A cross-dimensional analysis of nanotechnology and equality: examining gender fairness and pro-poor potential in Canada’s R&D landscape

ABSTRACT

This study provides a cross-dimensional analysis of two equity concerns related to Canadian nanotechnology, investigating the relationship between the development of nanotechnology applications that benefit the poor and the gender gap in the scientific workforce. Many affluent countries, like Canada, aspire to use R&D to reduce inequality in both economic and gender dimensions, which makes cross-dimensional analyses essential for responsible innovation to fully understand how technologies affect equality and to guide policy. Relevant publications and patents are analyzed to explore if Canadian nanotechnology was addressing the needs of the poor and then to examine gender disparities in research and innovative advancements of pro-poor applications of nanotechnology. Only a small percentage of analyzed articles and patents reflected pro-poor priorities and Canadian workplaces involved in pro-poor nano-applications were largely male-dominated. Results suggest that coordination between pro-poor and gender-responsive policies is needed to promote both more equitable and more inclusive forms of innovation.

KEYWORDS:

• Nanotechnology
• gender
• pro-poor
• bibliometrics

Introduction

Since its inception, scientists and policymakers have promoted nanotechnology ¹ as a platform that can provide benefits for a wide range of industries and create significant economic and societal benefits (Roco and Bainbridge 2001). For roughly two decades, firms, governments, and universities across the world have been increasingly investing in this technology for reasons of technological competitiveness (Bae et al. 2016; Forster, Olveira, and Seeger 2011; Islam and Ozcan 2017). However, also from its inception, nanotechnology policy and scholarly literature has included a strong emphasis on the responsible development of nanotechnology, which broadens analysis beyond the economic realm, focuses on understanding complex distributions of outcomes, and examines how governance mechanisms can be designed to steer innovation towards just and sustainable outcomes (Fisher 2019). More recently, the normative governance of nanotechnology has continued as an explicit theme in the scholarship on RI (Fisher and Rip 2013; Pandza and Ellwood 2013; Radatz et al. 2019; Wiek et al. 2016). We contribute to this ongoing body of work, particularly work on nanotechnology, equity and RI in the global south (Beumer 2018; Foladori and Invernizzi 2018; Harsh et al. 2018; Hartley et al. 2019; Vasen 2017) and work on nanotechnology and gender (Ghiasi, Harsh, and Schiffauerova 2018; Meng 2018; Villanueva-Felez, Woolley, and Cañibano 2015) as well as work on RI in the global south (De Hoop, Pols, and Romijn 2016; Macnaghten et al. 2014; Valkenburg et al. 2020). More generally, we note that inclusion and gender are central concepts in prominent frameworks for RI (Owen and Pansera 2019). We provide a recent historical case study from Canada examining the relationship between the development and commercialization of nanotechnology applications that benefit the poor (pro-poor) and the gender gap in the scientific workforce. Conceptually, our analysis provides a way to examine inequality across economic and social dimensions. Empirically, the focus on Canada provides an examination of a country often overshadowed by its neighbor to the south and looks at a critical period for Canada’s nanotechnology and research and development (R&D) strategies. Methodologically, the combination of bibliometrics with employment data provides a novel approach to contextualize R&D analysis.

This study applies Cozzens’s theory of equity and nanotechnology (Cozzens 2010) to provide the first quantitative, cross-dimensional analysis of nanotechnology and inequality. Vertical inequalities refer to unequal distribution of wealth and income between counties or individuals (rich-poor), and horizontal inequalities concern differences between groups in terms of culturally-defined categories (e.g. gender and ethnic inequalities). Both types of inequalities are important for RI because they are factors that hinder social cohesion ² and inclusion ³ (Cozzens et al. 2007; Cozzens and Wetmore 2011). Cross-dimensional analyses are crucial to more fully understand the relationship between nanotechnology and inequality because exploiting nanotechnology’s potential for the poor could lead to a less equitable global society if it widens inequalities at horizontal dimension while trying to rectify vertical inequalities.

Cross-dimensional analyses are especially important to inform policymaking in advanced economies, like Canada, that are actively trying to use R&D to promote equality in both vertical and horizontal dimensions by improving gender fairness in the scientific workforce (NSERC 2017), and by using R&D as a tool to help international development (IDRC 2015). We examine a critical time for nanotechnology in Canada’s recent history: the first decade of this century. The 2000s were a ‘nano-decade’ when Canada (and many other countries) had a strong focus on nanotechnology. For instance, Canada’s National Institute for Nanotechnology opened in 2001, along with the analogous NanoQuebec initiative, and by that time there were already five National Research Council facilities conducting nano research in Ontario (nanowerk 2006). Furthermore, this decade was also an important time for pro-poor technology in Canada. In 2004, Canada set a target of using 5% of R&D investment to address challenges in developing countries (Salamanca-Buentello et al. 2005).

To examine this key period, the study combines bibliometrics with workforce analysis in a novel way. We use bibliometric data from publications (Scopus) and patents (US PTO) to assess whether nanotechnology R&D addresses issues facing poor communities in developing countries and provide an improved understanding of women’s contribution to these advances over time. ⁴ We then use Statistics Canada data to further examine where potentially pro-poor nanotechnology-focused industries stand in the Canadian economy in terms of gender gap in wage and employment.

The paper is structured as follows. We first present more background on pro-poor, gender and employment dimensions of nanotechnology. We then detail the methodology. The next three sections provide findings and analyses of publication, patent and employment data respectively. Only a small percentage of Canadian nanotechnology articles and patents reflected pro-poor priorities and Canadian workplaces involved in pro-poor nano-applications were largely male-dominated. The conclusion then reflects on implication for method, theory and policy.

Background

Pro-poor nanotechnology

Nanotechnology has become a widely debated case study for whether and how emerging technologies might have positive impacts on poor and marginalized communities (Foladori and Invernizzi 2018; Harsh et al. 2018). In an often-cited study, Fabio Salamanca-Buentello and colleagues (2005) ascribed the most promising nanotechnology applications to help meet the United Nations (UN) Millennium Development Goals (MDGs), which are now incorporated into the Sustainable Development Goals (SDGs). The study argued that nano-applications in energy, agri-food, and water are the most likely to benefit the poor, improve livelihoods, and contribute to achieving several of the MDGs: eradication of poverty, decrease in child mortality, improvement of maternal health, control of HIV/AIDS and other diseases, and environmental sustainability. Nonetheless, none of the top pro-poor nanotechnology applications contribute to the third UN MDG, namely gender equality and empowerment of women. This underlines that it is not possible to reach global equity by introducing novel nano-applications that might reduce the gap between the poor and the rich while also increasing gender disparity. Hence, it is of great importance to look into the cross-cutting relations of both dimensions of inequality and understand how the development and commercialization of these pro-poor applications affect gender equality.

