made a protected disclosure and suffered a detriment as a result (whistleblowing legislation) and may make additional claims, either as the matter of their whistleblowing or as a form of retaliation. Out of the four categories of additional claims a whistleblowing case may bring-internal procedures, wage regulation, non-standard forms of employment, discriminationthe discrimination claims are based on complex legislation (BEIS, 2020a). We thus assume that: A4: whistleblowers Who also Claim Discrimination Will be Overrepresented at ET Noting that whistleblowing is typically comprised of a number of attempts to raise a concern rather than a oneoff event, Vandekerckhove and Phillips (2019) researched whistleblowing sequences, to find that patterns depend on the whistleblower's formal power in the organisation and how recipients respond. The overall finding, however, suggests that successful whistleblowing depends on 'the whistleblower's ability to break the organisation's control over the process' (Vandekerckhove & Phillips, 2019, p. 201). Thus, what drives a whistleblower to blow the whistle again, to a different recipient, is the search to get out of the power imbalance that makes their whistleblowing ineffective and unsafe. Thomas (2020) writes that one of the characteristics of whistleblowing is its appeal to a central or external power. Park et al. (2020) researched motivations of external whistleblowers using a means-end-chain approach. They find that 'power to change' was the most important attribute of whistleblowing. Hence, our assertion here is that whistleblowers take their employer to ET in a call upon the ET judge to speak justice, which is an 'appeal' to a central power (Thomas, 2020) to even out the power differences that has made the whistleblower ineffective and retaliated against.
The ET is designed to do just that. It operates according its 'Rules of Procedure' (National Archives, 2013), a UK Statutory Instrument. This legislation stipulates-in Schedule 1 Rule 2-that the overriding objective of these Rules is to enable Employment Tribunals to deal with cases fairly and justly. Dealing with a case fairly and justly includes, as far as practicable-(a) ensuring that the parties are on equal footing, […] One of the aims of this paper is to test whether the British ET is able to meet its overriding objective in whistleblowing cases. We build our hypotheses on the basis that it does, namely that all of the power attributions we used to formulate our set of assumptions will not show significant patterns in relation to the outcome of the whistleblowing cases. That is, if ET is able to 'deal with cases fairly and justly' by 'ensuring the parties are on equal footing', we expect the outcome to relate to the merit of the case, but not to any of the power imbalances that the whistleblower comes to ET with. We thus hypothesize: H1 outcomes of whistleblowing claims at ET do not depend on the claimant's gender, tenure, sector, or a combination of claims The 'equal footing' the ET promises as its overriding objective does not merely relate to how the claimant comes to ET-i.e. driven by power imbalances in the employment relationship-but also how a power imbalance might manifest itself during the ET hearings. In other words, we also need to look at how claimants themselves attempt to address the power imbalance and how ET is able to deal fairly and justly with that imbalance.
One way to address the power imbalance is for a claimant to hire legal representation. Legal representation be that bar or solicitors have expert power and legitimate power in their mastery of the case. While having such expert power is desirable for claimants, the ability of claimants to secure such support can be problematic due to claimants not being able to afford to pay for legal advice (BEIS, 2020a) and a lack of legal aid. However, recent research by BEIS (2020a) found that claimant legal representation has increased from 33 to 41% for cases generally. Furthermore, when a claimant does seek advice, lawyers are the most common source of advice (BEIS, 2020a). However, the majority of claimants do not have legal representation. Pyper (2017) notes that for the period April 2011 to March 2013, only 46% of claimants had some form of representation. Public Concern at Work (PCAW, 2016) found similar results when examining whistleblowing cases only; 56% of claimants in whistleblowing cases did not have legal representation. Adler (2006) and Urwin et al. (2014) suggest there is an 'arms race' where opposing parties are keen to replicate the others' representation in order not to give an advantage to the opponent. The notion 'equality of arms' is an expression of the extent to which the type of representation of claimant and defendant in an ET case is balanced in terms of expert power and mastery of ET procedures. In this paper, we use 'relative representational strength' (RRS) to capture equality of arms. We define RRS as the claimant's representational strength relative to the employer's representational strength. Whilst legal representation is seen as the strongest (Latreille et al., 2005;William et al., 2019) it is also expensive. Hence it is likely that employers will be able to afford this type of representation more often than whistleblowers. However, if the ET can-as it promises-ensure that in the way it deals with cases it treats the parties 'on equal footing', then RRS should not affect the outcome of a whistleblowing case. We thus hypothesize: H2 relative representational strength (RRS) at ET will not affect the outcome of whistleblowing claims We noted that the reason a claimant might struggle to find legal representation and thus achieve the same representational strength as the employer, is financial (BEIS, 2020a); it is expensive to get legal representation. We will assume that is the main reason and test for that. One way to operationalise this is to test whether any of the power attributes we used in the previous hypotheses as independent variables, determine the RRS in whistleblowing cases at ET. If they do, that would suggest that power imbalances in the employment relationship are reproduced at ET through the type of representation. Consistent with our previous line of reasoning, we hypothesize: H3 in whistleblowing cases at ET, the claimant's gender, tenure, sector, or a combination of claims will not affect the relative representational strength (RRS) Figure 1 summarises the assumptions and hypotheses developed in this section.
