Act

Complaint/Dispute Reference No.

Date of Receipt

Complaint seeking adjudication by the Workplace Relations Commission under section 77 of the Employment Equality Act, 1998

CA-00005635-001

04/07/2016

Complaint seeking adjudication by the Workplace Relations Commission under section 77 of the Employment Equality Act, 1998

CA-00005692-001

06/07/2016

The complainant filed two complaints against the respondent, alleging discrimination on the grounds of gender and age, in terms of access to promotion, contrary to S. 8(1) of the Employment Equality Acts 1998-2015, on 4 July 2016. I held a joint hearing into these complaints on 2 November 2016 and 11 January 2017, with both parties present. The final piece of correspondence was received on 2 February 2017.

The complainant works in a relatively small expert section within a larger public sector body. He submits that from 2009 onwards, men and older workers have been systematically disadvantaged in terms of the respondent’s promotion policies. His extensive written and oral evidence can be summarised as follows:

The complainant, in all these competitions, was never empanelled to be possibly promoted. In one competition, his marks showed a clear dis-improvement, which according to the complainant is unlikely, as with his greater experience, he should have improved instead. He puts this down to a change in the interview panel.   He further states that in all these years, all women who competed for promotion were empanelled. With regard to the respondent’s most recent promotion competition in 2015, he stated that five of the eight participants were empanelled, and that all of these people were under 50 years of age.   They were three women and two men. The candidates not empanelled were the complainant, who is 60, a man due for retirement in 2026 and hence somewhere between 50 and 55 years of age, and an external candidate who, in the estimate of the complainant, may have possibly been older than the complainant, that is, over 60.   He pointed out that all of the three women who participated in this competition were first empanelled and subsequently promoted, and that the youngest of the participating men was also promoted.

The complainant argued that since staff come in through the same appointments process, they should have similar abilities, but that in his observation, the career trajectories of his colleagues do not bear this out.   In cross-examination, however, he accepted that he does not have any knowledge of the previous experience of the other candidates.  Nevertheless, he stated in evidence that at least one successful woman candidate had shown him work she was given to do, which would have been at the higher grade level and have prepared her very well for promotion. The complainant also alleged in more general terms that the women candidates were receiving personalised coaching from one member of the interview board.

The complainant also took issue with aspects of the competition, as well as the fact that seniority is no longer a criterion for promotion.   He complains that he was unable to do parts of the test on paper as opposed to online, although he accepts that he was offered extra time to complete the test. Similarly, he complains about aspects of the feedback process, but states himself that whilst written feedback was first refused to him – instead of oral feedback, which he says was offered to him, with the offer repeated several times – it was later provided.

In cross-examination, the complainant accepted that the promotion panel was still open and that the gender ratio of promotions could still be 60/40. The complainant also accepted that the grading of the written test was indeed anonymous and done by an outside assessor who only was only given candidate numbers from 1-8.

The complainant also made use of statistical evidence to support his case.

He combined all male and female candidates from the promotion competitions from 2009 to 2015 into a contingency table and performed the X²-test on it.   A brief summary of what the X²-test is and what it entails will be given the conclusions below. It is important to note, however, that one of the expected values, the meaning of which is also explained below, was just 3. The gist of the complainant’s statistical evidence is that the promotion pattern observed in his section between 2009 and 2015 has a likelihood of less than 1 in 1000.

With regard to the respondent’s objection that the complaint was brought outside the time limits set down in S. 77(5) of the Acts, the complainant explained that he had been on holidays when the results of the most recent competition came out on 4 January 2016, and that he learned of same only on 7 January 2016.   He also made the point that he first appealed the matter internally.

The respondent in its written submission raised the preliminary objection that the complaint was brought outside of the time limits specified in S. 77(5) of the Employment Equality Acts. It states that the results of the 2015 promotion competition were published on 4 January 2016, yet the complainant’s complaints were only received by the WRC on 4 and 6 July 2016, respectively. According to the respondent, the last day on which the complaints could be received by the WRC was 3 July 2016.

In terms of its defence of the substantive case, the respondent’s main points are that its entire recruitment and promotion systems had been extensively audited by the Commission on Public Service Appointments before the 2015 promotion competition and been found to be sound and transparent.

The respondent’s HR director stated in her evidence that the interview board had been extensively trained, including in equality matters and unconscious bias The respondent further stated that while until recently, the gender balance in the section in which the complainant served had been 52% male and 48% female, this had only changed to a 50/50 split, which, the respondent asserts, does not show a preference for women staff. The respondent further stated that age is not asked for in competitions.     An individual’s experience is subsumed under “competencies”; the respondent is particularly interested in the quality of a candidate’s experience.

It was the HR Director’s evidence that promotion on seniority and suitability is gone from the Irish public sector and that all promotions are now merit-based.   She also pointed out that one man over 50 would have been empanelled had he not failed the written test, which the complainant passed. The interview board would not have known at this point that the man had failed the written test.

The respondent also takes a very serious view on the allegations raised by the complainant in terms of the coaching of individual women candidates and maintains that the complainant has not adduced any concrete evidence in support of same.     The respondent pointed out that it is ultimately for an employer to decide how to develop individual staff members.

Specifically with regard to the statistical evidence adduced by the complainant, the respondent points out that one of the expected values in the complainant’s table is below 5. It states that it is generally accepted among statisticians that expected values of less than 5 render the results of the X² test unsafe. The respondent also argued, and opened case law in this respect, that statistical evidence on its own cannot be decisive in a complaint of discrimination and that such evidence cannot be used to “backward engineer”, in the words of their counsel, the outcome of a promotion competition.