Invernizzi and Foladori (2005) found Salamanca-Buentello’s argument too ‘optimistic’ and claimed that new technologies are designed for the advantaged rather than the disadvantaged. Cozzens et al. (2013) further empirically tested the argument of Salamanca-Buentello et al. (2005) by examining nanotechnology research advances in energy, agri-food, and water that have pursued the MDG-related priorities and argued that both developed and developing countries benefit from these applications. The last UN SDG which envisions creating a global partnership for development highlights the social responsibility of developed countries to address the needs of developing countries and to ensure that developing countries can benefit from the new technologies. This study accordingly seeks to corroborate the work of Cozzens and her colleagues (2013) and specifies how nanotechnology R&D in Canada (i.e. one of the top affluent countries) during a key recent period was relevant to the context of developing countries. We examine scientific and technological advances in the top potential ‘pro-poor’ nano-applications (energy, agri-food, and water).

Nanotechnology and women

Issues related to women in science, technology, engineering and mathematics (STEM) are often seen as threefold (Schiebinger 2008). Firstly, there is a research focus on investigating women’s participation in STEM at an individual level where the concepts of ‘glass-ceiling’ (Hymowitz and Schellhardt 1986) and ‘leaky pipeline’ (Berryman 1983) have been coined to address gender barriers women face in their career path. A second focus is on analysis of gender in the cultures of STEM at an institutional level, which reflects structural and cultural changes through the introduction of gender-related policies at national or institutional level. Third, there is a focus on gender and the outputs of STEM, which addresses how incorporation of gender dimension into research leads to new discoveries and hence stimulates scientific excellence.

When it comes to nanotechnology specifically, the works of Smith-Doerr (2011) and Meng and Shapira (2011) shed light on nanotechnology and challenges of gender equity. By employing feminist theories and comparing the nanotechnology development to biotechnology, Smith-Doerr (2011) attempted to explore the probable place of women in nanotech-related research and production and highlighted that nanotechnology was a field where more research was needed to uncover inequalities in STEM. Meng and Shapira (2011) further tried to explore one of the areas Smith-Doerr (2011) highlights, namely nanotechnology patenting by women, and concluded that despite being highly male-dominated, there was a gradual reduction in the gap distance between female and male patenting. Meng (2018) further suggested that women might be better situated in nanotechnology because more collaboration opportunities were available to them in the field of nanotechnology. Nanotechnology’s scientific production is also male-dominated: female to male ratio of publication productivity in nanotechnology is 0.28 worldwide and is 0.26 for Canada (Larivière et al. 2013) and it could take more than 25 years for women and men to be equally represented in nanotechnology authorship (Holman, Stuart-Fox, and Hauser 2018). Scientific collaboration ties are weaker (Ghiasi, Harsh, and Schiffauerova 2018) and less informative for women in nanotechnology (Villanueva-Felez, Woolley, and Cañibano 2015). Smith-Doerr (2011) calls for further research into gender equity in the emerging nanotechnology sector, underlining nanotechnology’s importance to policies that try to halt emerging gender inequity in STEM and promote equality. This paper contributes to this literature by tracing women’s scientific and innovative productivity and impact in pro-poor nanotechnology applications.

Nanotechnology and employment

It has long been argued that nanotechnology’s promise for job creation is huge (Palmberg, Dernis, and Miguet 2009). Perhaps most well-known are the estimates of Roco (2011, 2017, 2018), who found that employment growth rate is 25% in nanoscience and nanoengineering, increased from 60,000 to 600,000 workers worldwide over the period 2000–2010 (Roco 2018, 22). He (2017) estimated that nanotechnology-based labor and market would double every three years, creating 6 million jobs by 2020 and 60 million jobs (Market value of 30,000 billion dollars) by 2030 worldwide. For Canada alone, it was predicted that the country would own up to 10% of the nanotechnology market share by 2020 (Alberta Advanced Education and Technology 2007). These forecasts vary based on different definitions of nanotechnology, the degree of optimism, and the degree that nanotechnology can add value to final products (Hullmann 2007). Shapira and Youtie argue that the forecasts for total jobs have been shrinking since 2004 (Shapira and Youtie 2015). The evidence on how many jobs have been created since these predictions were made in the early 2000s is mixed. Stephan, Black, and Chang (2007) and Freeman and Shukla (2008) found the number of nanotech-related jobs is ‘few’ and job growth in nanotechnology ‘modest’. However, the state of New York, which has invested heavily in nanotechnology, has exceeded its 2008 job projections (Wessner and Howell 2020). Moreover, Roco (2017) showed that market value in 2014 exceeded the market value he predicted for 2015 in (Roco 2011).

When considering workforce equity, the types of nanotechnology jobs are just as important as the number of jobs. The majority of the nanotech-related workforce is formed by highly educated scientists and engineers (Invernizzi 2011), and thus not positions that are open to most people in the world. However, the broad spectrum of nanotechnology creates the need for workers with a range of skills at different levels of the production chain (e.g. in manufacturing, sales, marketing, and distribution) (Invernizzi 2011). Hence, nanotechnologies might still have a potential to create growth in jobs open to more of the population and not just the highly educated. This study helps us understand the kind of jobs nanotechnology created in Canada by examining industries where companies involved in Canadian pro-poor nanotechnology R&D in over a decade and identifies the gender gap in employment and wage.

Methodology

Data

The dataset used in this study is taken from Ghiasi, Harsh, and Schiffauerova (2018). Scientific publications are extracted from the Scopus database. Scopus accumulates and presents one of the largest article abstract and citation databases spanning science, technology, medicine, social sciences, and arts and humanities. Scopus provides an extensive coverage of data about peer-reviewed scientific publications in multiple disciplines including abstract, citation (per each year) and author information data (for each co-author), which gives Scopus significant advantages over other databases, as these data are very valuable and useful for assigning different attributes (e.g. gender and sector). The patent data come from the United States Patents and Trademarks Office (USPTO) which has the largest coverage of patents registered in North America and provides patent data information, including inventor name, assignee name and location, international patent classification (IPC) code, application and grant date, and geographical location of the residence for each inventor. The nanotechnology-related article and patent data are extracted using a full-text keyword search strategy. More details can be found in Barirani, Agard, and Beaudry (2013) and Moazami, Ebadi, and Schiffauerova (2015).

