OGGO Committee Meeting

     Good morning, everyone. Today's meeting, as we had scheduled, is on the name-blind recruitment pilot project.

    We have before us from the Public Service Commission Patrick Borbey, the President; Stan Lee, the Vice-President; and Charles Tardif. From the Treasury Board, we have Carl Trottier, Assistant Deputy Minister.

    Welcome to all.

    Mr. Borbey, do you have any opening remarks?

     Thank you, Madam Chair, for the opportunity to appear before the committee to share what we have learned as we have explored the use of anonymized applications and its impact on the screening and recruitment of visible minorities within the Canadian public service. As you said, I'm accompanied here by colleagues from the Public Service Commission, whom you introduced, but I also want to note that I'm pleased to have Carl Trottier here.

    The Public Service Commission and the office of the chief human resources officer work closely together on many files. Mr. Trottier has been involved in this pilot project since day one. I would also be remiss if I did not acknowledge the help and support of the many departments and agencies who participated in this pilot. It is through these types of partnerships that we can get a government-wide perspective on important staffing matters.

    The Public Service Employment Act states that:

    It goes on to recognize that:

    These are powerful words, and it is because of them that the Public Service Commission of Canada considers diversity and inclusion as fundamental components of its mandate. Therefore, we are always exploring innovative ways to improve recruitment methods. This pilot project is one of the many activities we have undertaken in our attempt at doing so. For example, over the years we've implemented a series of tools in support of barrier-free recruitment. Some of these include universal design for testing of candidates; employment equity targeted recruitment programs; machine-scored testing to eliminate subjectivity; training to avoid bias during selection processes;

    a recruitment system that offers automated screening and random selection of candidates; policies that provide the option to restrict recruitment to members of employment equity groups, to improve representation where gaps are identified; and an attestation form—which I've attached—that is signed by all managers with staffing delegation.

    At the Public Service Commission, we believe in the power of experimentation to drive evidence-based innovation. Our work on anonymized application is an example of this experimentation.

    You will note that I am stressing the “anonymizing” aspect of our work. While we originally referred to this as “name-blind recruitment”, because this was the term recognized internationally, it is our view that the idea of anonymizing better reflects the approach we have taken. In other words, it is the removal of all personal information that could lead to the identification of a candidate's origin, as opposed to merely the removal of an individual's name.

    I will take a moment to quickly recap the features and findings of the pilot.

    It included 27 positions advertised to the public from 17 departments and agencies. This resulted in a sample of over 2,200 candidates, of which 685, or 31%, self-identified as visible minorities.

    In keeping with research standards, our pilot project was reviewed by three external experts—two members of academia and an expert in methodology. In addition to complying with research standards, we considered peer review essential to ensure that the conclusions from our project were warranted and reasonable.

    Overall, the pilot found that there was no net benefit or disadvantage to using the anonymized screening method for visible minorities. It also showed that the anonymized method reduced the screen-in rate for all other applicants. Not surprisingly, the results showed a strong correlation between previous government experience and screen-in rates for all candidates.

    There are always limitations when it comes to research methodology. Being open about these is considered a professional obligation, and so you will notice that our final report identified several limitations related to this particular project.

    It should be noted that when we designed our methodology, we tried to specifically address some of the limitations reported in previous research papers on the subject, such as using fabricated resumes and fictitious staffing processes. We also recognized that our pilot would shed additional light on anonymized recruitment and that further research or work would be necessary.

    Our report was clear that the findings provided one additional source of evidence, but would not provide the complete answer on the applicability of anonymized recruitment in the federal public service. In other words, this pilot was never intended to be a silver bullet solution; rather, it was developed to contribute to, and complement, the existing body of evidence.

    With regard to next steps, the Public Service Commission is undertaking a formal audit to examine the success rate of employment equity groups at key stages of the recruitment process. This approach will provide us with additional evidence while addressing some of the limitations—namely, that managers were aware of their participation in the pilot project and that we were dealing with organizations who had volunteered to participate. The audit will also examine hiring practices to identify areas that may contain potential barriers or that may, for one reason or another, be more inclusive.

    The Public Service Commission will also share its methodology with departments and agencies who may decide to anonymize applications for their staffing processes.

    We will also explore how anonymizing principles could be included in the design of any future technology changes to our recruitment systems.

    In conclusion, Madam Chair, I want to assure this committee that the Public Service Commission is fully committed to diversity and inclusion, and will continue to take steps to keep these values at the forefront of public service recruitment and staffing. As an example, we will continue to provide targeted recruitment programs, such as the indigenous student employment opportunity, which offers indigenous full-time students work experience in the federal public service.

    We would be pleased to answer any questions you may have.

    Thank you.

    Thank you.

    Committee members, we will be going till a quarter to one, after which we will consider committee business.

    We'll start the first round with Monsieur Ayoub.

    You have seven minutes.

    Thank you, Madam Chair.

    Once again, gentlemen, thank you for being here.

    The pilot project is certainly compelling.

    You may have mentioned it, but why did the Public Service Commission undertake a pilot project with a view to improving practices? Do you know whether anyone else in Canada or the world has ever undertaken a pilot project of this magnitude? If so, how do those results compare with yours?

