Thank you, Madam Chair.
Allow me to continue the discussion on the important request we made for data. I would like to share with you the work of Professor Julien Larrègue, associate professor in the department of sociology at Laval University and a member of the Centre interuniversitaire de recherche sur la science et la technologie, who appeared before the committee.
By way of introduction, Mr. Larrègue discusses the stratification of the academic world, which is a central theme in the sociology of science. A minority of universities and researchers concentrate the bulk of resources in the scientific field, whether defined in terms of symbolic, economic or social capital. In addition to this unequal distribution, a well-known Matthew effect is also at work, as Harriet Zuckerman demonstrated empirically in her classic study of Nobel Prize winners, and as Robert K. Merton also demonstrated by expanding on the issue. Researchers and institutions that have accumulated more scientific capital tend to be disproportionately rewarded for their work, whilst those less advantaged receive less reward than their actual contribution would merit. In other words, early success increases the likelihood of future success, which further reinforces existing hierarchies.
Although a growing body of work shows that research funding is also subject to concentration phenomena and a Matthew effect, the precise extent of institutional inequalities in the distribution of research grants remains relatively difficult to establish. A significant limitation in the literature lies in the fact that most studies only take into account accepted grant applications, which makes measurements and explanations imprecise at best.
To date, only a few studies have had access to comprehensive data from funding agencies—you will appreciate that this touches on the issue of the initial motion. Yet this is essential because, without a clear understanding of the total initial number of applications, whether accepted or not, it is simply impossible to determine the cause of any disparities in the distribution of grants among researchers and institutions. In particular, it is not possible to determine whether the observed inequalities are due to the fact that certain groups submit fewer applications in proportion to their size, whether they result from the practices of evaluation committees, or whether both are true. This data is also necessary to make an analytical distinction between success rates and the size of the grants awarded. Inequalities may stem not only from the number of grants awarded to researchers at different levels of the academic hierarchy, but also from the amount of the grants obtained.
Another significant limitation lies in the analytical disconnect between quantitative and qualitative approaches to scientific inequalities. Quantitative analyses of funding distribution tend to treat science as a homogeneous space from which generalizations can be drawn, whereas we know it is fragmented and marked by struggles. Whilst such macro-level studies are important for documenting general dynamics, it is absolutely crucial to understand how these dynamics vary across disciplines. Although this aspect has been extensively studied from a qualitative perspective, further research is needed to understand how the evaluative culture of disciplines relates statistically to the reproduction of funding inequalities.
In this article by Professor Larrègue, an original dataset comprising 56,680 grant applications submitted to the Social Sciences and Humanities Research Council of Canada between 2000 and 2021 is combined with 43 interviews with former council members to analyze funding inequalities among Canadian universities.
There are three main questions. Firstly, how does the success of funding applications vary across institutions? Secondly, do the amounts awarded vary accordingly? Thirdly, are these trends consistent across disciplines?
University affiliation plays a predominant role in the allocation of funding, even when controlling for other factors. Researchers employed at more prestigious and larger institutions are more likely to secure grants and receive higher amounts.
It is important to emphasize that the study’s findings demonstrate that this effect varies across scientific fields. Whilst disciplines such as management or economics exhibit clearly elitist patterns, academic hierarchies play a somewhat less significant role in disciplines closer to the humanities, such as history, anthropology or fine arts. Drawing on the sociology of science and the sociology of quantification, the study argues that evaluators use academic affiliation as a judgmental tool to mitigate the uncertainty inherent in the assessment of grant applications.
This uncertainty manifests itself at two levels: at the applicant level, where the proposed research involves promises that are difficult to verify, and at the committee level, where evaluators may be uncertain about the limits of their own expertise in judging proposals. In this context, university affiliation reflects not only symbolic capital, but also a set of easily interpretable and institutionally recognized signals. Consequently, applicants affiliated with more prestigious institutions receive higher scores for their CVs, which improves their overall chances of securing funding. This dynamic is particularly pronounced in disciplines with strict and consensual definitions of scientific quality, such as economics, where university affiliation closely corresponds to other recognized signals of value.