For this study, we analyze articles published in 1996–2011 and patents granted over the period 1996–2009. As mentioned above, the 2000s were years where Canada made large investment in nanotechnology R&D and a period when Canada was pledging to conduct R&D aimed to address issues in developing countries. 1996 was selected as the start date to allow analysis of any ‘ramp-up’ in nano R&D before 2000s. 1996 is also the first year in which Scopus has full coverage. 2011 was selected as the end year for publications because the National Research Council Canada reoriented its national strategy to become more industry-focused organization and made a transition from research to technology and product development in 2011 (NRC 2016). 2009 was used as the end year for patent analysis because we utilized a specialized database of nanotechnology patents (details can be found in Barirani, Agard, and Beaudry (2013)) for which we only had access to data through 2009. The specialized database is necessary because simply using the USPTO classification code for nanotechnology (977) drastically underestimates nano patents, as discussed in Barirani, Agard, and Beaudry (2013).

Canadian nanotech-related articles involving water, energy, and agri-food applications are identified using the keyword filters proposed by Cozzens et al. (2013), and those where at least one author is affiliated to a Canadian institution are extracted. Similarly, patents are classified into those three applications using the same keyword filters while looking only into titles and abstracts.

The gender is assigned to each author using the GenderChecker name and gender database, which uses 2001 and 2011 UK Census Data, and multiple other sources. ⁵ The main advantage of this database is that this database includes names that are assigned to only one category (Female, Male, or Unisex) with a high degree of certainty. For those authors where GenderChecker assigns a unisex designation or is not able to assign a gender designation at all, gender is manually assigned based on author’s academic, professional or Linkedin profiles. Note that gender is assigned to all the authors and inventors identified. Academic, governmental and industrial sectors are assigned to authors based on their affiliations. Similarly, provinces are assigned based on where affiliated institutions are located. For patents, different sectors are assigned based on the assignee sector type, i.e. whether it is a university, governmental agency or a company.

Canadian author-inventors (A-Is) are Canadian inventors whose names also appear in the Canadian nanotechnology articles database. In this study, A-Is are assigned to those who were involved only in publishing and patenting in pro-poor nanotechnology applications. For this purpose, first or middle initials are removed from given names. Afterwards, inventors and authors are paired by matching their (main) given names and last names. A-Is are identified as those pairs with patent(s) and article(s) similar in title(s) and abstract(s), or those pairs whose main subject area(s) of publications (in Scopus database) are similar to the international patent classification (IPC) code(s) on their patents, or those pairs whose province(s) of their current and previous affiliations in the Scopus database.

A total of 1157 articles and 2528 authors, 365 patents and 608 inventors, and 43 A-Is are identified. GenderChecker was able to identify 85% of authors and 92% of inventors as either female or male. For the rest of the authors and inventors, gender is assigned manually using their academic or professional profiles and other sources. For validation, 1000 random individuals among authors and inventors were selected, and their gender and sector were independently identified and confirmed.

Bibliometrics

The analyses are based on bibliometric indicators of scientific and innovative activities. Bibliometrics is a method commonly used to assess innovative and scientific research excellence through quantitative analyses of patent and research publications. This study deploys quantitative bibliometric indicators to conduct large-scale analyses to measure the scientific excellence and innovative potential for pro-poor nanotechnology applications in Canada.

To evaluate the productivity of researchers or groups of scientists, this study uses the number of publications as the main indicator. Average number of citations per year is used to address the scientific impact. The SCImago Journal & Country Rank (SJR) is also used as a journal impact indicator. SJR is a journal prestige metric, which uses Google’s PageRank™ to rank nearly 17,000 journals based on Scopus data. SJR weights received citations based on the subject field, quality, and prestige of a citing journal (Guerrero-Bote and Moya-Anegón 2012). Due to its specific features, such as being subject-field normalized, exclusion of journal self-citations (maximum 33% of journal self-citations is counted), broader coverage of journals, and compatibility with Scopus data (García-Peñalvo et al. 2010; Guerrero-Bote and Moya-Anegón 2012), SJR is chosen in this study as a measure of journal’s scientific impact.

All of these measures are used based on fractional counts of articles, assigning each author 1/x count of authorship where x represents the number of co-authors in an article. This means that if a paper with five authors has the same citation impact as a sole-authored paper, the author of the sole-author paper is considered to be involved in higher impact work than any of the five individuals of the co-authored paper.

A similar approach is applied to patents. The volume of patents is used as an indicator of technological output and impact of patents is evaluated based on the average number of citations received per year and the number of claims. Each of these indicators is measured as a fractional count of patents where inventorship is measured as 1/y where y is the number of inventors listed in a patent. Further, each of these bibliometric indicators is measured for female scientists and their male counterparts to map gender disparity in the development of pro-poor technologies.

Thus, this study focuses on article- and patent- level analysis to map involvement of women in the field, regardless of their level of contribution to the field. Therefore, authorship is not weighted based on the byline position and contribution of authors. The author- and inventor- level analysis (including both contribution and collaboration pattern analysis) is published in a complementary study (Ghiasi, Harsh, and Schiffauerova 2018).

This study uses independent t-tests to address gender differences in citation impact, journal ranking, and patent claims, for which the two samples are required to follow a normal distribution. Since the distribution of citation, SJR and patent claims are positively skewed, this study uses bootstrapping techniques (1000 replications) – a technique used to reduce the effect of outliers and other data problems by resampling and replacing observation – to calculate the 95% confidence interval (CI) of the means. Since this study focuses on the bootstrap distribution of the means, with 1000 resamples, the distribution of the mean of the resamples are approximately normal according to the Central Limit Theorem and therefore can be used in the statistical t-tests. More details on means and t-tests can be found in Appendix.

Table 1 shows the number of Canadian nanotechnology articles and patents with energy, agri-food, and water applications. Nanotechnology-related articles with energy applications represented only 3.7% of all the Canadian nanotech-related publications, which was the highest among the three pro-poor application areas of nanotechnology. Ontario had the highest share of nanotech-related publications in all the three applications followed by Quebec. As for patents, the shares of nanotech-related patents with energy and agri-food applications were higher (respectively, 5.8% and 2%) compared to their shares of publications (3.7% and 1.2% respectively). Ontario held the highest share of patenting, and British Columbia and Saskatchewan showed prominence in energy and agri-food patent applications, respectively. Given that the numbers of papers and patents were very small in Saskatchewan, the cross-province analysis in this research is confined to Ontario, Quebec, British Columbia and Alberta.