Our assumptions and H1 relate to power imbalances relevant to whistleblowing prior to proceedings at ET. H2 and H3 relate to power imbalances at ET.

British Whistleblowing Legislation and the Route to Adjudication
The purpose of whistleblowing legislation is to protect employees from detrimental treatment and unfair dismissal which could occur because they have made protected disclosures (Lewis & Bowers, 2018). Such treatment has legal protection in many countries in the form of whistleblowing legislation that stipulates what constitutes a protected disclosure and how protection can be claimed. In Britain, whistleblower protection was enacted in 1999 by the Public Interest Disclosure Act 1998 (PIDA). It protects workers from detriment and automatic unfair dismissal if they have raised a concern which they had reasonable grounds to believe was in the public interest, either to their employer or a prescribed person (e.g. regulator, police or MP), and under certain conditions also for wider disclosures (for example to journalists). PIDA is implemented through the Employment Rights Act 1996. Claimants can seek adjudication before an ET.
In order to bring any claim to an ET a claimant must have knowledge of the law (Meager et al., 2002), the capacity to claim (Corby, 2015) and not fear retaliation (Lewis, 2017). Particular issues arise for whistleblowing claims because research shows the majority of people are unaware of whistleblowing legal protection and also fear retaliation (Mangan, 2014). Irrespective of legal knowledge, all claimants must follow a prescribed process. Initially a claimant must notify the Advisory, Conciliation & Arbitration Service (ACAS) that they wish to take a case to ET. ACAS will then attempt to negotiate an 'early conciliation' settlement between employer and claimant. Our study, however, does not include any data on early conciliation. Our sample only included whistleblowing cases that were lodged at the ET and went to a preliminary hearing or beyond between January 2015 and December 2018 inclusive in England and Wales.
If early conciliation is unsuccessful the claimant can take their claim further by submitting an ET1 form, which is a daunting process (McDermont & Busby, 2012). The ET1 form must be submitted, in the case of whistleblowing claims, within 3 months minus 1 day of the incident occurring. If the claim is lodged after 6 months, the judge makes a decision whether they will allow the claim to proceed depending if it was reasonably practicable for the claim to have been lodged in the time limit. After the ET1 has been lodged, the employer has a chance to respond using another official form, the ET3. At this point, the case will proceed to a preliminary hearing where a judge, sitting alone, will rule on preliminary issues such as if the case is brought within the strict time limits. If the case is not dismissed at a preliminary hearing, the case will progress to a full hearing. The full hearing is held in front of a judge and two lay members, one representing employers and one representing employees. Despite recent changes to the ET rules stating judges can sit alone for certain cases, such as wage related cases and unfair dismissal, whistleblowing cases still require a tripartite panel of a judge and two lay members.

Data Collection
The data includes all ET cases in 2015-2018 that claimed Public Interest Disclosure and went to a preliminary hearing or beyond in England and Wales. Public Interest Disclosure claims are categorised as open track claims and claimants had to pay a £250 lodging fee and a £950 hearing fee. Fees were introduced to shift the cost burden of the ET system onto those who use it and to discourage vexatious claims (Dickens, 2014). Despite the desire to reduce the number of vexatious claims, research shows that success rates of claims have fallen since fees were introduced, suggesting that genuine cases were put off going to ET as much as vexatious claims (Tetlow, 2017). However, due to a landmark case taken by Unison in late July 2017, fees were ruled to be unlawful and the fee system was scrapped.