Furthermore, focusing on generic characteristics of applications such as university affiliation, the number and type of publications, and previous funding, among others, also allows evaluators to ensure the comparability of applications. Faced with the challenge of ranking unique and seemingly incomparable projects, committees rely on cross-cutting indicators of past success as predictors of future performance. Much like credit ratings in the financial sector, these appear to be seemingly disinterested, impartial and objective measures.
We now come to point two: the work of evaluating grant applications. Much research has been devoted to understanding the factors that influence the peer review process in the context of scientific funding. A famous experiment conducted at the National Science Foundation in the United States demonstrated that outcomes depended largely on chance, and in particular, on the individual appointed as the reviewer. Although chance clearly plays a role in funding outcomes, structural mechanisms such as cumulative advantage often reinforce and amplify initial inequalities over time.
As DiPrete and Eirich explain, a strict process of cumulative advantage implies that current resources, such as past publications or previously obtained grants, increase the chances of securing new ones. Conversely, a cumulative disadvantage may result from prolonged exposure to a position of lower status, leading to direct and interactive effects on outcomes throughout one’s career. For example, individual characteristics and institutional affiliations, such as a university’s prestige or gender, can influence both initial opportunities and the long-term returns on acquired resources.
A researcher affiliated with a less prestigious institution may be penalized not only because of perceptions regarding the institution’s quality, but also due to material constraints such as reduced research time or less administrative support. A study of research grants from the U.S. National Institutes of Health showed that working at an institution receiving more funding increased the probability of securing a grant by 9.7 percentage points. Although each application may be viewed as a one-off assessment, over time, the symbolic capital, procedural knowledge and institutional support accumulated by certain individuals and organizations create a system that structurally favours them in grant competitions, a dynamic that clearly reflects the operation of cumulative advantage mechanisms.
Furthermore, the literature clearly shows that disciplines and research fields influence and shape evaluation processes. This is particularly evident in studies on gender inequalities in funding.
For instance, professors Larrègue and Nielsen have studied the working of an interdisciplinary social sciences committee to show how gender inequalities in scientific funding are partly reproduced and mediated by knowledge hierarchies. Compared to other projects, tendentially feminine research topics and methods had more chances of being discredited by reviewers. Similarly, while women are significantly less likely to be funded by the Canadian Institutes of Health Research, the extent of this disadvantage varies across research domains. A natural experiment indicated that this gender gap might be most consequential when evaluators focus on applicants' CV instead of their project.
The capacity of some disciplines to develop their own, autonomous definitions of scientific quality plays an important role in how peer review separates the wheat from the chaff. For instance, the so-called superiority of economists, which translates into a strict ranking of journals and institutions as well as in widely shared views on the unequal worth of different methods and research areas, leads to very clear outcomes: Projects and scholars that deviate from this orthodoxy are much less likely to be funded. Hence, funding distribution is not only a material mechanism but a symbolic process through which scholars and fields of research are ranked and given a certain value.
Because such orderings are context-dependent and may vary across disciplines, it is unclear how generic factors like university affiliations and their associated prestige factor in these processes. While some studies report clear correlations, others do not find institutional prestige to be a decisive factor in grant outcomes. In Canada, it has been found that the prestige of universities influences the size of grants among successful Social Sciences and Humanities Research Council, or SSHRC, beneficiaries, but it is unclear whether this pattern is stable across disciplines or if it matters during the evaluation process.
In her ethnographic study of how U.S. panels distribute fellowships and grants, Lamont observed that “evaluators are most concerned with disciplinary and institutional diversity, that is, ensuring that funding not be restricted to scholars in only a few fields or at top universities.” However, committee members' explicit commitment to diversity might not necessarily translate into increased funding opportunities for applicants from less prestigious universities; there could be a gap between individual reviewers' meritocratic beliefs and the aggregate result of their work.