    Thank you for the question.

    Other countries and even some Canadian provinces have undertaken similar pilot projects and studies, but our pilot was the largest and used actual staffing processes. As I mentioned, some other projects relied on fictitious processes and resumes. Consequently, they did not test real staffing processes used by real managers making staffing decisions about real applicants and cases.

    I believe our pilot was slightly more comprehensive than anything that had been done previously. We did look at pilot projects and studies undertaken in Australia and France, and that research helped us in developing our methodology. The answer is yes, we did take into account other projects that had been done abroad. In some cases, the experts who reviewed our pilot had conducted studies, but those were much more limited in scope.

    I'm going to stay on the same topic.

    Certain parts of the public sector no doubt have an interest in increasing the number of women and Indigenous people on staff, and even men, in areas where they are under-represented. The goal is to achieve a better balance by ensuring that every member of the population has equal access to employment opportunities.

    If I understand correctly, your preliminary finding was that the method used in the pilot offered no real benefit and made little, if any, difference. Any impact was quite minimal. Did this method of using real, but anonymized, applications lead to any sector-specific findings? For instance, did you find that the method would have made it possible to hire more women in a given sector? Conversely, were you unable to arrive at those kinds of conclusions?

    The methodology has its limitations. Our main focus was on visible minorities. We didn't take gender equity into account.

    In Canada, the process is more complicated, because it's actually very hard to anonymize or defeminize a resume or job application in French. It's easier in English, but it's very hard in French.

    Could you kindly explain why it's harder to do in French? Is it due to gender?

    Yes, that's exactly it. One of the problems is the constant use of “il” and “elle”—or “he” and “she” in English. That's why we didn't take into account gender equity. Instead, we really focused on visible minorities.

    It's not impossible. Unless I'm mistaken, it's a bit harder, but it's doable.

    If it's all right with you, I will ask our methodology expert to comment on that.

    Technically, it's impossible. In French, the phrase “Je suis diplômé”—“I am a graduate”—would take an additional “e” at the end in the case of a female graduate. The removal of that “e” automatically indicates the information being excluded. Even if the resume were to be anonymized, the text would reveal the gender with which the person identified.

    I see. I was under the impression that the idea behind the pilot was to remove certain pieces of information in order to level the playing field somewhat. Therefore, the document could undergo an initial pass, resulting in a draft to which changes could be made, either technologically or manually, so that the resume did not identify the applicant's gender.

     Either there is objectivity, or there isn't. The idea is to ensure that the process is as objective as possible. When all is said and done, I'd like to know whether this tool would make it possible to achieve near-total objectivity in specific areas.

    The methodology can be applied in certain circumstances. We are ready to share the methodology with departments that have gaps in representation. As you know, employment equity groups are generally overrepresented across the public service. Some departments or agencies, however, may need to improve representation in their workforce. Certainly, the methodology could be used in specific cases where increased representation is sought, in a given department or job category.

    I'm going to touch on another topic now. I realize this isn't necessarily the focus of the study, but it has to do with human resources.

    I'd like to know whether achieving a balance between francophone and anglophone employees is a problem in the public service. Does representation pose a challenge on that level?

    I'm going to make an assumption and say that unilingual English speakers have access to positions that unilingual French speakers would never have access to because they aren't functionally bilingual.

    Am I mistaken?

    The figures actually show a slight overrepresentation of francophones in the public service. That means there isn't really any issue with representation in that regard.

    Managers are the ones who decide whether to designate a position as bilingual, unilingual English, or unilingual French. They determine what their requirements are.

    From our standpoint, there doesn't seem to be a problem. That said, Mr. Trottier may have something to add.

    Time is up. You could probably answer that when I give it to the next guys.

    Mr. McCauley.

    Gentlemen, welcome back.

    Mr. Lee, I'm sure you've heard this before: I was so excited to think that, finally, the government had delivered for us. It's the wrong Stan Lee, but at least I can tell my kids I met Stan Lee.

    Mr. Borbey, what is the origin of this report? Was it felt that there was need for direction from someone? How did the report origin come about, please?

    I have to admit that it predates my arrival at the Public Service Commission. This was a decision that was made about a year ago to take a look at whether name-blinding could be applied in our jurisdiction. It was based on studies that have been conducted in other countries.

    Carl, I know you were part of the process from the start. Do you want to add anything?

     Could you explain how it came about, please? Was it felt that there was a need in terms of an inequality with certain groups, or...?

     We are always striving to be able to create the best means to recruit individuals. We are always looking to—-

    How did the report come about, then? If it's pre-Mr. Borbey but during your tenure, was it a directive from someone? Or was there a belief—-

    The report came out during Mr. Borbey's tenure. It began before his tenure and came out during his tenure. What I was getting to was that if there is a doubt that there might be unconscious bias going on, I think it is our responsibility—-

    Was there a doubt expressed?

    We didn't know. We had to do the study to be able to figure that out.

    The reason I ask is this. I think it's your third time with us, Mr. Borbey. Each time I've noted that the previous government and the current government both have done a very, very good job. It's a success story of hiring in the public service. I actually went to the Library of Canada probably six months ago. We show that their employment equity designated groups—women, aboriginal people, people with disabilities, members of visible minority groups—are all overrepresented compared with workforce availability. Again, it's a success story of the previous government and the current government.

