“Construction is different” and the sites are a risk environment with a limited access. There is a lack of exposition assessment, making no possible sound conclusions (Swuste, Frijters, & Guldenmund, 2012). Studies at task level represent a very low percentage of all research, it is necessary pay more attention to safety conditions on site (Zhou, Goh, & Li, 2015).
Spanish construction sector has the highest incidence rate of accidents (EUROSTAT, 2014) and their study is one of the main research line (Cambraia, Saurin, & Formoso, 2010). Nowadays accident rates, as lagging indicators, are proposed to renovate by leading ones (Hinze, Thurman, & Wehle, 2013). From compulsory accidents notification, ESAW (European Statistics on Accidents at Work) reporting system (European Commission, 2002), researches are able to manage variables related to worker, business, temporal, material and geographic variables (Camino López, Fontaneda, González Alcántara, & Ritzel, 2011; Camino López, Ritzel, Fontaneda González, & González Alcántara, 2011; Camino López, Ritzel, Fontaneda, & González Alcantara, 2008; López Arquillos et al., 2012; Pérez-Alonso, Carreño-Ortega, Callejón-Ferre, & Vázquez-Cabrera, 2011). Following these authors, issues as the age of the injured person, company staff, accident day, absence, deviation from accepted practice, injury and location variables are included in contingency tables to test the relationship between them and to accident severity. Other actual lines deeper analysed accident rates and different processes of construction or worker’s characteristics (López-Arquillos, Rubio-Romero, & Gibb, 2015; Rubio-Romero, Suárez-Cebador, & Abad, 2014). The source of information, coming directly from firms, limit the quality of the data and the subsequent analysis based on them (Salguero-Caparros, Suarez-Cebador, & Rubio-Romero, 2015).
Other research line connect site conditions with likely of accidents (Hinze et al., 2013; Memarian & Mitropoulos, 2013), but it is a limited source of research to low studies. There is another type of information based on surveys as well to big scale, European, national, or reduced to specific studies, but again we do not obtain direct information of site (INSHT, 2015; Eurofounf, 2015)
Limited research lines connect some organizational factors on site and risk conditions assessed on site (Fang, Huang, & Hinze, 2004; Forteza, Carretero-Gómez, & Sesé, 2017) and fewer have linked safety conditions on site with site typologies.
Previous research connect some project characteristics with certain safety performance elements. Table 1 shows the main researches that connect some site characteristics that define different site typologies attending safety results.
Taking in account previous studies, our work analyses a sample of sites assessing site levels of risk using a new toll named CONSRAT (Forteza, Sesé, & Carretero-Gómez, 2016) and classify attending main site typologies coming from leading elements that compose these typologies. In doing so we have used to the main site typologies used in the sector to refer and classify construction sites (Ministerio de Fomento)
A quantitative study of the classification of site typologies attending risk levels of sites is conducted in this study.
Values of risk level rank from 0.00 (no risk) to 1.00 (high risk) following CONSRAT methodology, as we mentioned before. As we can see there is general level of risk fairly high in all categories. Despite that, we can observe differences levels of risk in some issues. We omit categories of classification without differences in the level of risk.
- Data and sample
Our sample is representative from building construction sites in Spain. Using CONSRAT (Forteza et al., 2016) we assessed the levels of site risk and site characteristics on visiting a total of 957 sites. Data were collected from 2003 to 2009. All sites were selected following a random criteria. In doing so, we used together two different notification forms to Labour Authority coming one from promoter and another form contractor. We selected approximately 4 per cent of sites coming from these notifications. The sampling error of our sample was approximately 2.7% with a confidence interval of 95%.
Table 2 shows a main classification of sites of our sample compared with the total number of sites of Balearic Islands and Spain available from Spanish Ministry (Ministerio de Fomento). Our sample has a similar distribution of sites typologies comparing with total sites coming from official data. On the one hand our sample has a total of 43.9% of single family houses (including insolated and infilled) while total populations achieves to a 65.95% in Balearic Islands and 77.65% in Spain. On the other hand our sample has more level of multi-family, 44.6%, than the total population with a 26.60% and 15.74% respectively in local and national environment. Finally services and other uses buildings represent a total of 11.05% in our sample and just a 7.44% or 6.6% in local and state reference.