Preliminary issue: Time Limits

With regard to the respondent’s objection that the complaint is out of time, I am satisfied that the complainant’s explanation that he learned of the outcome of the competition on his return from his holidays both explains and justifies the slight delay which occurred in terms of filing his complaint. It must be noted that the complainant is, at the most, three days over the six month time limit set down in the Acts.   Furthermore, the complainant is unrepresented, and the time of the publication of the results was right after Christmas, when many workers take longer holidays.   To restrict the complainant’s access to justice for this fairly small infringement of a statutory time limit, which did not prejudice the respondent in any way, would not be conscionable.   Accordingly, I am extending time pursuant to my powers set out in S. 77(5) and find that I have jurisdiction to investigate this matter.  However, only the respondent’s 2015 promotion competition is thus within time and can be investigated, as I stated to the parties at the hearing of the complainant and as was accepted by them.

The Substantive Case: Statistical Evidence

The complainant is entitled to adduce statistical evidence in support of his claim. He stated that the figures he based his calculations on came from taking notes on the outcomes of promotion competitions in his section since 2009. The respondent did not dispute the veracity of the complainant’s figures.

The X²-test (chi-square test), which the complainant employed in this context, is a hypothesis-testing statistical tool to examine the independence, or otherwise, of two variables – as here, gender and promotion.   In brief, if the two variables are independent, then people should succeed or fail on promotion roughly in proportion to their numbers in each category. For example, if staff in a section is 70% male, then 70% of men or thereabouts should succeed in getting promoted.   These numbers are called the expected values, and they are compared with the actual figures, which are called the observed values.   The sum of all squared differentials between observed and expected values, divided by the expected values, is called the X² test statistic. In very general terms, the higher this number, the less likely is the independence of the two variables.    

I am satisfied that the X²-test, which the complainant used, is indeed the correct hypothesis test for the independence of two variables, as in the case on hand, gender and promotion. Furthermore, I have no reason to believe that the complainant’s figures or computations are fundamentally wrong, although I do note that he gives his expected values as whole integers which they are normally not, given that they depict frequencies rather than counted objects.   There are, however, other significant problems with the statistics which the complainant represented.

The most serious of these is that the complainant represented his figures as an aggregate, taken from promotion competitions from 2009 to 2015. However, each promotion competition must be seen as a statistically independent event in terms of the chances of any of the participants to obtain a promotion. Two events are said to be statistically independent if the occurrence of one has no effect on the likelihood of the occurrence of the other.   I am satisfied that this is the case here.   One person being promoted in one competition has no effect on the chances of another person being promoted in the next competition, and each competition represents a fresh chance for all participants to again obtain promotion.   Furthermore, and in more everyday terms, it may be that an employer changes criteria from one competition to another. This may be in terms of changing competencies in response to changing business needs or, for example, because an employer had become aware that some test criteria used were disadvantaging certain groups and subsequently amended its approach: To analyse the results in aggregate would greatly and unfairly prejudice a respondent. For all of these reasons, the independence of the variables of gender and promotion will need to be analysed for each competition separately, both in terms of mathematics and common sense.     I therefore cannot accept the statistics adduced by the complainant as valid evidence.

For the sake of completeness I also want to state that I accept the respondent’s objection that a X²-test with expected values of less than 5 ought not to be accepted.   This is a generally accepted rule because the calculation of the critical values, which determine whether two variables are independent or not, is based on approximations which hold only when the expected values are reasonably large. Whilst I am aware that there are less restrictive rules which are sometimes applied, the complainant made no argument for any of these. However, nothing turns on this point in light of the analysis above as to why the complainant’s statistical evidence is not validly adduced.

The Substantive Case: Other Conclusions

The complainant’s evidence in support of his allegations, outside of statistics, is quite weak.   In terms of his complaints about the process, not only was this process the same for all candidates, but by his own evidence, he was facilitated by the respondent in various ways. One example would be the extra time offered to the complainant for the written test.     Therefore, these misgivings cannot ground a complaint of discrimination, which is less favourable treatment connected to a specific ground, after all.  

The complainant was not able to offer any supporting evidence for his allegations that a member of the interview board coached women staff on the process – over and above the interview training provided by the respondent to all candidates – or that the head of the section influenced the promotion process.

Even if I take his evidence that one woman colleague showed him work she was undertaking at a higher level at its height, I find it cannot establish a case for more favourable treatment.   First of all, employers do need to be granted some amount of discretion in how they develop their staff and second, it is just as easy to fail when entrusted with a higher responsibility as it is to succeed, so this is in itself not proof of more favourable treatment. Neither is there the slightest evidence that gender was the decisive factor in the allocation of tasks. It might have been just as well that the worker had an educational background or other experience which made her appear suitable for it in the eyes of her managers.  

I certainly do not accept the complainant’s contention that just because staffs are recruited through the same appointments process, they all have similar abilities, and hence should have similar career trajectories and receive similar development.   That is virtually never the case, whatever the employment in question, and has nothing to do with any protected characteristics.   In fact, one might say that the Employment Equality Acts exist so that people of ability are not held back in the workplace because of bias against them because of their protected characteristics.

The weakness of the complainant’s evidence, together with the respondent’s undisputed evidence that a man over 50 passed the interview and would have been empanelled but for his lack of success in the written exam – given that the complainant’s main grievances were with the alleged manipulations by members of the interview board – mean that no prima facie case has been established that the respondent public body discriminates against the expert staff in the section in which the complainant works on the grounds of gender or age in terms of promotions. The complaint must therefore fail.

I find that the public employer who is the respondent in this case did not discriminate against the complainant, on the ground of either gender or age, in terms of promotion contrary to S. 8(1) of the Employment Equality Acts 1998-2015.

Time limits - discrimination – access to promotion – gender –age – statistical evidence – probabilistic reasoning – statistically independent events – hypothesis testing – chi-square-test – expected values – no prima facie case.