The case judgments for 2015-2016 are held at a central archive in Bury St Edmunds, England which has a basic electronic database of all claims in England and Wales. This archive was searched by the first author for all cases claiming Public Interest Disclosure. Searching the database resulted in box files where the paper judgment for each case was stored and retrieved. This judgment was scanned, and a copy made. For the cases from 2017 and 2018 new Ministry of Justice electronic database was searched and cases downloaded.

Data Coding
After collecting all judgments, they were sorted into cases that were withdrawn before a preliminary hearing and those that went to a preliminary hearing or beyond (n = 603 cases). The cases that did not go to a preliminary hearing were not included in the sample as they contained very limited information; only respondent, claimant and jurisdiction codes. The sample cases were subject to content analysis using a code book developed by the authors. The code book was based initially on the coding used in the Survey of Employment Tribunal Applications (Buscha et al., 2012) and expanded to include additional interesting variables specific to whistleblowing cases, such as'was the claimant subject to a detriment?' Each case was coded and after data cleaning, was analysed using SPSS. The coding was carried out by the first author and research assistants. The interrater reliability was 97% and any disagreements between coders was resolved through discussion involving both authors.
In order to assess RRS each case was coded for type of representation from the least powerful type of representation i.e. no representation to the most powerful representationlegal representation. We then aggregated the representation codes into five matched categories for employer and claimant: not attended or represented, self-represented, internal representation (this included family and friends for claimants and senior person in the organisation for employer), external representation (lay rep, Citizens Advice Bureau or Trade Union for claimant and lay rep, consultant for employer) and finally legal representation for both claimant and employer. Each category was given a matched value from zero to four. Claimant representation was then subtracted from employer representation to give a score between − 4 and + 4. We further collapsed the calculated score into RRS categories of low, neutral and high. A negative score indicated that the claimant had stronger representation than the employerwe call this 'high RRS'. A positive score indicated that the claimant had weaker representation than the employer-we call this 'low RRS'. If claimant and employer had the same representational strength, the score was zero-we call this 'neutral RRS'.
In evaluating combination of claims made in addition to a whistleblowing claim we coded for type of claim using the ET jurisdiction codes. We then aggregated these codes into categories for analysis. The categories used were: (1) Claims related to discrimination: disability, sex, age, sexual orientation, race, maternity, religion or belief. There were no claims for marriage/civil partnership or transgender.
(2) Claims related to wages: failure to pay national minimum wage, unauthorised deductions/ failure to pay wages, redundancy pay, equal pay issue.
(3) Claims related to non-standard forms of employment: less favourable treatment for being part time, fixed term, flexible working, dismissal as temporary or agency worker. There were too few of these cases to include in the multinomial logistic regression analysis. (4) Finally claims related to procedures 1 : failure to be accompanied at grievance or disciplinary, failure to provide written reason for dismissal, failure to provide written pay statement, failure to consult before transfer, detriment for trade union membership, redundancy for acting on health and safety regulation, exercising statutory right, breach of contract, failure to provide written statement of terms and conditions, breach of Working Time Regulations, unfair dismissal.

Checking Multicollinearity
We test H1, H2 and H3 using multinomial logistic regression analysis in SPSS. This kind of analysis allows to test statistical significance of relations between nominal independent and dependent variables that have more than two nominal values. It is, however, necessary to ensure there is no multicollinearity when using multinomial logistic regression analysis. We first checked whether there were any correlations among the independent variables. The highest significant correlation was − .198 (tenure and sector), which is nowhere near the .8 which would suggest multicollinearity.
We then did linear regressions for the two dependent variables and the independent variables. Collinearity diagnostics indicate there is no collinearity, i.e. Tolerance is > .1; VIF < 10; Condition Index < 15. An overview is given in Tables 1 and 2. Other conditions for doing a multinomial logistic regression analysis were also met. All dependent and independent variables were coded nominally, with the exception of the RRS calculation, which can be argued is an ordinal variable. The data is a result of independent observations (i.e. the cases did not influence each other) and the dependent variables have mutually exclusive and exhaustive values. However, we did not have all variables for all of our cases. For assumptions 1-4 we indicate the respective n for each variable. For H1, H2 and H3 we were able to test a model for n = 505, but tenure was tested separately (because we had information on claimant's tenure in only 271 of the cases).