    The library says, you know, here's the proof, so I'm curious to know what led us to say, well, maybe it's not—even though, again, the previous government and the current government have done a very good job in this matter.

    Yes. Again, I think we have to go beyond some of those statistics and take a look at where there are some gaps and some differences. For example, I'm concerned when I look at our recruitment campaigns that we're getting fewer indigenous people applying to our programs and fewer persons with disability. We need to understand whether there are some barriers, perceived or real barriers, preventing us from accessing the best talent pools in the country.

    Even with applications dropping, your departments are doing a very good job of hiring, because they are increasing in these items you mentioned. Are you able to provide us with some of the information that shows there's a reduction in the applications?

    Yes.

    The next part is that, again, I want to state, as I have stated to you before, that the current government and the previous government have done an excellent job. It's a success story, and it should be heard, as we only hear the bad stuff. It has been a success story in the hiring. But I do note that, just on an off chance just a couple of months ago, I filed an Order Paper about the cost of the project, and it was almost $200,000.

    That's what I'm getting at. Why would we spend $200,000 of taxpayers' money when the proof is clear that this has been a success story and current and previous governments have done a great job and it's improving and on the right track? I'm just curious; did we spend the $200,000 for a real purpose without doing the research, or was it a real need? Has someone brought forth some real shortfalls that somehow aren't reflected in the statistics, or is this going on in the government and it has to be corrected?

    Again, this is one initiative among a number that we have undertaken—I listed them earlier—to ensure that our recruitment system is bias-free. I have to be able to give that assurance to Parliament that our systems are bias-free. The way we can do that is through experimentation and testing to make sure there are no biases.

    To me, this is one initiative in the context of that overall effort. As I said, we're going to continue to do some work in this area using audit methodology, because this did not prove or disprove whether there are barriers or discrimination within our hiring practices in the public service.

    I'm thinking that results and success would prove that they're doing the right job, and maybe we should focus on the different parts of encouraging the applications.

    When you say you're going to be doing some more audits, would that lead federally to another report, similar to the current one?

    This will be an audit that we will conduct at the Public Service Commission, so it's not the same as a research study. The audit will take a look at samples of staffing actions across the public service.

    When will that be published?

    We're designing the methodology right now; this process will probably take some time.

    It will probably be at some point next year, Stan?

    Mr. Stan Lee: Yes.

    Okay. So it won't be that long from now; within 12 months?

     It will be a little more than 12 months, we estimate.

    Okay.

     When you were doing the pilot project, some applicants were culled out, as you mentioned. Were they contacted, the ones who were disapproved because...or didn't make it through the process of the pilot project? Will they be contacted and invited to reapply?

    No one was disadvantaged in the pilot project—

    I didn't say disadvantaged, but you did mention that there were...or it did not prove there was any discrimination against some, but you said there was a lower rate for some. Were those from the lower rate approached and offered to—

    Yes. We took precautions to ensure that we would have a fair approach in this process. One of the things we did, where there was disagreement or somebody was being screened out, was ask hiring managers to sit down and look at the application again together and determine if the person—

    But if someone was screened out, would they—

    Thank you. Your time is up.

    Mr. Blaikie, you have seven minutes.

    Thank you very much.

    The thing that stood out to me in your report was that, first of all, there was no net benefit or disadvantage to using the anonymized screening method, and then there were reduced screen-in rates for all other candidates. Can you help me understand how it is that the screen-in rate for other candidates would be affected simply by having an anonymous application?

    We don't really have an answer for that. We have to admit that it's a very intriguing result. We are discussing with some academics, for example, why that would be. At this point, we have no real defined reason why that would be. Any explanation would be speculative at this point.

    Is that claim based on the fact that there was a lower rate of screening-in than usual or is it based on an assessment after the fact of applications that didn't screen in that were then assessed in a non-anonymized fashion and then it was determined that, actually, we would normally screen those people in?

    That's a very good question. In fact, what you see in the report are the initial decisions. What happens to the candidate after is a different matter, because hiring managers, where they disagree, will sit together and look at it factually and see whether or not the person will go ahead.

    So the claim here that there was a lower screen-in rate is based on an average percentage of candidates who would normally screen in, and it was lower than usual in this case?

    No, it's compared with the traditional method. If you look at the table, you'll see that there are two methods to the left, which is the traditional method, which is all the information available, and the second row, which is the anonymized method. It's a strict comparison of the two methods.

    Okay. So the same candidates were run through the parallel process and there was a lower screen-in rate. That is definitely an interesting outcome.

    Minister Brison, either at this committee or another committee, had talked about the name-blind hiring process. He had pointed to information from other jurisdictions where name-blind hiring had a salutary effect on the hiring of candidates from visible minority groups. Do you guys have any sense as to why this pilot program didn't produce comparable results? Do you think that this is a strong counter-example to other research out there or that this doesn't really speak to the authority of research done in other jurisdictions?

    That's a terrific question. In fact, when we looked at the research, we saw a mixed bag of results. Sometimes you see a beneficial effect. In other instances you'll actually see an adverse effect. There is a paucity of research as well. We looked at maybe 10 to 12 serious studies on the issue. That's not many.