Moreover, we can expect institutional prestige to play a different role, and have a different meaning, across disciplines. If excellence and diversity are indeed additive considerations for grant reviewers, diverging conceptions of how scientific excellence relates institutional placement might influence review processes.
This article suggests that the role of university prestige partly depends upon the weight that reviewers give to applicants' CV versus their project. While SSHRC rules explicitly state what the respective weight of the CV and project should be, in practice committees implement discipline-specific practices that give precedence to one or another.
I will now turn to section 3.1, which covers funding data and is in the third section, which deals with methods and data.
These investigations are grounded on a database of 56,680 grant applications that were submitted to the SSHRC between 2000 and 2021.
This dataset, obtained through a data-sharing agreement, includes information on the submission year, the language of the application, the primary discipline of the project, the outcome of each proposal, whether acceptance or rejection, the scores given to application for each of the main evaluation criteria and the amount awarded when successful. The data include applications to three funding programs.
Standard research grants, from 2000 to 2011, can be valued at up to $250,000 over three years. The success rate over the studied period was 38%. With regard to insight development grants, from 2012 to 2021, they were valued between $7,000 and $75,000 over one to two years. The success rate was also 38%. In the insight grants category, from 2012 to 2021, they could be valued at up to $400,000 over two to five years. The success rate over the studied period was 34%.
Additionally, the dataset provides information on the institution, gender and age of all applicants at the time of submission. Only the main applicant was considered for analysis. The primary applicant is the one formally responsible for the submission and administrative coordination of the project, and is typically the main contributor to the design, writing and overall direction of the proposal. Moreover, the study tells us that they only have information about co-applicants if they have applied as the main applicant for another application.
However, this focus introduces certain limitations. It does not capture the potential influence of team composition, interdisciplinarity or collaborative dynamics that may also affect funding outcomes. Using the initial dataset, several additional independent variables were designed.
At the applicant level, they calculated the total number of applications that a given professor has submitted, the number of SSHRC grants obtained in the past and the size of the team for each project.
At the university level, they further coded and included the geographical location, namely, the province; the primary working language, English, French or bilingual; the level of prestige, categorized as U3, U12 or non-U15; and the size, namely the number of students.
For this last variable, the following classification was applied.
There are large universities, that is to say the universities with 20,000+ students, such as the University of Toronto, the University of British Columbia and McGill University. There are medium-sized universities with 10,000 to 20,000 students, such as Queen's University, University of New Brunswick and Wilfrid Laurier University. There are small universities, with fewer than 10,000 students, such as Acadia University, Bishop's University and Mount Allison University. There are very small universities with fewer than 1,000 students, like Trinity Western University, Tyndale University, and Pontifical Institute of Mediaeval Studies.
From these various steps, a binomial logistic regression was conducted to estimate the effects of the following 11 independent variables on the funding success of SSHRC applicants: gender of the main applicant, age of the main applicant, total number of SSHRC applications of the main applicant, team size of the project, funding program, year of application, university prestige, university province, university language, university size and language of the project.
The dependent variable is a binary measure of whether a given application received funding or not. The regression was conducted on RStudio with the function glm.
They then focused their analysis on 30 research-active universities—those with the highest share of SSHRC applications during the period. They compared their overall success rates, followed by the individual applicants' chances of success, controlling for the previously mentioned variables. They also examined the average grant amounts awarded to applicants at each university. Finally, they analyzed success probabilities by university ranking within each discipline, using average marginal effects derived from a logistic regression model.
Similarly, they assessed differences in grant amounts by disciplinary ranking, this time relying on descriptive statistics. The main limitation of this quantitative analysis is that it does not take into account applicants' publication records.
Such data could, in principle, be retrieved from bibliometric databases such as Web of Science. However, the expected analytical benefit would be limited relative to the considerable effort involved in collecting and matching this information at scale.