Findings
For assumptions 1-4, we ran simple frequency tables (cf. Tables 3, 4, 5 and 6). With regard to tenure, we took 5 years as a marker because for most of the statutory employment rights in Britain, a tenure of 2 years is the threshold. Whistleblowing legislation is an exception, and those rights can be claimed from the first day in employment. We find that almost 6 out of 10 whistleblower claimants had at least 2 years tenure (cf . Table 3). Hence, most of those who bring a whistleblowing claim at ET have their full employment rights. Our findings indicate that we cannot simply assume that whistleblowing legislation will attract everyone who seeks justice at ET when they have less than 2 years of tenure and hence do not qualify for other employment claims. The authors have often heard this assertion when discussing whistleblowing legislation. We find that a majority of those who seek justice at ET for a whistleblowing claim have more than 2 years of tenure and hence would qualify to make other claims as well.
With regard to gender, we find that 42% of whistleblowing claimants were women. The Office for National Statistics 2 indicates that of those in employment in the UK, 46.9% are women. A t-test suggests this is significant (p < 0.05). Of course, we do not know the gender distribution of all whistleblowing at the workplace, nor that of unsuccessful whistleblowing. Nevertheless, compared to the gender differentiation of the UK workforce, our findings suggest that women are not overrepresented at ET. Table 5 shows our sector finding. There are more claimants from the private sector (70.5%) than from the public sector (29.5%). However, the Office for National Statistics 3 indicates that in the UK, 21.5% work in the public sector and 78.5% in the private sector. Hence, we find that the public sector is significantly overrepresented in our sample (29.5% vs 21.5%, p < 0.05).
With regard to combined claims we find that 41% of whistleblowing claims are brought in combination with a discrimination claim (cf. Table 6). This is much more that what we found for combinations of whistleblowing claims with wage related claims (25.4%), procedural claims (11.9%) or claims relating to non-standard-forms-of-employment (1.2%).
To test the hypotheses H1, H2 and H3, we did multinomial logistic regression analysis for two dependent variables, outcome of the whistleblowing claim (H1, and H2) and RRS (H1, and H2). Results are summarised in the Tables 7 and 8 for the outcome of the whistleblowing claim as dependent  variable (H1, and H2), and Tables 9 and 10 for RRS as the dependent variable (H1, and H2).
There is a small gender effect, namely that male whistleblowers are more likely to see their case dismissed at preliminary, and less likely to settle in their case. There is also an effect of claiming discrimination alongside whistleblowing retaliation. Whistleblowing cases without a discrimination claim were more likely to be successful, and less likely to be dismissed at preliminary hearing. Hence, H1 is partially confirmed.
The biggest effects were for the RRS variable. Where the claimant had the same representational strength than the employer (neutral RRS), the case was less likely to be dismissed at preliminary hearing, more likely to be settled, withdrawn after preliminary hearing, or successful for the claimant. Where the claimant had stronger representation than the employer (high RRS), the case was more likely to be either settled or successful for the claimant. Hence H2 is rejected.
Representation has the most effect at ET in whistleblowing cases. However, bringing a whistleblowing claim to the ET in combination with a discrimination claim does not bode well, and being a man seems to put the whistleblower at a slight disadvantage.
Our findings show that public sector workers are less likely to have more power than their public sector employers. Tenure also has an effect on RRS. Workers with less than 2 years of tenure are less likely to have a neutral or high RRS.
In terms of the kind of whistleblowing cases, we found that whistleblowers without a discrimination claim were more likely to have a neutral or high RRS. And whistleblowers without a wage related claim were less likely to have a high RRS. Hence H3 is only partially upheld.

Discussion
This paper's main question is whether adjudicating institutions can even out power imbalances relevant to whistleblowing. Undeniably, the employment relationship is unequal, with the employer holding more power than the employee and in order to make employment rights effective this power imbalance needs to be addressed (Dickens, 2012). In Britain, ET are the adjudicating institution for whistleblowing legislation. Its 'Rules of Procedure' stipulate that the overriding objective of the ET is to 'deal with cases fairly and justly', which includes 'ensuring that the parties are on equal footing' (National Archives, 2013). We built our hypotheses on the assumption that those using the legislation to seek justice through ET, would be whistleblowers that did not have enough power to avoid detriment and see their   employer take their concern seriously. So, we expected the less powerful to be the whistleblowing claimants at ET, and we expected ET to be able to 'undo' that power imbalance.