    What we can say is that so far, researchers are pointing to two things. One is that whether you see positive or negative effects will depend largely on the organizational context. That is, is there systemic discrimination currently in the system? If there is systemic discrimination, in all studies you tend to see a positive effect with anonymized recruitment.

    The other one is whether or not you have positive policies in place. If you have positive policies in place and you hide information as to the employment equity, it prevents the application of these positive policies. That's when you'll see a decrease of the effect.

     I think that speaks to my next question. On the one hand, some of the arguments for a name-blind or anonymous hiring process come from the point of view of wanting to reduce discrimination in the hiring process. There's also a positive argument. If we say we want to foster a more meritocratic hiring environment, then we want to remove things that are not based on performance and see people go through.

    I'm just wondering if there isn't a positive argument for anonymous hiring based on trying to leave out factors that aren't related to performance, and then how that interacts with any programs or principles that would have to do with purposely trying to hire equity-seeking candidates. Do you have policies in place? Have you looked at how you weigh those different values against each other in the hiring process?

    The Employment Equity Act provides a certain number of tools that managers can apply to the hiring process when they're trying to, as you say, effect positive change. There are some departments, and some parts of departments, that are struggling with representation, including of women. For example, one of the areas we're concerned about is women in technology. Through our post-secondary recruitment campaign this year, we gave preference to first consideration of candidates who self-declared as women when we were referring for new positions, because we know we have a significant shortage there: 25% of our computer scientists in the government are women. To try to bridge that gap, we can use the Employment Equity Act to bridge some of those measures.

    Again, on the other hand, there are some organizations that probably could benefit from using name-blinding, because at the end of the day, when they look at their representation, there's clearly a problem; there's an issue. Even when we looked at some countries where name-blinding seemed to be positive in terms of its impact, these were countries where there was a significant amount of systemic discrimination against the group that was being targeted.

    In terms of where it didn't work, as Stan was saying.... For example, in Australia, when they were looking at the impact on the hiring of women, there was the opposite effect. In fact, managers were actually predisposed to giving consideration to women in the screening decision, and anonymizing did actually have a detrimental impact on that.

    These are all the factors that we have to look at, and I think case by case we have to determine whether using this technique is the right approach to address whatever circumstances may exist in a particular organization.

    Thank you. Your time is up.

    We'll have Mr. Peterson for seven minutes.

    Thank you, Madam Chair.

    Thank you, everyone, for being with us once again.

    I want to take a higher look at this. Before the name-blind process, the department was clearly committed to the goals of diversity and inclusion. A number of tools were already in place, including universal design and employment equity programs. I want to focus on the training to avoid bias during selection processes.

    So there's already training built into the system. Can someone briefly explain to me what the nature of that training is?

    The training around diversity and inclusion is training for hiring managers to find a way to put aside their biases. Part of that includes concrete things like ensuring that you have what we call an “inclusive selection” committee. By having visible minorities sit on the selection committee, for example, you'll actually work through the various selection criteria together, collectively. That's one example.

    Another example is allowing for what we'll call cultural sensitivity. Some hiring managers may not be aware that looking someone in the eye may be considered offensive in some cultures. We try to get hiring managers to make abstraction of that and to focus on the answers that are provided by the candidate.

    These are examples.

    Looking at the results of the study and what existed before, can the inference be drawn that the tools already in place were adequate to achieve the goal that a name-blind process also achieves? Is that a correct inference, or is there not enough data to make that inference?

     I don't think we can conclude that the tools we have, the approaches we have, are sufficient. We're constantly testing new approaches, seeing what works and what doesn't work.

    We've also been in consultation with the employee groups that represent visible minorities, indigenous people, and persons with disabilities, and they have a number of recommendations that they've directed at us in terms of how our systems or processes can be further improved. For example, we listened to some pretty substantial feedback on how accommodation requirements were addressed at the recruitment stage and we were able to see that our processes were actually acting as barriers. Last year we did a “lean” exercise where we reviewed our accommodation approach, and we reduced about half of the steps. As a result, we were able to assess candidates for post-secondary recruitment who declared that they needed an accommodation, at the same time as we assessed other candidates.

    There's an example of continuous improvement. We're not claiming that we're perfect.

    No, no; it's just another tool.

    We know you report to Parliament. If Parliament asks you whether PSC's hiring practices are bias-free, are you able to give an answer to that question?

    Under the Public Service Employment Act, absolutely that is what we have to pursue—bias-free selection processes according to merit, non-partisan principles, and all of the principles associated with the act. That is what our mandate is. Are we there? Do we do it on a consistent basis across the system, which is talking about 200,000 employees across more than 80 organizations? No, we're not perfect. That's why we want to continue to improve. That is also the reason we said we're going to do an audit now to see what else we can find out about what's going on and how these various groups are affected through the recruitment process.

    On that audit framework that I think you said is being designed right now, we're not in a position to know exactly when—

    Yes. It will take about a year after we complete the design. Probably in late spring of next year we'll be able to report on the findings.

    Thank you for that.