Moreover, their objective is not to assess individual academic productivity per se, but rather to highlight broader institutional hierarchies—hierarchies that already reflect, among other factors, differences in publication patterns and research visibility of their members.
I'll move on to section 3.2, on interview material.
To understand the mechanisms underlying the unequal allocation of funding, these statistical analyses were complemented by semi-structured interviews with professors who served on SSHRC evaluation committees between 2014 and 2024 for the insight and insight development programs.
The composition of the review committees for each type of funding is publicly available on the SSHRC website. Members were contacted through their institutional email addresses.
They aimed to achieve a balance across disciplines, career stages, such as assistant, associate and full professors, university prestige, geographical location and socio-demographic characteristics such as gender, age, ethnicity and language. The goal was to ensure a diversity of perspectives.
Interviews were held in English or French depending on the interviewees' preference. The interviews primarily explored the practical organization of the evaluation process and the criteria employed by committee members to assess funding applications.
The relative weight given to applicants' CVs versus the content of their research projects and the nature of the discussions that take place during committee meetings were compared. They lasted between 42 and 88 minutes. All interviews were recorded, fully transcribed and anonymized. They focused on the interviewees' experience within the SSHRC committees and asked them to provide concrete examples whenever possible.
As underlined by Orupabo and Mangset in their study of academic hiring practices, focusing on practical information serves as a methodological tool to address social desirability bias. Interviewees are generally less preoccupied with presenting themselves in a favourable light when recounting processes and events compared to when they are directly asked about their opinions, meanings or values.
The qualitative material was thematically coded using NVivo, with particular attention to passages concerning perceptions of the applicant's university and the ways in which institutional affiliation—both directly and indirectly—influenced the evaluation process.
These themes were then analyzed in relation to the disciplinary background of each interviewee to identify potential variations across fields. A total of 43 researchers were interviewed, spanning six disciplines. Fifteen of them were in political science, 13 were in sociology, seven were in economics, six were in history and one was in management.
The four main disciplines were selected due to their contrasting positions within the scientific field, the diversity of their evaluative cultures and their divergent attitudes toward symbolic and status hierarchies.
Economics holds a dominant position in the social sciences, characterized by its quantitative orientation and strong formalism. Its evaluation methods emphasize a strict hierarchy of publications and institutional affiliations, alongside highly centralized and internationalized recruitment processes.
Conversely, history—a literary discipline primarily employing qualitative methods—features predominantly national patterns of scientific production and recruitment. As we shall see, historians often resist hierarchies based on institutional positions.
Political science and sociology occupy intermediate positions between these extremes, although political science aligns more closely with economics than sociology does. Both disciplines exhibit internal polarization in Canada—translating into divisions between qualitative and quantitative approaches, French and English literatures, and tensions regarding national and international dynamics of publication and recruitment.
Despite efforts to ensure a diverse and balanced sample, this qualitative approach presents certain limitations. The study relies on retrospective accounts, which may be influenced by memory biases or selective recollection. Moreover, while the interviews aimed to elicit concrete examples, participants may still under-report practices perceived as problematic or controversial.
Finally, although the sample includes disciplinary and institutional variety, it remains limited in size and cannot fully capture the breadth of experiences and perspectives present across all SSHRC committees.
I'll continue with section 4 on the adjudication process.
Before presenting the findings, it is important to briefly describe the process and organization of the SSHRC evaluations. For each of the three programs, applications are peer reviewed by committees constituted according to disciplinary expertise. While some committees for insight development grants can be interdisciplinary, they are typically focused on one or two proximate disciplines, such as political science and public administration or sociology and demography. Before meeting collectively to decide on the final ranking, each committee member conducts a preliminary review of a subset of applications, with two or three evaluators assigned per file. For the insight development grants, SSHRC also seeks external reviews to support the committee's deliberations.