Table 1. Number and share of nanotech-related publications and patents with energy, agri-food, and water applications.

		Energy	Agri-food	Water
Number of articles		726	244	271
% of nanotech-related papers		3.7%	1.2%	1.4%
Share of authorship of Canadian provinces	Ontario	45.9%	38.2%	49.2%
	Quebec	23.0%	25.0%	27.9%
	British Columbia	12.5%	9.5%	4.4%
	Alberta	11.6%	11.7%	10.2%
	Saskatchewan	2.8%	9.6%	2.6%
	Others	4.2%	5.9%	5.7%
Number of patents		241	82	53
% of nanotech-related patents		5.8%	2.0%	1.3%
Share of inventorship of Canadian provinces	Ontario	50.3%	51.2%	62.7%
	Quebec	17.8%	6.3%	11.3%
	British Columbia	24.0%	5.5%	9.4%
	Alberta	5.7%	13.3%	9.0%
	Saskatchewan	0.1%	15.8%	5.7%
	Others	2.1%	7.9%	1.9%

Employment and wage analysis

This research uses author affiliation data and patent assignee data to identify nanotech-related companies to which at least one publication is affiliated or to which at least one patent in water, energy or agri-food applications is assigned (185 unique companies are found in our article and patent databases). These companies are further classified based on their primary North American Industry Classification System (NAICS) code using the Mergent Online database. These steps allow us to identify industries associated with nanotech-related companies in water, energy, and agri-food.

The identified industries are then examined with regard to the number of female employees and their hourly wages by using the Labour Force Survey estimates (LFS) table of Statistics Canada. Statistics Canada’s LFS table ⁶ provides the list of industries and the data on the number of men and women employed in Canada in those industries and their hourly wages since 1997. Employment and wage data is further assigned to the list of publications and patents published by those companies at the time of the publication or granted patent, ensuring the company was active at the time. The rate of employment and wage defined in this study, therefore, considers both the number and the year of publications (both papers and patents) produced by a company involved in nanotechnology pro-poor applications. Gender gap in employment is further defined as the difference in the share of men and women employed, while gender wage gap is defined as the difference between average hourly wage of men and women to that of men.

Publications

Universities held the highest share of publications across all the provinces. The University of Alberta and the University of Toronto were among the most active universities. Ontario had a high share of governmental publishing, which is due to the existence of the National Research Council (NRC) and Health Canada related laboratories in Ottawa, Ontario, the national capital (Figure 1(A)). The laboratories include the NRC Institute for Microstructural Sciences, the Steacie Institute for the Molecular Sciences, the NRC Institute for Biological Sciences, Health Canada’s Healthy Environments and Consumer Safety Branch and Health Canada’s Biologics and Genetic Therapies Directorate. Agri-food applications held the highest share of governmental involvement in scientific research, in which the NRC’s Plant Biotechnology Institute (in Saskatchewan) and Agriculture and Agri-Food Canada’s Saskatoon Research Centre played an important role (Figure 1(B)). The private sector and hospitals were involved in low rates of publishing across different provinces for pro-poor nanotechnology applications. However, AB Sciex LP and Medicago Inc. were among the top firms engaged in publishing.

Figure 1. (A) Share of different sectors in pro-poor nanotechnology authorship by province; (B) Share of different sectors by field. Abbreviations for provinces are: ON: Ontario; QC: Quebec; BC: British Columbia; AB: Alberta; Abbreviations for sectors are: U: University; G: Government; I: Industry; H: Hospital.

Nanotech-related publications with energy applications were the main focus across most of the dominant provinces. The number of publications in water and agri-food applications was similar but low across different provinces (Figure 2(A)). Figure 2(B) shows the average annual growth rate over the years 1996–2011 and reveals that the growth rate of the number of papers for all three pro-poor applications was higher than that of the overall nanotechnology papers.

Figure 2. (A) Number of publications by province in pro-poor nanotechnology; (B) Average annual growth rate of publications in pro-poor nanotechnology.

Gender disparities in publications

Authorship of pro-poor nanotechnology papers was largely male-dominated. Women held merely 18.6% of total authorship (Figure 4(C)) in all the pro-poor applications. This share was lower than the share of women authorship across all the sciences and all the engineering, which are 30% (Larivière et al. 2013) and 20% (Ghiasi, Larivière, and Sugimoto 2015), respectively. Figure 3 shows that despite the growth in the number of articles and authors, the share of female authorship did not change noticeably over the 15-year period.

Figure 3. Number of articles and authors (left axis) and share of female authorship (right axis) over the years in pro-poor nanotechnology. Abbreviations for gender are: F: Female, M: Male.

Narrowing the focus to disparities across different sectors (Figure 4(A)), the share of women authorship was higher in universities and governmental agencies and was the lowest in industry. Women affiliated with governmental agencies published in similar (no significant difference) ranked journals, and their publications received a similar (no significant difference) number of citations compared to their male counterparts. Women in academia, despite the higher overall citation and journal impact of their papers compared to other sectors, were involved in papers that had significantly lower citation impact and were published in significantly lower-ranked journals in comparison to their male peers in academia. It might seem that authorship in industry provides an equal context for women in the sense that, on average, they published in only similar ranked journals, and their publications received similar citations (no significant difference). However, it should be noted that the share of industry in pro-poor nanotechnology authorship is very small (∼3%) and involves only 111 authors with only 19 distinct female authors. Therefore, the findings on citation and journal impact for women in industry might be due to a selection effect, in the sense that women authors in a male-dominated sector are required to be extremely competent and recognized in the field, or they leave the sector or the field altogether (Ghiasi, Larivière, and Sugimoto 2015).

Figure 4. Share of female authorship (left axis) and citation and journal impact (right axis) of the papers published by female and male authors (A) by sector (B) by application area and (C) by province (***, **, * show significance of gender differences at the 1%, 5% and 10% levels: details can be found in the appendix (Tables A1–A3)).