    Were there any shortcomings or limits to this test itself? The participants knew they were participating. They were voluntarily participating. Could this skew the results? Is there any way to test that? If there is, I assume it has been tested.

    We acknowledged the limitations. Our methodology was reviewed by three outside experts, and the results were also reviewed. They confirmed that these are the limitations. However, in our study, we also addressed some of the limitations of other projects previously. I mentioned, for example, using real processes, real candidates, real applications, rather than fictitious ones, in some cases, where it was just a paper exercise. We've gone beyond some of the limitations, but there are still some limitations in terms of our methodology.

    Okay. You were initially calling it the name-blind pilot project, and then you changed it for the obvious reasons. Did the scope of the pilot project change when you got under way? Were all these features, such as name, address, languages spoken, and all that, already meant to be excluded, or was it a process as you were working through it?

    In some of the previous research projects or experiments that were done elsewhere, only a certain number of fields of information were removed. We felt that it still allowed the assessor, the hiring manager, to guess or find out the origin of the individual. That's why we went a bit further. We looked at all the different factors and we decided to anonymize all of the information. In the example of the CV and application at the back of the research study, you can see how much of the information is blacked out as a result.

    Right.

    How is my time, Madam Chair?

    You have one minute.

    Perfect.

    Some of the characteristics, if we can call them personal characteristics, are perhaps key competencies of the role. How do you factor for that when you're doing these?

    That's a very good question, and we struggled a little with this. For example, if somebody went to Harvard University, one might say, well, that's probably a feature that we would like to encourage in terms of staffing. However, because you went to Harvard University does not mean that you are a white person. You could be a visible minority. You could be from another country and you just happened to attend the university. At the end of the day, we had to take out that information in order to be fair, because again, if somebody has a diploma from a university in another country, where typically they might guess that this is a person who is a visible minority, obviously that becomes a factor in the decision, so we had to apply that element universally.

    In the future, when we have technology that allows electronic screening, the fact that you have a master's degree in biology should be all that counts.

     Thank you.

    We will now go to the five-minute round.

    Mr. McCauley will be sharing his time with Mr. Blaikie.

    Thank you.

    I have a question, Mr. Borbey. When you were answering either Mr. Peterson or Mr. Blaikie, you talked about problems or circumstances that might exist. How are you determining circumstances that might exist when you don't know they exist?

    I appreciate what you're trying to do here, but the results are very clear that the government is doing a very good job. I'll give you compliments, but it almost looks like this was $186,000 of taxpayers' money just to prove the government is doing a good job. We've discussed it already. It almost looks like some of this project is searching for a problem that doesn't exist. It's almost like you're trying to prove something exists that may not.

    I'm wondering if the focus should instead be more on known hiring problems we might have as opposed to looking for circumstances that might exist.

    I'm not sure, Madam Chair, what the question is. I mean, I'm referring to the fact that—

    Well, it looks like a solution looking for a problem. Again, the government has done a great job, and I compliment you. The current government and the previous administration did a very good job of hiring for the target equity groups. We've heard that the project shows that there is no real issue, so it looks like you are continuing to try to find a problem that may not exist. I just wonder if that is the best use of your resources or your best focus.

    When I talked about circumstances, I think each department, each deputy head, has a responsibility to examine the hiring practices within their organization and to measure representation. Even though we have good representation numbers across the public service, that does not mean that in some organizations there may not be under-representation—

    But can't I just go to the library and ask them for a breakdown? They gave me a breakdown for overall. They'll give me a breakdown department by department. We can figure out who is not doing that and then target it that way instead of —

    All I was saying was that the methodology could be applicable for some of those departments that may want to increase representation or that maybe suspect that there is unconscious bias in their staffing decisions. We're just prepared to share the methodology. We're not going to dictate to people to use it or not.

    Should we dictate to them, then, if we think the name-blind project...?

    We have a very delegated, decentralized staffing regime that we've put in place over a number of years. The commission has an oversight role, but the commission does not dictate to departments.

    Okay. That's a fair answer.

    We do use, for example, audit methodology to be able to determine whether departments should be doing a better job or maybe whether we should be a little bit more engaged on certain issues.

    Have you identified the departments that need to do a better job?

    That was not the purpose of this work.

    Outside of this work, have you identified those that need to do a better job?

    I think we are always examining the data on representation, having discussions with deputy heads where there are clearly some gaps, and encouraging them to take proactive measures to be able to bridge some of those gaps.

    Okay.

    Mr. Blaikie, go ahead.

    You have one and a half minutes, Mr. Blaikie.

    I just want to ask a question about the recruitment side. Obviously, you can have really good hiring practices. That's only going to operate on the candidate pool. I'm just wondering about the extent to which trying to encourage a more diverse initial candidate pool upon which to implement this hiring process is part of the scope of these efforts to have better hiring practices and whether your study included an analysis of the initial pool as opposed to just how the hiring process worked on the candidates who happen to apply.

    Every year we conduct a series of outreach activities targeted at specific groups. This is part of the whole idea of ensuring that people are made aware and of generating interest in applying from within various employment groups. We also have these career streams and various inventories that are dedicated to employment equity groups.

    The audit will also be looking at what you referred to, which is what happens from the original candidate pool to the pre-selection, but also at the key steps in the hiring process that go beyond the pre-selection.