Our findings suggest that the power imbalances that bring whistleblowers to ET are not related to tenure or gender, but rather to sector and perhaps other protected characteristics (for example race or disability). The results of our analysis suggest that ET can even out a power imbalance female whistleblowers might face. However, we do believe some caveats need to be mentioned here. We find a slight underrepresentation of women as whistleblowing claimants. We are unable to say why that is. Extant literature suggests that women are less likely to blow the whistle but that those who do are more likely to suffer detriment. Wider research (Merger et al., 2002) also note men are more aware of their rights and, therefore, more likely to make a claim than women. It is possible, therefore, that this particular equation works out to see less female whistleblowing claimants, or that women are less likely to pursue a whistleblowing claim at ET, reflecting the wider trend for claims overall where women bring fewer claims than men (BEIS, 2020a). We did not find a significant effect of gender on RRS, but other analysis (APPG, 2020) suggests that slightly more women than men are litigants in person. The implication is that what on the surface looks like an 'equal footing' might very well be different experiences. Our findings do suggest that male whistleblowing claimants are more likely to see their claim dismissed at preliminary hearing, and less likely to settle than they are to see their claim dismissed after preliminary. We did not see those differences for women. Other potentially gendered differences on which our analysis here remains inconclusive, relate to whether or not compensation is granted and how the monetary value compares to the cost of pursuing the claim. Finally, we measured the outcome of the whistleblowing claim as dependent variable. However, if the claim is not related to automatic unfair dismissal, then analysis elsewhere (APPG, 2020) suggests that female whistleblowers see 20% less ET rulings acknowledging unfair dismissal.
With regard to sector, our analysis does find the public sector whistleblower claimant is overrepresented. We know that in Britain trade unions are more powerful in the public sector (BEIS, 2020b). Although none of the whistleblower claimants had union representation at ET, they might have been advised by their union-data which is not captured in the ET judgment. Another possible explanation is that public sector employers are less likely to settle. Our findings on the relation between sector and RRS suggest that public sector whistleblowers rarely had high RRS. We might derive from this finding that public sector employers are more likely to take on the fight at ET, and seem to be well prepared.
What we want to turn the focus of the discussion to now is the RRS findings throughout the different analyses we undertook for this paper. Our findings suggest that RRS at ET does matter for whistleblower claimants. However, in developing our hypotheses we argued that the stronger case might find it easier to have better representation for the claimant. It is, therefore, useful to see what power imbalances are reflected in differences in representation, and which representation imbalances the ET are not able to even out.
Our findings allow us to suggest that tenure, sector and combination of claims have an effect on RRS of whistleblower claimants. However, our findings also suggest that ET are able to see through those imbalances in representation that stem from tenure and sector, but not those stemming from a combination of claims. More specifically, we find that claimants who made a whistleblowing claim in combination with a discrimination claim had low RRS. Since 41% of whistleblowing claims at ET are made in combination with a discrimination claim, the importance of this finding should not be underestimated. The ET seems to have difficulty in putting those claimants 'on an equal footing', as we found that when brought in combination with a discrimination claim, the whistleblowing claim is less likely to be successful and more likely to be withdrawn after preliminary hearing. We believe this finding points at a problem for the ET in relation to access to justice for cases involving complex legislation, and also a problem for the legal profession, not only in terms of how the legal profession is incentivized to pursue certain claims but also in terms of how it is subsidized to provide early legal advice, i.e. legal aid.
The purpose of ET, as set out by the Donovan Commission (1968) is to ensure speedy, non-adversarial, accessible, informal and cheap resolution to employment disputes. This goal coupled with the over-riding principle to ensure fair and just treatment to place parties on an equal footing (National Archives, 2013) should result in little difference between type of claims made either separately or in combination. Our findings, show, however, that the ability of the ET to meet these lofty goals depends on the type and combination of claims made, suggesting combination and complexity of legislation can override these principles. We find that claims where whistleblowing was made in combination with discrimination were adjudicated significantly differently than other claims. That is an important finding because these claims amount to 41% of all whistleblowing claims. A possible explanation is that combining whistleblowing claims with a wage related claim lowers the complexity of a whistleblowing case because the evidence is more likely to be tangible and, it is therefore, easier to separate the disclosure and detriment. In contrast, those claiming discrimination and whistleblowing are more likely to self-represent and, therefore, lack the legal expertise to separate out these factors. Furthermore, discrimination and whistleblowing legislation are complex (BEIS, 2020a; Lewis, 2013) which could further hinder the case. However, is it not close to a scandal that just because legislation is complex, unethical practices remain inadequately adjudicated?