Nanotechnology authorship with energy applications was the most male-dominated application where women published articles in significantly lower-ranked journals (Figure 4(B)). This may be due to the fact that research in energy occurs largely within the highly male-dominated fields of engineering and physics, whereas agri-food and water applications are associated with less male-dominated fields, namely earth sciences, biology, biotechnology, chemistry and health sciences, according to the work of Larivière (2014), who mapped the level of male-dominancy in scientific authorship of different disciplines in Quebec, Canada and worldwide. Authorship in agri-food and water applications was less male-dominated, and no significant differences were found between citation and journal impact of papers published by authors of each gender.

Compared to other provinces, Quebec showed more gender equity. Women in Quebec had a higher share of authorship than the authorship share for all Canadian women. This is in line with the analysis of Larivière (2014) across all scientific fields. Moreover, Quebec women were involved in papers published in the similarly ranked journals that also received a similar citation rate than their male peers from Quebec (no significant difference) (Figure 4(C)).

In general, female authors published in journals with similar ranking as that of men, while their authorship received significantly lower rate of citations (Figure 4(C)). This finding could correspond to the ‘Matilda effect in science’ (Rossiter 1993) at the citation level, where the number of citations of papers published by women is lower than the number of citations expected to be received by papers published in a given journal (Ghiasi, Larivière, and Sugimoto 2015; Larivière 2014) (Ghiasi, Larivière, and Sugimoto 2015; Larivière 2014).

Patents

As one might expect, industry holds the lion share of patent inventorship (73%), followed by universities (16%) and governmental agencies (9%). In terms of provinces, Ontario held the highest share of patent inventorship followed by British Columbia and Quebec (Table 1). British Columbia had the highest share of industry patenting, and universities played a major role in Quebec and Alberta (Figure 5(A)).

Figure 5. (A) Share of different sectors in pro-poor nanotechnology inventorship by province; (B) Share of different sectors by field.

Government played a large part in the development of nanotechnology patents in agri-food applications compared to the other applications (Figure 5(B)). Top governmental agencies involved in pro-poor nanotechnology patenting were the NRC, the Ministry of Natural Resources, and top companies are the Xerox Corporation, Nortel Network limited and Zenon Environmental Inc. Top universities were the University of Saskatchewan and Laval University.

Similar to the publication analysis, the volume of patents with energy applications was the highest among the dominant provinces, and the inventorship in agri-food and water applications were similar (Figure 6(A)). The annual growth rate of energy patents was higher than the overall nanotechnology patent growth rate (Figure 6(B)). Water had the lowest growth rate, which could imply that technological advancements with water applications were undeveloped in Canada.

Figure 6. (A) Number of patents by province in pro-poor nanotechnology; (B) Average annual growth rate of patents in pro-poor nanotechnology.

Gender disparities in patents

Similar to the authorship analysis, despite the growth in the number of patents and inventors, the share of female inventorship only increased slightly over time (Figure 7). Women were involved in only 11.3% of inventorship in pro-poor applications of nanotechnology (Figure 8(C)). This share was slightly higher than the findings of Sugimoto et al. (2015) who analyzed all technological fields worldwide and found a proportion of 10.3% in 2013. It is also higher than the findings of Mauleón, Daraio, and Bordons (2014) who focused on Spain and found that women were involved in only 9% of patents. This could be due to the interdisciplinary nature of the nanotechnologies, based on which Meng and Shapira (2011) justified the narrower gender gap they found in nanotechnology patenting in the US.

Figure 7. Number of patents and inventors (left axis) and share of female inventorship (right axis) over the years in pro-poor nanotechnology.

Figure 8. Share of female inventorship (as a percentage) and citation impact (as a decimal) (left axis) and average claims (right axis) of the patents granted to female and male inventors (A) by sector (B) by application area and (C) by province (***, **, * show significance of gender differences at the 1%, 5% and 10% levels: details can be found in the appendix (Tables A4-A6)).

When focusing on women’s participation in patenting across different sectors, the findings reveal that women’s share of inventorship was slightly higher in industry compared to universities and governmental institutions. Women were involved in industry patents with the same citation impact as their male counterparts, however, at universities and government institutions, there was a significant gender difference in the citation impact (Figure 8(A)). These gender differences in citation impact and number of claims were significant in the government sector (Figure 8(A)).

Contrary to female authorship, the share of women inventorship was slightly higher in the energy and agri-food sectors (12%) than in the water (7.5%) (Figure 8(B)). Women contributed to Canadian patents with a significantly lower number of claims, while their patents received similar citation rates as those of their male peers (no significant difference) (Figure 8(C)). However, the gender difference in citation impact of patents in agri-food application area is significant, while this difference is not significant for the number of claims (Figure 8(B)). The share of women inventorship was highest in Alberta, and female inventors contributed to patents with the same citation impact and claims as their male peers in Quebec (Figure 8(C)). This also corresponds to the publication analysis where papers published by women in Quebec have shown to have the same impact as the papers by their male counterparts.

This study also provides a cross-gender analysis for the production and research impact of Author-Inventors (A-Is). Author-inventors are here Canadian-affiliated individuals who published at least one paper and registered at least one patent in subfields of water, agri-food, and energy. Women represented 9.5% and 2.5% of A-I’s total authorship and inventorship, respectively. They published in similarly ranked journals (this difference is weakly significantly higher) and articles receiving similar rate of citations. They were also involved in patents with similar number of claims and with similar number of citations than men. Therefore, no significant differences were found in the scientific and technological impact of A-Is of each gender. This might also be associated with a selection effect, according to which women are required to be highly qualified and competent to become A-Is. Therefore, their scientific output is of the same impact.

Figure 9. (A) Share of female A-Is authorship (left axis) and citation and journal impact of papers published by female and male A-Is (right axis); (B) Share of female A-Is inventorship (as a percentage) and citation impact (as a decimal) (left axis) and average claims (right axis) of the patents granted to female and male A-Is (***, **, * show significance of gender differences at the 1%, 5% and 10% levels: details can be found in the appendix (Table A7)).

Employment and wage

Industries that are comprised of at least one company contributing to the scientific and technological development of pro-poor nanotechnology applications were identified. 52% of the companies were focused primarily on manufacturing, followed by professional and scientific services (24%) (Figure 10(A)).

Figure 10. (A) Share of companies contributed to the scientific and technological development of pro-poor nanotechnology applications by industry; (B) Gender gap in employment and wage by different pro-poor applications.