Table 3 shows the comparative of sample’s promotor typology and total of the sector in order to situate our sample sites.
- Analysis design
Coming from literature review, we rank each element selected attending risk level assessed on site. Finally we compose the main types of sites joining different compatible elements.
- Main descriptive statistics
In addition to the first possibility of site classification in agreement with table 1, and in order to understand the main characteristics of our sample, next we introduce the principal features of elements that composes the sites. In doing that we follow CONSRAT methodology (Forteza et al., 2016), showing those characteristics more descriptive of sample.
- Site typology and stage of work.
Our sample is characterised by small new construction sites as 89% of sites are new construction and 61% are sites with a building composed by grand floor and one or two more floors, follow by sites from 3 to 5 floors (34%). Most of sites are single family houses, 44%, followed by multi-family, 34%.
Regarding site stage, main stages are structure and foundation (34%) and brickwork (28%), most of them just with one simple work in progress (64%), and with workers situated on perimeters of floors or roofs (60%).
- Promoter and contractor characterization
The promoter or our sample is characterised by professional one (57%), that means they are companies that promote to sell the buildings. On most of sites (81%) there is no notice of designation of health and safety (H&S) coordinator, which is a duty of promoter based on CE Directives. There is no documented work of H&S coordinator (95%) and the promoter contract just one contractor in most of the visited sites (83%).
Main constructors are companies (86%), followed by self-employers with or without workers (14%). The 92% of sites present from one1 to 5 companies, while there is subcontracting at the 55% of sites. Most of construction sites (54%) have from one to five workers folowed by sites from 6 to 15. Site management structure on site present a 26% of sites with nobody in charge, 26% of sites with a current worker, 25% of sites with foreman and 25% with business owner at the head (see Fig. 1 for further details). Attending preventive functions of site structure, nobody assumes these functions in the 58% of sites while they are assumed on secondary way in the 37% of sites.
.jpg)
Fig. 1. Site management structure.
Concerning to specific H&S issues, there is not H&S plan on 63% of sites and the appropriateness of its previsions are unknown or not applicable on site about the 88%. The compliance with H&S plan is deficient or very deficient in the 96% of sites.
- Level of risk attending the different site characteristics
- Level of risk based on Site typology.
.jpg)
Fig. 2. Number of floors
To assess risk based on site typology CONSRAT uses basically site size, measured by number of floors, building configuration and type of site. We can observe that the level of risk decreases by size, except the most small sites with just one floor, that have the lowest level of risk (see Fig. 2). Regarding building configuration, we did find differences between single family houses (Isolated and infill, around 0.80) and infill multifamily (0.75). We did not find differences of risk among others issues.
- Level of risk based on Promoter characterization.
CONSRAT promoter characterization takes in account the type of promoter, designation of health and safety coordinator, her/his documented work, type of contracting and environment conditions. Figures 3 to 5 show the characteristics with differences of risk among them. As we can see, the level of risk on site increases when we find a private-individual promoter, when there is no H&S coordinator or his/her work is not documented. When promoter decides to contract more than one contractor the level of risk on site decreases (See Fig. 6).
.jpg)
Fig. 3. Level of risk attending type of promoter
.jpg)
Fig. 4. Level of risk attending designation of health
.jpg)
Fig. 5. Level of risk attending health and safety coordinator documented work
.jpg)
Fig. 6. Level of risk attending Type of contracting
- Level of risk depending of contractor characterization.
Following with contractor characteristic, CONSRAT analyses this issue in three aspects: type of firm, concurrence and H&S plan. Type of firm includes role, structure and resources on site and functions. Concurrence includes number of firms, subcontracting and number of workers. H&S plan includes presence, adequateness and compliance.
Starting by type of firm, sites with a company at head have lower level of risk than the others with a self-employed with or without workers (See Fig. 7). Figure 8 shows the level of risk by management structure on site, taking in account personal resources of contractor on site. As we can observe, the level of risk increases from more to less resources. In general, more assumption of preventive functions means lower levels of
risk on site, despite we find high levels of risk on site when there are the top of functions (See Fig. 9). Although the differences are small, we have find that when subcontractor or promoter-contractor are at heat of site, there is more level of risk than when there is the contractor (From 0.815 to 0.805 respectively).