In the context of an increasingly adversarial ET system (Corby, 2015), legal representation and high RRS remain the most powerful way to secure success at full hearing for a claimant. This situation poses a threat to claimants who find it difficult to secure legal representation due to the expense (BEIS, 2020a) and lack of knowledge about their legal rights (Meager et al., 2002). The Low Commission, in 2014, highlighted the urgent need for early legal advice and when this is not available there is a cost to the individual (distress) and to the system (an increase in those without representation). When this occurs, as it has done for whistleblowing (APPG, 2020), the ET system must adapt to deal with those without legal representation (Low Commission, 2014). However, our results show this adaptation has not occurred which threatens the ability of the ET to deliver on its overriding objective of 'ensuring that the parties are on equal footing' (National Archives, 2013). As a result, we call for legal aid to be provided to all claimants at ET at the earliest opportunity. An alternative could be to establish a specialist agency that can make limited adjudication-e.g. whether someone is disabled, whether a practice is discriminatory, whether someone has made a protected disclosure-and allow the ET to take over positions of that agency. In Britain, the mandate of the Equality and Human Rights Commission could be expanded so that it could have that expertise function for discrimination; and an All Party Parliamentary Group is currently advocating a specialist agency for whistleblowing.

Conclusion
The overriding objective of the British ET is that it hears cases 'fairly and justly' by putting parties 'on an equal foot'. This paper inquired whether ET in Britain even out or reproduce the power imbalances whistleblowers face as they call upon the law to seek justice. Our findings suggest that ET to an extent do seem to deliver on their objectives but appear to fail dramatically when it comes to an important subset of whistleblowers.
Although power imbalances relating to whistleblowers' tenure or employment sector have an effect on the representational strength of whistleblowers, they do not seem to have an effect on the outcome of the whistleblowing claim. We saw that a substantial part (41%) of whistleblowing claims are made in conjunction with a discrimination claim. Our findings suggest that for these cases, ET are not able to hear the parties 'on an equal foot' and thus jeopardizes one of the assumptions of rule-of-law.
The ET themselves are not the only ones to blame here. Our findings suggest that representational strength does matter for whistleblowers. It remains possible that cases with more complex merits find it harder to attract legal representation, but that does not seem a plausible explanation for our findings on whistleblowing and discrimination. Here, we saw both an effect on representational strength, as well as on the outcome of the whistleblowing claim. We argued that as both whistleblowing legislation as well as discrimination legislation are complex, more needs to be done to avoid a 'double whammy'.
Our findings suggest that access to justice remains an issue for whistleblowers who face discrimination. Providing legal aid is one option, changing legislation another, and creating an expert body to facilitate adjudication at ET might be a third. We note that an All Party Parliamentary Group in the UK-the APPG for Whistleblowing-is advocating the establishment of an Office of the Whistleblower, to that end.
Taken together, the evidence presented in this paper suggests that the ET is not delivering on the overriding objectives of the Donovan Commission, to make ET speedy, nonadversarial, accessible, informal and cheap. The evidence in this paper suggests they have become formal, as noted elsewhere (see Corby, 2015), adversarial and inaccessible, with the need for representation central to the success of a case.
Our paper also allows us to identify three areas for further research. First, the main finding of this paper points at ET cases that involve a combination of discrimination and whistleblowing claims. Qualitative research could contribute to deepening our insight presented here. More precisely, research would be helpful that examines access to legal aid, as well as investigating a more granular set of variables that are specific to these combined cases.
Second, there are indications that suggest a gendered experience for whistleblowing claimants at ET. However, our study used methods of analysis that are not designed to characterise these differences in experience. We nevertheless believe this is an important dimension of justice to which ET as an adjudicating institution needs to pay attention. Further ethnographic research would be useful.
Third, we suggested that improving the adjudication of whistleblowing can happen through interventions in legislation, in the provision of legal aid, or in the establishment of an independent agency for whistleblowing. Research into the hopes, aspirations and appetites for each of these reforms from the perspective of users, professionals and policy makers is needed to further delineate the changes that this paper calls for.
Funding British Academy, Leverhulme Trust.

Conflict of interest
The authors declare there is no conflict of interest.
Research Ethics For this paper, only publicly available data was used Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.