     Thank you.

    Ms. Mendès.

    Thank you, Madam Chair.

    Thank you all for being here today.

    Before I go into French, I have a question that has been bugging us here. What on earth does “screen-in rates” mean? I couldn't find a French translation.

    In the selection process, the very first step after a candidate's application is received is for managers, supported by HR professionals, to determine, out of, say, 500 applicants, which ones meet the essential qualifications for experience. After that, they can go to the next phase. That's a first screening. It does not involve using testing methodology, a collective or individual test, or an interview , which is normally later on in the process. And of course, there are always references at the end as well. Those are the various steps.

    So it's that very, very first point where you determine that, out of the 500 applications you've received, the following 200 meet the experience qualifications and the other 300 or whatever do not. That's a decision made by staffing managers.

    So that is screening.

    Thank you very much. That helps clarify things for us.

    Mr. Lee, I'd like to make a quick comment about the use of gender in French. In the example you gave, the phrase “Je suis diplômé” would take an extra “e” at the end in the case of a female applicant. That is true, but you could capture the same meaning with the phrase “Je détiens un diplôme de”, which would be equivalent to the phrase “I have a degree in” in English. Ways of getting around gender-specific structures in French do exist. Given how rich the language is, it is possible to get around the rules.

    I completely understand what you were trying to achieve with your study. I also I understand that you took it upon yourselves to determine whether your methods were valid and effective. However, I can, to some extent, appreciate what Mr. McCauley was getting at with his questions. In his estimation of the government's approach, the public service is very proactive when it comes to recruiting a diverse workforce. Let's call it diversity so as not to include one group and not another.

    I think it's valid to call into question some of the factors you ruled out of the pilot project, factors such as knowledge of foreign languages. I say that because knowing a foreign language is useful for a number of jobs in the public service. I mean beyond English and French, our official languages. I'm referring to the ability to speak a foreign language. Obviously, it's important for people who work at Global Affairs Canada, say, to be able to speak a foreign language.

    Why did you remove that information from job applications?

    Again, the point was to ensure our methodology was as fair as possible. However, the information was not removed in cases where knowledge of a foreign language was clearly listed as an asset for the job.

    All right, I see. The information was not removed across the board. It depended on the position in question.

    Exactly. That's how it worked, as the president explained. When information was listed in the statement of merit criteria, it was not redacted.

    Very well. Information matching the job description was therefore retained during the process.

    Absolutely.

    Take, for example, a technical position requiring expertise in Great Lakes water analysis. When an applicant had experience doing that in Lake Erie, for instance, we did not redact the name Lake Erie. Similarly, that information would be revealed in the case of an applicant who had done that type of work in, say, Lake Titicaca, even though we know it is in Peru and refers to a geographical location. The information would all depend on the job criteria.

    I see. It was just to gain a clearer understanding.

    Now, I'd like to pick up on one of your findings. You indicated that the pilot may have shown a reduction in screen-in rates, and I'd like to know why.

    We don't really have an answer to that question. It is nevertheless a good one. What other studies have suggested—and not shown, in my view—is that, redacting information from job applications makes it more difficult to assess the extent or breadth of the experience in question, qualitatively speaking.

    That is all the more important in the public service, given that managers have to be sure the applicant meets the merit criteria in order for that person to be screened in to the process. That's harder to do when a significant amount of information is missing.

     That ties in with the point I was making earlier.

    Thank you.

    Mr. Kelly, you have five minutes.

    I don't know that I'm going to need five minutes; I have just a couple of thoughts. I know that perhaps our colleagues on the other side might have a bit of a busy day today, with some travel and a convention ahead of them. In the interest of time I'll keep it short.

    I agree with Mr. McCauley that upon reviewing the data we have, this really would seem to be a success story in the overall achievement in employment by the various categories in this study. This is an achievement Canadians can be proud of in terms of the governments of different parties over time achieving the levels of participation of women, aboriginal peoples, disabled persons, and visible minorities in the public service. The study is perhaps a worthy undertaking to examine the question of whether or not the name creates bias, but the road to an $18-billion deficit is paved with one good idea at a time that piles up into these kinds of large expenses.

    I agree that we seem to have achieved quite a bit of success as a country in this area, and we have to be conscious of never allowing that to lapse or to see reversals in this. I think this is something we can be proud of, and I'd like us as a committee to move on to other subject matter.

    I just want to clarify that we're dealing with both today's public service and the public service of the future. We are going through a pretty important renewal. There will be a lot of departures over the next few years. This is an opportunity for us to make sure that we will be attracting the best talents from across the country—again, in terms of the experience and skills and also the representation across the country, including geographical representation. So for us, this is an important investment into the kind of future public service that we can create.

    I have shared with the committee that I am concerned that in some categories we're not attracting enough candidates. That's of concern to me, because today's successful representation may change overnight, very quickly, as we bring in a lot of new public servants over the next few years. That's why I think it's important to study these issues and make sure that our processes are bias-free, that they work, are transparent, and serve all Canadians.

    Thank you.

    Mr. Drouin, you may go ahead for five minutes.

    Thank you, Madam Chair.