The repartition of the applications among the committee members is an administrative task handled by an SSHRC officer. Reviewers may and, in practice, often have to assess applications that are not related to their own research interests or fields of research. They assign a score for each of the three main criteria: challenge, feasibility and capability. For every application assigned to them, researchers in the study indicate that they only had access to the scores for a subpart of their dataset. The challenge criterion refers to the purpose and importance of the project. Feasibility refers to the methods and material means used to carry it out. Capability refers to the applicant's expertise, as demonstrated by their CV. In addition, the scoring is weighted, with the challenge and capability criteria each accounting for a larger portion of the score than the feasibility criterion. Preliminary scores are used to establish a provisional ranking when committee members meet collectively. They do not typically review or discuss all the applications. Unless important discrepancies in members' assessments are noticed, top-scored and bottom-scored applications are rarely examined. The discussions focus on the intermediate applications that are around the funding line. In the event of a persistent disagreement between reviewers regarding the evaluation of a particular application, a collective vote may be held after discussing each of these applications. The committee reviews and finalizes the ranking. This final list divides the adjudicated applications into those recommended for funding and those that are not.
I will now tell you about section 5, which looks at the results and the cumulative advantages of prestigious universities.
It's fascinating, Madam Chair, so I hope my colleagues are listening carefully.
Descriptive statistics show that success in funding applications to the SSHRC is indeed correlated with institutional affiliation. Professors who are employed in bigger and more prestigious universities have a higher likelihood of getting grants. It's there in black and white. This is an independent researcher, and that's what it says. Applications from candidates affiliated to a U3 university, the U3 universities being McGill University, University of British Columbia and University of Toronto, represent 19.2% of all applications, but I'd like to draw your attention to the fact that they represent 24.6% of those funded, which shows there's an imbalance. Applications from candidates outside the U15 represent 46.4% of all applications, but only 39.3% of those funded.
These gaps are therefore even wider when considering the broader structure of academic employment in Canada.
Between 2016 and 2020, U3 university teaching staff, so those teaching at University of British Columbia, University of Toronto and McGill University, represented 15.2% of all university staff and 19% of all applications, while non-U15 university staff represented 54.7% of all staff, but only 46.4% of grant applications.
In short, it's very interesting and relevant to note that U3 professors—those at University of British Columbia, McGill University and the University of Toronto—tend to apply more and, when they do, they have more success. The success rate is also correlated with university size. The largest universities have the highest success rate, 40% for the whole period, followed by medium-sized universities at 34%, with the very small universities and small ones exhibiting the lowest success, at 27% and 25%, respectively.
Interestingly, the drop-out rate, that is to say those with no reapplication after one non-funded grant proposal, is structured similarly, with professors affiliated to less prestigious universities giving up more often, at 28% for U3 institutions, namely University of Toronto, McGill University and University of British Columbia, 32% for U12 institutions and 35.7% for extra U15 institutions.
The effect of university prestige and size persists even when accounting for other factors. A binomial logistic regression was performed to predict the success of funding applications between 2000 and 2021. Just as a reminder, through the SSHRC, the researchers were provided with 56,680 applications. Results are presented using average marginal effects and predicted probabilities. All the variables related to the characteristics of the main applicant's university are significantly correlated with funding outcomes.
First, all the variables related to the characteristics of the main applicant's university are, again, significantly correlated with funding outcomes.
Second, candidates from U3 universities—McGill University, University of British Columbia and University of Toronto—have the highest predicted probability of success, at 43.8%. By contrast, those from U12 universities have a lower probability of success, at 37.7%. Finally, those from non-U15 institutions have the lowest predicted probability of 32.6%. These differences are statistically significant, with average marginal effects of -0.6 for U12 institutions, -0.11 for non-U15 universities, compared to U3 universities, which are, of course, in the reference indicator, less than 0.001.
A similar pattern is observed when it comes to university size. Applicants from large universities have the highest probability of success, at 37.8%, while those from small universities, at 29.9%, and those from very small universities, at 31.3%, have a significantly lower chance of securing funding.