.jpg)
Fig. 7. Level of risk attending type of constructor at head
.jpg)
Fig. 8. Level of risk attending site management structure
.jpg)
Fig. 9. Level of risk attending preventive functions of structure
Following with concurrence on site, we can see than not always more companies means more risk. Figure 10 shows than site from 1 to 5 firms have similar levels of risk than sites with more than ten firms, while medium sites (from 5 to 10 sites) risk decrease. Subcontracting do not affect negatively the level of risk, as we can see at figure 11. Similar behaviour has the number of workers, as we have obtained lower levels of risk on site when there are more workers.
.jpg)
Fig. 10. Level of risk attending number of companies on-site
.jpg)
Fig. 11. Level of risk attending level of subcontracting
.jpg)
Fig. 12. Level of risk attending total number of workers on site
Continuing with the H&S plan, we can observe than the presence, compliance and appropriateness of this document decrease the level of risk (See Fig. 13, 14 and 15)
.jpg)
Fig. 13. Level of risk attending presence on site of H&S Plan
.jpg)
Fig. 14. Level of risk attending compliance with H&S Plan
.jpg)
Fig. 15. Level of risk attending appropriateness of the previsions of the H&S Plan
– Number of sites attending main characteristic classified by size.
Taking into account the results of risk level obtained in our field work, below is shown the different composition of the sample of sites using as first classification the size, evaluated by number of floors, and as secondary all the other characteristics.
Figure 16 shows the distribution of firms attending type of promoter according number of floors. As we can see there are more proportion on private promotors on small sites with ground floor plus one or two floors (GF+1/2, 302) than in medium sites (GF+3/5), where most promoters are professional (282). Similar proportion of promoters are obtained if we increase the size (GF+5 and infrastructure sites), where private promoter decreases or disappear. Continuing with characteristics related with promoter and size we can see than in most of small sites there is no constancy of H&S coordinator (See Fig. 17, 503 sites in GF+1/2) while the proportion of H&S coordination designation increases when site is bigger.
.jpg)
Fig. 16. Number of firms attending size and type of promoter
.jpg)
Fig. 17. Number of firms attending size and designation of coordinator
Below is shown the distribution of constructor characteristics classified by size of site. Comparing again small and medium sites (GF+1/2 and GF+1/23/5) we can see that the most of sites do not have nobody in charge followed by sites with foreman, business owner and worker with some functions at head of site (See Fig. 18). For its part, most of medium sites present a distribution with lower proportion with nobody in charge and higher with business owner and foreman at head on site. When we increase the size, the proportion of sites with more resources increases too. The preventive function of constructor mentioned resources has a similar behaviour. As we can see, at figure 19 the proportion of sites without assumption of preventive functions is higher in small sites.
.jpg)
Fig. 18. Number of firms attending size and management structure
.jpg)
Fig. 19. Number of firms attending size and preventive functions
Focusing in the relation between size and specific resources in H&S, we can observe that most of small sites do not have the H&S plan on site (397, see Fig. 20), with more proportions than medium (190). On bigger sites the proportions is reversed and we observe more sites with H&S plan. Similar behaviour has the appropriateness and compliance of the mentioned H&S plan (See Fig. 21 and 22).
.jpg)
Fig. 20. Number of firms attending size and presence of H&S plan
.jpg)
Fig. 21. Number of firms attending size and appropriates of H&S plan
.jpg)
Fig. 22. Number of firms attending size and compliance with H&S plan
Our first results on site typology seem not agree with previous literature that connect more size with more accident (Fang et al., 2004; Liao and Perng,2008), but they do agree with other studies that connect accidents with small projects and low budgets (Cheng et al., 2010) or just with small size. We believe that size is just the context were develop other characteristics related with the level of organizational resources than have special connection with risk or accidents as previous research find (Fang et al., 2004; Forteza et al., 2017; Mohamed, 1999; Teo & Ling, 2006; Wu, Liu, Zhang, Skibniewski, & Wang, 2015). The rest of the levels of risk according promoter and contactor characterises are in this line. Then we are going to analyse the environment of size analysing the others issues.