    Thanks to the committee for doing this study. I think you should be doing more of that. It's the right thing to do. I mean, any criteria to measure is I think good news for us, as parliamentarians, in order to understand. Unlike my colleague Mr. McCauley, I wouldn't necessarily wait until the Edmonton Oilers were out of the playoffs before measuring their success.

    Voices: Oh, oh!

    Mr. Francis Drouin: You want to be proactive.

    In short, I just want to thank the committee.

    I don't know whether I will use the full five minutes, but I do want to ask you something about your study.

    People in my riding, especially visible minorities, often have the same complaint: they are able to get a job in the public service, but, once they are in, they aren't able to get promoted.

    Did your study take into account internal promotions?

    No, it looked only at external staffing processes, in other words, to gain entry to the public service.

     We, too, receive that kind of feedback, so I understand what you're saying. Members of certain employment equity groups feel that it's harder for them to advance their careers than it is for other people. We need to examine the measures in place to make sure people have access to training and development programs. We also have to make sure that internal staffing processes, such as those for management positions, are completely unbiased and equally accessible to everyone. That's something we need to keep an eye on.

    Unfortunately, I missed part of the meeting. What comes next for this pilot project? What are you going to do in the months and years ahead?

    We will be doing three things.

     First of all, as I mentioned, we are ready to share our methodology with departments looking to address areas where certain groups are thought to be under-represented. We are ready and willing, then, to share that methodology with them.

    Second, we are going to examine the technological results. If the screening is done electronically, for instance, it would eliminate human intervention and thus all bias.

    Third, and most importantly, we are going to conduct a full audit in the next year to measure the success rates of employment equity groups throughout staffing processes.

    Very good.

    I want to be sure I'm clear on the technological aspect. Are you saying that the applicant selection process would be entirely electronic, even the interview?

    No.

    It would just be to screen applicants in or out of the process, then.

    Yes. For instance, we talked about female versus male representation. If we had a technological tool, it would be no problem. In other words, the software would not identify the applicant's gender, since it would not detect French occurrences of “ée” indicating that the applicant was a woman. Those possibilities don't currently exist, but we are including them as we define our next recruitment platform. That's one solution that could be applied.

    Exam-based solutions could also be put in place. After the screening stage, the 200 or 300 screened-in applicants could be asked to take a test that would be administered electronically, in other words, over the Internet. That would help to further screen out applicants and ensure that only qualified candidates were invited to an interview by the manager.

    That's great. Thank you very much.

    I completely agree with you. We need to examine the processes currently in place in order to attract members of my generation and train them to become managers, assistant deputy ministers, directors, and so forth.

     We need people who can replace me as well.

    Indeed. They have to start somewhere.

    Thank you very much for your work.

    Mr. Blaikie, you have three minutes.

    You have no questions? Okay.

    Mr. Jowhari, you indicated that you had a question.

    Yes.

    You have seven minutes.

    Okay.

     Thank you to the witnesses.

    I'm referring to the final report that's been tabled, and I'm looking specifically at the statistical significance of the different tables or the different categories that the study dealt with. I notice that there were 27 positions and 17 departments. I managed to find the name of the departments. When it came to the positions, I was successful in finding the occupational categories, and then I found classifications under each one of these.

    For example, under administrative foreign services, I found AS1 and AS2. Can you give me an idea of what type of positions these are in general? Unless I go in and do a search on AS1 and try to get an understanding of what the scope was.... That will help me, because there are some significant statistical differences that I'd like to be able to probe, but before I do that, I want to understand the nature of the positions.

     If you want, I can go through all of the categories that are presented.

    Just in general, are they entry level or management level? Are they analytical level? I'm trying to get an understanding of the level of positions and why they were specifically chosen.

    The idea of the process was to get a good representation of the jobs currently available in the government. We wanted to have, as you mentioned, entry-level positions but also more senior positions and different types of categories.

    We have a number of administrative assistants. When you refer to AS, those are administrative assistant positions, but we also have biologists, engineers, and economists. We also have people working in general trade. The idea was to represent all these categories of employment, so that's why you see the variety here in the tables that you referred to on page 13.

    You had picked five occupational categories. Can you tell me, overall, how many occupational categories exist? It's five out of how many: five out of 100, five out of 10...?

     It's five out of six. The only one not included was for the executives.

    Perfect. So from an occupational category, we are covering almost 90%. From the 27 positions within these categories, what percentage are we covering?

    That's a good question. I cannot give you a—

    That's fine. Could you perhaps find that information and send it to us?

    Yes.

    In general, one could say that from an occupational category we are covering a lot. We are covering the comprehensive, and on classification it looks like we're covering different levels.

    Coming specifically to table 6, when you look at the traditional versus the name-blind recruitment, we've gone from 46.3% under operational, for example, to 36% under name-blind. As I look at it, there seems to be a large statistical difference. The other ones are averaging probably about 3% to 5%. Why is there such a significant difference? Compare it with table 8, which is specifically talking about visible minorities. There's almost no statistical difference. I'm trying to reconcile this. Why is there such a statistical difference in one category? When it comes to visible minority, I don't have any statistical variance.