The corresponding average marginal effects are -0.08 and -0.07, respectively, reinforcing the persistent influence of institutional capacity on funding outcomes.
Geographic location also plays a role. Compared to Ontario, at 37.6%, applications from Alberta, at 34.4%, and from the rest of Canada, at 31.5%, show significantly lower predicted probabilities of success.
That's noteworthy, Madam Chair.
So for those with significantly lower predicted probabilities of success, the data show average marginal effects at -0.03 and -0.06, respectively, while applicants from British Columbia, at 40%, have slightly higher chances. Quebec shows no significant difference from Ontario, and we're very pleased about that.
Interestingly, the role of language is not straightforward and individual level factors must be distinguished from institutional dynamics. Applications written in French are associated with a lower probability of success.
You did not hear this from the Bloc Québécois. It comes from an independent researcher, funded by the federal government, namely by the Social Sciences and Humanities Research Council.
This researcher tells us that applications written in French are associated with a lower predicted probability of success than applications written in English—33.2% versus 37.4% for English. Affiliation with a francophone university corresponds to a higher predicted success rate compared to English-language universities. Although it is beyond the scope of this article to provide a thorough analysis of this apparent paradox, these findings demonstrate the need to differentiate between individual and institutional levels when accounting for the effect of language in science.
Gender and age are also associated with differences in predicted funding success. Predicted probabilities show that women, at 36.2%, have slightly lower chances of success compared to men, at 37.1%, while non-binary applicants, at 38.8%, exhibit the highest predicted success rate. However, these differences are modest in magnitude. The average marginal effect for women is negative and statistically significant, at -0.01, while the effect for non-binary applicants is not statistically significant due to wide confidence intervals.
Predicted probabilities decrease with age: Younger applicants, at age 26, for example, have a predicted success rate of 39.7%, while the rate declines steadily with age to 35.7% at age 54. This relationship is captured by a small but statistically significant negative marginal effect of age on funding success, suggesting a slight but consistent age-related disadvantage over time.
That's not reassuring, Madam Chair. In Canada, the older you are, the less likely you are to obtain funding from the Social Sciences and Humanities Research Council. One might say that's ageism.
So, while being affiliated with a prestigious university is generally a key factor in securing funding at SSHRC, U15 universities do not necessarily exhibit the same levels of success in grant applications.
We can obtain a more fine-grained picture by looking at the performances of 30 research active universities—representing 83.7% of the full sample—constituted of all U15 institutions and 15 non-U15 institutions that are most represented in SSHRC applications.
Eight universities, seven of which are from the U15, appear to be overfunded. They receive more grants than their proportion of the applications would allow. The vast majority of universities, however, proportionally receive less funding than one could expect from looking at their share in the full sample of applications.
Again, there are important differences across institutions. While some are close to equilibrium, others are starkly disadvantaged. For example, the University of Manitoba's share of funding is 16.6% lower than that of all universities. Are any members from Manitoba following our proceedings? I hope they're listening. An independent researcher is telling them that their university is funded as part of the research funding allocation. The University of Saskatchewan receives 26.4% less than all universities. Are any members from Saskatchewan following our proceedings? That confirms that they are less likely to get funding. In the case of Brock University, the variance from all applications is 33.4%.
These trends are confirmed if we analyze performance across these 30 institutions while controlling for a few independent factors, as figure 1 clearly shows: I would nevertheless take the liberty of explaining it to you. There are wide disparities in funding success across universities, including within the U15 group. That's quite astonishing.
In line with the researchers' hypothesis, three universities lead the way when it comes to obtaining research funding in Canada. They are University of Toronto, which has a 49.1% predicted probability of success; McGill University, at 48.5%, and University of British Columbia, at 47%.
While most U15 universities are indeed located in the top half of the distribution, others, like the University of Calgary, are less likely to secure funding. Again, are there any members here from Alberta who are following our proceedings? No. The University of Calgary has a predicted probability of success of 32.4%, ranking 19th and occupying a median position within this top 30.