All three application areas were shown to be largely male-dominated in terms of both wages and employment. However, the agri-food industry showed a lower gender gap in both employment and wages in comparison to energy and water applications. Energy and water subfields were almost equally male-dominated in terms of both employment and wages (Figure 10(B)). As discussed above, nanotechnology authorship and inventorship in agri-food applications were also less male-dominated. Together with the employment analysis, this implies that women in this application area were subject to less inequality in comparison to the other two subfields.

Acknowledging that defining benchmarks shapes the interpretation of data, it is useful to compare the gender gap in companies pro-poor nano R&D to multiple benchmarks. Research and development in nanotechnology is a distinct category in Statistics Canada data (code: 541713), which is a subcategory of a more comprehensive industry titled ‘Professional, scientific and technical services’ (code: 54). This broader category approximates the STEM workforce in Canada, and here the gender gap in employment was 3.6% and in wage was 26% (Figure 11). When using this as a benchmark, companies involved in pro-poor applications (regardless of the application) were located in industries with distinctively higher gender gap in employment and slightly lower gender gap in wage. However, when considering all the industries in Canada as a benchmark (where the gender gap in employment and wage is 1.9% and 17%, respectively (Figure 11)), these companies were located in highly male-dominated industries in terms of both employment and wage. However, in terms of percentages, it should be noted that publishing and patenting activities in all the pro-poor applications were even more male-dominated than what women face in terms of wage gaps and employment gaps in workplaces in these areas.

Figure 11. Gender gap in employment (left) and wage (right) in Canadian industries.

Conclusions

This study analyzes the relationship between Canada’s progress in pro-poor nanotechnology R&D and gender inequality in the scientific workforce during a key recent historical period for nanotechnology development. The findings reveal that only a narrow spectrum of Canadian nanotechnology articles and patents reflected pro-poor priorities for energy, agri-food and water applications, which is significant because these are the three application areas of nanotechnology that have been identified as having the most potential to create pro-poor technologies. However, this study found a higher share of nanotechnology publications in these three pro-poor applications for Canada than the shares found by Cozzens et al. (2013) in their global analysis. This shows that Canada had a larger focus on the development of pro-poor nanotechnologies than the worldwide average, even if only slightly so.

Ontario had the largest share of governmental publishing and patenting activities, which is due to the fact that many of the governmental agencies – including NRC related agencies – are located in Ottawa, the capital. Among the three application areas, Canada put more focus on publishing and patenting on energy applications, which is in line with the worldwide analysis of pro-poor nanotechnology applications (Cozzens et al. 2013). The growth rate of publications in these applications was higher than the development of all Canadian nanotechnology papers, which implies there was an increasing focus on publishing on pro-poor technologies during this period. However, in terms of patents, the growth rate of nanotechnology patents with water applications was very slow and was lower than that of all nanotechnology patents, which was contrary to the case of energy and agri-food applications.

Canadian women researchers accounted for only 18.6% of total nanotechnology authorship in the three application areas, and industry was found to be more male-dominated compared to government and academia in terms of publications. Papers authored by women received a lower rate of citations while being published in similarly ranked journals as their peers. This is often associated to the ‘Matilda effect in science’ (Rossiter 1993) at the citation level, where the number of citations of papers published by women is lower than the number of citations expected to be received by papers published in a given journal (Ghiasi, Larivière, and Sugimoto 2015; Larivière 2014). Women in the government and private sector contribute to papers with similar impact as their male peers. Authorship in agri-food applications is less male-dominated in comparison to the other applications. Quebec showed the least amount of gender inequality compared to other provinces in terms of authorship and scientific impact of their work.

The patent analysis in this study reveals that the share of Canadian women inventorship was only 11.3% and the share of women inventorship across different sectors was quite similar. Patent citation impact reveals that inventions by women were equally important as those created by men in Canada and across all the major provinces. Inventorship in water applications was the most male-dominated. Female A-Is contributed to publications and patents with similar scientific impact as male A-Is, which might relate to a selection effect: unless being extremely competent, women do not get involved in both scientific and inventive scientific activities. In terms of workforce analysis, all the companies involved in R&D for all three application areas are active in highly male-dominated industries in which women are subject to lower rates of employment and lower hourly wages than average.

These findings have implications for method, theory and policy. One main contribution of this study is methodological in that it combines the use of bibliometric analyses of articles and patents with workforce analysis of industries, proposing a novel approach that could be applied to better understand the socio-economic dimensions of any technology that is in the development stage. While this study uses the Mergent Online database to identify NAICS codes of companies, other databases (such as LexisNexis Academics) could be used and other classification codes could be identified, such as the Standard Industrial Classification, or International Standard Industrial Classification. And while this study focused on gender, public socio-economic data at the time of publication or patent (or for a specific period) could be further utilized to better analyze other labor market outcomes. For example, public data on employment and wage across various types of work, ethnicity, age, and regions are provided by the Canadian and the US governments for each class of industry, in the Labor Force Survey and Current Employment Statistics, respectively. This could provide several avenues for analyzing technology impact on employment for people in different age groups or in terms of ethnicity, economic class, province of residence, type of job, etc..

The paper also contributes conceptually and methodologically to the field of RI. Given that core goals of prominent RI frameworks revolve around gender (European Union 2012) and inclusion (Stilgoe, Owen, and Macnaghten 2013), it is important to bring these two conceptualizations of equity together more explicitly. This paper contributes to the field in that it provides an approach to conceptualize a cross-dimensional analysis of emerging technologies and equity. Our approach allows for analysis of distributional outcomes of nanotechnology in both an economic dimension (pro-poor) and a social dimension (gender). Furthermore, while early efforts to bring about RI were often discussed as national endeavors, this paper contributes to more recent efforts to take into account consideration of its transnational and global connections (Doezema et al. 2019; Ludwig and Macnaghten 2020), as efforts to generate pro-poor technologies that might create benefit for poor communities in one country, could exacerbate inequalities in the STEM workforce in another country.

This brings us to Canadian policy. During the period we examine, Canada was not only investing a great deal in nanotechnology R&D, it was also committed to using science and technology as tools for international assistance. In 2004, the government of Canada devoted 5% of its R&D budget to address the challenges faced by developing countries (Salamanca-Buentello et al. 2005). In 2006, the collaboration between the Association of Universities and Colleges of Canada and Canada’s International Development Research Centre was formed to support scientific and technological advancements that bring benefits to developing countries and help solve development problems (AUCC 2006). Despite all these efforts, this study finds that only a narrow spectrum of nanotechnology research and development efforts in Canada are in pro-poor areas that might contribute to the needs of developing countries. This shows that while Canada was heavily investing in what, at the time, was one of the ‘latest’ emerging technologies, and was creating policy efforts to try to use R&D as tools for international assistance, the two efforts appear to be disconnected.