    It's all about the sample size itself. Usually the larger the sample size and the more candidates, the easier it will be to predict the right information. If you have a very small sample size, a large difference can't be determined or can't be identified as easily.

    How would you categorize your sample size?

    For that operational category, if you refer to page 13, we have only one process that ended up in that category. We are much lower here compared with all the others. As you can see, we have only 87 candidates.

    Okay.

    That's much smaller, for example, than for administration and foreign service, where we have hundreds of candidates. That explains the difference.

    Thank you.

    In your final report, just before your conclusion, you talk about some of the challenges, let's say, associated with the pilot project. You list about five items. One of them is that the process is very cumbersome.

    You talked about introducing some of the automated processes, and I know there's a plan to do another study in the future. How can the lessons learned from this study help you streamline in the future some of the challenges you've highlighted here so that we're not spending another $250,000, or so that the $250,000 we're going to spend will give us a lot more in benefits?

    Yes. We're taking what we learned, including the fact that it's more labour-intensive than we had anticipated not just to anonymize the information but also to ensure that it's done consistently across—imagine—2,200 applications. It's pretty significant. It also added a certain amount of time to the process. We're all concerned about ensuring that our recruitment processes are as efficient as possible. One of the reasons we'd like to look at some technological solutions is to reduce some of that burden.

    How would you, at the end, summarize the result? I know that you did a conclusion and made comments in your presentation. Do we have an issue—yes or no?

    Do we have an issue? The study was not meant to determine whether there is an issue related to discrimination in the public service. The study was to determine whether the use of a technology or a tool, an approach called name-blind or anonymizing, has an impact on the screening rates. We looked only at screening.

    Does it have an impact on the screening rates?

    Well, we concluded that it doesn't, at the end of the day. It doesn't have a beneficial or a detrimental impact, particularly if you're worried about barriers, or conscious or unconscious bias, related to visible minorities.

    I find it a little bit hard, because I'm looking at the statistical difference in the other categories. When it comes to visible, we don't have any statistical variance. That's a challenge for me, but we can talk about that another time.

    Thank you, Madam Chair.

    Thank you very much.

    Are there any questions? No.

    I will take this opportunity to ask you a question. You were talking about sustainability and renewal of the public service, and saying that it is important because there is a lot of retirement. People were talking about how we've all done a good job, but as Mr. Drouin pointed out, people who complain are the people who are not in the upper echelons. It's very difficult for them to become deputy ministers or ADMs.

    Do you hire outside the public service for those positions, the way a CEO would be hired? Would you do that? And would that be a name-blind or anonymous application?

     Thank you, Madam Chair. That's a good question.

    When it comes to the upper echelons of the public service—deputy ministers, deputy heads, and associate deputy ministers—that is under the responsibility of the Privy Council Office. Those are Governor in Council appointees. They fall outside the jurisdiction of the Public Service Employment Act. All the way up to senior or deputy minister falls under our jurisdiction. Yes, certainly I know—my colleague might want to talk about talent management, for example—that diversity is an extremely important factor that we consider in talent management.

     Do you want to talk a little bit about it?

    Sure.

    With regard to talent management, it's actually quite a rigorous process at the public service in terms of determining the learning plans for the individuals, the skill gaps, and how we can support those employees to finally develop and access the higher echelons and move through the ranks of the public service. It's a process that happens yearly. It's a process that is done in quite a rigorous manner by committee, by department, and finally by the public service at large through each year.

    With regard to your previous question about hiring, my colleague mentioned that it is managed by PCO. It would be hard to provide some stats on that right now, but there are in fact signs of mid-career hires that are coming. I'll speak to the level below deputy, where in fact there are some mid-career hires happening. There are some efforts needed to increase those mid-career hires, though, that need to take place. That's what we're working on right now.

    From the Public Service Commission's perspective, we recognize that not only do we have to focus on external hires for entry-level positions—which is great—but also, in some cases, we need to find people at mid-career.

    The issue is that they're not usually on our website looking for job opportunities, so we're thinking about ways we could be more proactive. In fact, one of the experimentations we want to do next is to look at whether an employee referral program could be piloted. When you're looking at very hard-to-staff jobs.... I'll talk about my own parish. Psychologists are very important to our work. It's almost impossible to attract them through our regular recruitment programs. Again, it could be something where colleagues, psychologists who are currently working in the government, could refer professional colleagues—not their brother-in-law or their sister, but professional colleagues—who may not be thinking of a career in the public service but who could actually really contribute to our workforce.

    That's something we're looking at to see if we could do a better job at attracting at that level, as well.

    What are the next steps for this pilot project? Where does it go?

    The pilot project is complete. We will continue to share and discuss the results. There are colleagues in other countries who are interested in doing some comparisons. We actually are also interested in the U.K. and their model, which they haven't studied yet. They would like to use our study to be able to determine whether the work they've done on name-blinding is having the desired effect.

    We'll continue, of course, to want to find out more, but our efforts now will go in the direction of the three follow-up actions I talked about, particularly, in the immediate future, designing this audit work in the right way so that we can have a stronger body of evidence in a year or so that we could share with the committee.

    Thank you very much.

    Thank you to the witnesses. Thank you for being here.

    [Proceedings continue in camera]