I want to point out that other institutions are even less advantaged, even though they are U15: Dalhousie University, which has a 31.2% predicted probability of getting a grant; University of Manitoba, at 30.4%; and, again, University of Saskatchewan, at 25.1%.
This highlights that the U15 group, which gradually expanded since the early 1990s, only partially reflects the actual scientific hierarchies. Notably, among the five universities that joined the group in 2006 and 2011, which are University of Calgary, Dalhousie University, University of Ottawa, University of Manitoba and University of Saskatchewan, only the University of Ottawa falls in the top-performing half, with a predicted probability of getting a grant of 38.8%. As you know, the University of Ottawa is a bilingual university, which could explain the situation. I think that's something that would need to be verified.
In contrast, some institutions that are not part of the U15, possibly because they are not medical schools, appear among the best performers.
Applications from Simon Fraser University have a predicted probability of 43.4%, ranking fifth overall. Université du Québec à Montréal follows at 10th place with 37.2%. University of Victoria ranks 14th with 35%.
Researchers from the largest and most prestigious universities not only have higher chances of securing funding, but they also tend to receive larger grants when successful.
It's important that my colleagues understand that. Researchers from the most prestigious universities have more money and, not only that, they have it for longer.
From 2000 to 2021, U3 universities got 25.8% of the total funding when they represented 24.6% of the successful applications. U12 universities got 36.2% for 36% of the successful applications. The rest got 37.9% when they constituted 39.3% of the successful applications.
As we saw previously, disparities are also visible in figure 1, which I explained to you earlier. There is a subsample of 30 research active universities.
I'll take the liberty of again explaining a figure that presents the greatest disparities in the insight program. The amount of funding that went to U3 projects, $67,286, is 20.5% higher than what went to U15 universities, $55,817, and 38.8% higher than what went to non-U15 institutions, which was $48,488.
Let's recap. First, U15 universities are more likely to get funding. Second, they have it for a longer period of time. Third, they get higher amounts than other universities.
Usually, three strikes is a strikeout, or a home run, for those who are currently favoured in Canada's research funding system. The amount received by McGill University researchers in the insight program is $76,104. It is $54,354 for the University of Ottawa; $50,189 for Toronto Metropolitan University;$38,413 for the University of Calgary; and $23,003 for the University of Windsor.
The discrepancies are even more important when we divide the average amounts received by the number of applicants per project: for the insight program, the average individual amount allocated to U3 projects, $48,524, is 44% higher than those of other U15 universities, $33,783, and 69% higher than that of projects conducted at institutions outside the U15, $28,781.
This ranking closely matches the success rates analyzed previously. The correlation coefficient between universities' odds ratios and average amount received is 0.94. Put otherwise, the more successful a university is, the bigger the grants.
While the interviews conducted confirm that committees sometimes make budget cuts when reviewing and selecting projects for funding, they are generally limited to about 10% to 20% of the requested budget. It is therefore unlikely that the disparities observed are primarily the result of the work of the evaluation committees. This is important. The purpose of the motion was to request data from the evaluation committees so that it could be subject to an outside and confidential analysis. It was an independent, government-funded researcher who told us that. Those people have credibility and that's what they said. Evaluation committees have biases, so we have to try to mitigate those biases. To do that, we need to conduct an analysis with access to data.
To some extent, this can also be explained by the fact that candidates from more prestigious universities are more likely to apply for and obtain higher amounts. It's really quite incredible. These people know that they are at universities that are able to secure more grants. They have them for longer, so they ask for higher amounts than others do. If people thought there was no inequality in the research funding system, I am telling you today that that is not the case.
If I may, I would like to go to point 5.2, which pertains to university hierarchies according to the disciplines in the curriculum vitae and the disciplines of the project. In particular, it states that, to date, funding applications to the Social Sciences and Humanities Research Council of Canada have been approached as a homogeneous whole.