However, the study shows that the government faces added difficulty in producing more pro-poor R&D. Because the pro-poor scientific and innovative efforts we examined tend to be highly male-dominated in terms of both the scientific community and the workforce, the incentives to develop pro-poor nanotechnology applications might widen the gender gap and hinder social development. If the growth of interest in pro-poor nano-innovation could add to gender biases in the scientific workforce, international development efforts could have the unintended side effect of negatively impacting the social development of the nation. Any efforts Canada made to promote gender equity in STEM during the period we examined were not nearly as prominent or effective as the efforts to promote nano R&D.

Overall, our results imply that during the 2000s, there seemed to be a lack of coordination and connection between policies and mechanisms that aim to promote social (gender) and economic (pro-poor) dimensions of what is now understood as RI. One possible reason for this could be the siloed nature of Canada’s government, which is modeled after the Westminster system, where collaboration between departments and ministries is bureaucratically difficult, which could have stymied coordination between departments coordinating international assistance and science and technology policy. We also examined R&D in academia and industry as well as government, and while there was a commitment to spending 5% of R&D to address challenges, it is not clear how that was coordinated between the three sectors. Finally, during the 2000s, Canada also had three Prime Ministers and saw a shift in power from the Liberal to the Conservative party, which surely also made working across ministries and departments difficult. Policy coordination is necessary to make governance of any new technology more equitable in social and economic dimensions, but this study shows that these dimensions cannot be dealt with in isolation. Therefore, this study informs policymakers on the importance of mainstreaming cross-cutting concerns and brings attention to both gender and pro-poor concerns simultaneously.

Canada’s effort to mainstream gender in S&T policies is largely confined to NSERC initiatives and partial investments in programs such as PromoScience and CREATE (none of these programs has a mere focus on gender initiatives). Although NSERC shows policy concerns for underrepresentation and repression of women in sciences, its gender-related initiatives seem not to be very effective. According to Shendruk (2015), the share of women working in the STEM fields barely changed since 1987, and they are still underpaid (7.5% less than their male peers). Moreover, NSERC’s policies largely focus on participation rather than retention or progression of women in the science and engineering fields.

On the other hand, Canada’s R&D efforts in nanotechnology is long established and prestigious, which are credited with founding the Centre for Advanced Nanotechnology at the University of Toronto in 1997, and National Research Council (NRC) institutes in Alberta, British Colombia, Quebec, and Ontario, introducing large institutions such as the National Institute for Nanotechnology (NINT) and NanoQuébec (established in 2001). Despite all these efforts, initiatives to mainstream gender in nanotechnology R&D is nonexistent at national, institutional, and organizational levels.

With the right gender-responsive policies that help attract and support women in these nano-application areas, Canada might be able to reach a higher level of gender equality in its workforce and broaden its knowledge capacity and research performance in addressing developing world’s challenges. The results of this study are thus of great importance to policymakers to gain insight into the identification of leverage points to promote both gender equality and poverty alleviation in emerging science and technology policies, enhance success in new interdisciplinary environments (such as nanotechnology) and consequently promote a more equitable and inclusive society.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes on contributions

Gita Ghiasi is a Research Officer at the École de bibliothéconomie et des sciences de l’information (Université de Montréal) and the Institute for Data Valorisation (IVADO). She received her Ph.D. in Industrial Engineering from Concordia University and holds a Master’s degree in International Business from the University of Warsaw. Her research focuses on new technologies and the challenges of equity, equality, and development.

Matthew Harsh is an Associate Professor of Science Technology and Society and Director of the Center for Expressive Technologies at California Polytechnic State University, San Luis Obispo, USA. His work focuses on the structure and governance of science and innovation in Africa, especially related to new and emerging technologies.

Andrea Schiffauerova is an Associate Professor at Concordia Institute for Information Systems Engineering at Concordia University, Canada. Her main area of expertise involves economics f innovation and science, and management of knowledge and technology, with a particular interest in innovation networks and knowledge diffusion.

Notes

1 Nanotechnology is ‘the understanding and control of matter at dimensions of roughly 1–100 nanometers, the size-scale between individual atoms and bulk materials, where unique phenomena enable novel applications’ (NNI 2007).

2 Social cohesion refers to all efforts which ‘ensure that every citizen, every individual, can have within their community the opportunity of access: to the means to secure their basic needs, to progress, to protection and legal rights, and to dignity and self-confidence’ (Council of Europe 2001, 5).

3 Social inclusion is ‘people’s ability to participate adequately in society, including education, employment, and public services, social and recreational activities’ (Litman 2003).

4 The part of the analysis complements our previous study (Ghiasi, Harsh, and Schiffauerova 2018), which found gendered collaboration patterns among authors and inventors involved in pro-poor applications of nanotechnology and called for gender-related policies to support inclusion of women in scientific collaboration teams.

5 http://www.genderchecker.com

6 Table: 14-10-0064-01 (formerly CANSIM 282-0072) Employee wages by industry, annual (x 1,000); DOI: https://doi.org/10.25318/1410006401-eng

Appendix

Table A1. 95% Cls for average citation and journal impact of the papers authored by each gender and significance tests for their differences across different sectors (***, **, * show significance of gender differences at the 1%, 5% and 10% levels).

Sectors	Measure	Mean (CI) – Male	Mean (CI) – Female	Mean Difference (Male-Female)	Sig. (two-tailed)
University	Citation	0.92 (0.873–0.975)	0.808 (0.728–0.896)	0.112	0.026**
	SJR	0.575 (0.547–0.606)	0.52 (0.48–0.565)	0.055	0.037**
Government	Citation	0.635 (0.527–0.758)	0.664 (0.52–0.823)	−0.028	0.783
	SJR	0.383 (0.341–0.431)	0.401 (0.33–0.479)	−0.018	0.744
Industry	Citation	0.653 (0.502–0.831)	0.735 (0.378–1.176)	−0.081	0.712
	SJR	0.413 (0.338–0.509)	0.338 (0.233–0.448)	0.076	0.297

Table A2. 95% Cls for average citation and journal impact of the papers authored by each gender and significance tests for their differences across different application areas (***, **, * show significance of gender differences at the 1%, 5% and 10% levels).