I would point out that the researchers did not have access to data from the Canadian Institutes of Health Research or from the Natural Sciences and Engineering Research Council of Canada. It's not because they didn't want to do the work, it's because they didn't have access to the data.
We are now at a parliamentary committee and we have to ask the granting agencies to be transparent and to provide that data to certified researchers or to independent analysts from the Library of Parliament who have the expertise, knowledge and tools to conduct those analyses, so that we can then produce the best possible reports on our studies and in turn have the best public policy.
Evaluation committees from different disciplines do not necessarily give the university hierarchies the same weight when evaluating and ranking projects. It is therefore important to analyze the effect of institutional affiliations by discipline. That means that, depending on the university where people are located, their interests differ. In addition, their relationship with other institutions can influence the decision to award funding.
While candidates from prestigious universities in certain fields have a clear advantage, hierarchies seem to play a somewhat less important role in other fields.
In economics, for example, applications from non-U15 universities have a 29.6% probability of success as compared to 41.6% for U3 institutions, that is, the University of Toronto, McGill University and the University of British Columbia. The same is true in management, administration and business studies. The probability of success for non-U15 applications is 29.5% compared to 44.1% for U3 institutions.
Conversely, the gap is much smaller in certain disciplines such as anthropology, where non-U15 applications have an expected probability of success of 36.1%, which is quite close to the expected probability of 37.3% for applications from U3 institutions. We therefore commend scientific researchers in anthropology. They have the least bias by scientific field, and as indicated in this study that I am very pleased to share with you today.
Nevertheless, the best that candidates from non-U15 institutions can hope for is not to be disadvantaged. In no discipline do they outperform their counterparts at U15 institutions in terms of the predicted success rate.
Again, U15 universities are more likely to receive funding. They receive funding for a longer period of time. They receive higher amounts of funding. However, when you get into the details, as we're doing right now, the field of study doesn't matter. It's like saying that researchers outside those universities are not as good because they receive less funding.
I doubt it, but those researchers confirm hypotheses that we have been raising for a very long time at the current Standing Committee on Science and Research, which was created in 2021. This has not been the case since just 2021, but rather for many years. There is inequality in the distribution of research funding in Canada. It has been condoned, to be sure, by successive governments.
Once again, the discrepancies involve not only the success rates, but also the amounts awarded. In the disciplines with the greatest differences between U15 and non-U15 universities, there is a tendency to allocate higher amounts to the former. It's quite incredible.
Across all programs, the average amounts awarded to projects from non-U15 universities are lower than those awarded to projects from U3 universities: 62.6% in management, 54.2% lower in economics, 50% lower in criminology and 49.9% lower in urban studies.
Conversely, in the fine arts, literature and anthropology, the differences are smaller: 11.1% in fine arts, 24.5% in literature and 21% in anthropology. You will recall that I said that people in anthropology are the ones with the fewest differences. We applaud that.
Similarly, compared to researchers at U3 universities, faculty affiliated with the other U15 institutions consistently receive lower amounts. The only exception is for law. I think my colleagues know that. I have a lot of colleagues around the table who are interested in law. As a legislator, it is important.
As the researchers noted in this study, there is therefore a clear statistical correlation between success rates and the amounts distributed by discipline. There is a coefficient of correlation between the success rate of non-U15 universities by discipline and the funding gap between U3 and U15 universities. It is 0.67% and 0.50% when you include U12 universities.
Data on knowledge grants indicate that these disparities are based primarily on the assessment of candidates' CVs, which the Social Sciences and Humanities Research Council's documentation refers to as “ability”. While faculty at U3 institutions received an average score of 4.88 out of 6, applicants from non-U15 institutions received a score of 4.62, a difference of 0.26. Of the three evaluation criteria, this is the one where the effect of the university is most pronounced. For the “challenge” criterion based on the project assessment, the average difference between—