Areas	Measure	Mean (CI) -Male	Mean (CI)-Female	Mean Difference (Male-Female)	Sig. (two-tailed)
Energy	Citation	0.811 (0.763–0.862)	0.75 (0.665–0.844)	0.061	0.228
	SJR	0.539 (0.506–0.574)	0.468 (0.468–0.52)	0.071	0.019**
Agri-food	Citation	0.91 (0.783–1.049)	0.74 (0.603–0.899)	0.17	0.078*
	SJR	0.487 (0.429–0.553)	0.466 (0.402–0.54)	0.021	0.654
Water	Citation	0.984 (0.874–1.114)	0.887 (0.709–1.075)	0.097	0.374
	SJR	0.575 (0.53–0.623)	0.58 (0.501–0.656)	-0.005	0.903

Table A3. 95% Cls for average citation and journal impact of the papers authored by each gender and significance tests for their differences in Canada and across different provinces (***, **, * show significance of gender differences at the 1%, 5% and 10% levels).

Provinces	Measure	Mean (CI) – Male	Mean (CI) – Female	Mean Difference (Male-Female)	Sig. (two-tailed)
Canada	Citation	0.862 (0.816–0.904)	0.78 (0.711–0.855)	0.082	0.044**
	SJR	0.537 (0.512–0.562)	0.494 (0.458–0.534)	0.043	0.055*
ON	Citation	0.85 (0.791–0.91)	0.809 (0.695–0.927)	0.041	0.553
	SJR	0.492 (0.433–0.558)	0.555 (0.521–0.59)	−0.063	0.084*
QC	Citation	0.833 (0.74–0.934)	0.836 (0.704–0.98)	−0.003	0.976
	SJR	0.511 (0.467–0.559)	0.531 (0.467–0.598)	−0.02	0.624
BC	Citation	1.011 (0.865–1.168)	0.759 (0.52–1.044)	0.252	0.093*
	SJR	0.595 (0.514–0.68)	0.515 (0.388–0.668)	0.08	0.31
AB	Citation	0.899 (0.765–1.052)	0.743 (0.536–1.043)	0.156	0.31
	SJR	0.496 (0.406–0.595)	0.497 (0.403–0.604)	−0.001	0.99

Table A4. 95% Cls for average citation impact and claims of the patents of each gender and significance tests for their differences across different sectors (***, **, * show significance of gender differences at the 1%, 5% and 10% levels).

Sectors	Measure	Mean (CI) – Male	Mean (CI) – Female	Mean Difference (Male-Female)	Sig. (two-tailed)
Industry	Citation	0.076 (0.065–0.088)	0.07 (0.054–0.091)	0.006	0.612
	Claim	9.194 (8.526–9.932)	6.976 (5.726–8.787)	2.218	0.018**
University	Citation	0.091 (0.059–0.136)	0.02 (0.007–0.039)	0.071	0.028**
	Claim	10.673 (9.108–12.287)	7.979 (4.887–11.384)	2.694	0.137
Government	Citation	0.041 (0.031–0.052)	0.012 (0.002–0.025)	0.028	0.001***
	Claim	7.234 (5.883–8.757)	3.79 (2.57–5.073)	3.445	0.002***

Table A5. 95% Cls for average citation impact and claims of the patents of each gender and significance tests for their differences across different application areas (***, **, * show significance of gender differences at the 1%, 5% and 10% levels).

Areas	Measure	Mean (CI) – Male	Mean (CI) – Female	Mean Difference (Male-Female)	Sig. (two-tailed)
Energy	Citation	0.088 (0.074–0.104)	0.083 (0.055–0.119)	0.005	0.798
	Claim	10.465 (9.656–11.406)	8.265 (6.657–10.295)	2.2	0.039**
Agri-food	Citation	0.045 (0.033–0.059)	0.024 (0.011–0.042)	0.021	0.045**
	Claim	6.061 (5.22–7.005)	4.656 (2.966–6.847)	1.405	0.186
Water	Citation	0.061 (0.048–0.077)	0.049 (0.025–0.076)	0.012	0.396
	Claim	8.856 (7.536–10.257)	4.858 (3.689–6.069)	3.998	0.001***

Table A6. 95% Cls for average citation impact and claims of the patents of each gender and significance tests for their differences in Canada and across different provinces (***, **, * show significance of gender differences at the 1%, 5% and 10% levels).

Provinces	Measure	Mean (CI) -Male	Mean (CI)-Female	Mean Difference (Male-Female)	Sig. (two-tailed)
Canada	Citation	0.075 (0.066–0.087)	0.065 (0.047–0.091)	0.01	0.435
	Claim	9.296 (8.674–9.934)	7.002 (5.832–8.428)	2.294	0.004***
ON	Citation	0.08 (0.065–0.096)	0.07 (0.043–0.106)	0.01	0.635
	Claim	9.19 (8.394–10.044)	7.077 (5.826–8.39)	2.113	0.012**
QC	Citation	0.05 (0.04–0.061)	0.052 (0.033–0.077)	−0.002	0.872
	Claim	7.242 (5.894–8.784)	5.559 (3.204–8.368)	1.683	0.265
BC	Citation	0.106 (0.078–0.146)	0.079 (0.038–0.121)	0.027	0.374
	Claim	11.88 (10.26–13.61)	10.21 (4.42–18.42)	1.676	0.708
AB	Citation	0.049 (0.029–0.073)	0.054 (0–0.167)	−0.005	0.913
	Claim	9.353 (7.289–11.692)	4.114 (2.251–6.083)	5.239	0.007***

Table A7. 95% Cls for average citation and journal impact of papers and average citation impact and claims of patents of A-Is of each gender and significance tests for their differences (***, **, * show significance of gender differences at the 1%, 5% and 10% levels).

A-I Role	Measure	Mean (CI) -Male	Mean (CI)-Female	Mean Difference (Male-Female)	Sig. (two-tailed)
Author	Citation	0.6 (0.464–0.737)	0.71 (0.564–0.91)	−0.11	0.323
	SJR	0.505 (0.398–0.617)	0.686 (0.503–0.854)	−0.181	0.086*
Inventor	Citation	0.095 (0.055–0.144)	0.133 (0.1–0.167)	−0.038	0.271
	Claim	8.747 (6.96–10.802)	7.917 (3.1–3.833)	0.83	0.835