Looking at Fig. 16, we can observe lower organizational resources on small sites than the biggest ones as most of the small sites has a private promoter. If we look at Fig. 3 we can see than private or individual promoter has the highest level of risk, probably because it is related with lower level of resources as shows Fig. 4 and 5, were risk is related with no designation of H&S coordination and no documented work of this technician. The influence on site of promoter are in line of previous studies than linked this figure with level of safety on site (Ros et al., 2013; Fang et al., 2004; Hinze et al., 2013; Xinyu et al., 2006; Wu et al., 2015), and H&S coordinator is a Owners’ appointment (H&SC, European Directive 92/57 CEE) that develops safety inspections who complements the authority of the safety supervisor and the foreman (Ros et al., 2013).
Our empirical research shows evidence about this relevance and, in addition, we can see than in most of small sites of our sample there is no constancy of H&S coordinator, while in bigger sites this proportion decrees or is reversed, that is in the line to confirm the relevance of type and promoter organizational resources.
Following with the analysis of the characteristic of sites by size, we can see at Fig.18 that smaller sites have more proportion of sites with nobody in charge or a simple worker with some functions, while, bigger sites are related with more management level of resources. As we can see at Fig. 8 level of risk is related with site management structure as lower levels of resources shows higher level of risk on site. Constructor is the figure responsible of provide resources, define processes, foreman and crew on site with are related with the level of risk (Forteza et al., 2017). Other factor related with constructor resources are showed on Fig. 7 were we can see the level of risk related with type of constructor at head on site. This is in the same line of previous research (Baxendale & Jones, 2000; Fang et al., 2004; Xinyu et al., 2006; Wu et al., 2015). Following with the structure of constructor and the level of preventive functions of the structure assumed (See Fig. 19), we can observe that, on most small sites, there are not assumption of responsibilities or they are assumed secondary. As we can see at Fig. 9 more assumption of preventive functions of structure is related with lower levels of risk, except the highest level of assumption that has the higher level of risk. But if we observe this maximum level of assumption at Fig. 19, it is not representative as is corresponding just with four sites.
Finally, analysing size and promoter resources by H&S plan characteristics we can see that most of small sites do not have H&S plan on site, it is unknown or not apply and its compliance is deficient or critical (See Fig. 20, 21 and 22). As we can see in previous results (See Fig. 12, 14 and 15) these three items related with H&S plan have a relationship with the level of risk, as no presence, no compliance or the not appropriateness of H&S plan increases risk. This is in the line of several previous theoretical studies (Fang, et al., 2004; Ros et al., 2013) and empirical research (Forteza et al., 2017).
The other results obtained of risk level are related with concurrence on site. First one shows the relationship between number of companies and level of risk (Seer Fig. 10). As we can see, there is not a clear relation between them as sites with less firms on site have similar level of risk, while there is a clear decline of risk level on sites from six to ten firms. Following, we analysed the other factors of concurrence, subcontracting and number of workers. Our results point that the subcontracting is not related with more level of risk on site (See Fig. 11), and more works has the same result, as it is not related with more risk (See Fig. 12). These results are not accordant with previous studies that pointed about subcontracting and the coexistence of diverse firms, with their own processes and roles, add complexity and increase the difficult to manage (Forteza et al., 2017). Some researches connect more concurrence with low levels of safety (Hinze et al., 2013; López-Alonso, Ibarrondo-Dávila, Rubio-Gámez, & Munoz, 2013; Manu, Ankrah, Proverbs, & Suresh, 2013; Swuste et al., 2012; Yung, 2009), as more number of workers and the number of subcontractors are related with accidents (Manu et al., 2010), increasing the complexity of contractual system and agents involved on site.
The aim of the present research was to test the level of risk on site based on different characteristics that may affect this risk level. With these data we have obtained the main characteristics on site that are related with the level of risk. We studied three groups of factors related with risk on site; Site typology, type of promoter and type of contacting, analysing in each one the main characteristics related with then based on CONSRAT methodology.
Our results suggest than most risky type of site is characterized by: in relation to site typology, small sites (single family houses from one to two floors); in relation to promoter characterization, with individual or private promoter, without notice of designation of H&S coordinator and his/her documented work. We have obtained higher levels on sites with just one contractor; in relation to contractor characterization, most risky sites have a self-employed with workers at head, nobody in charge, without assuming preventive functions, with more than ten firms on site or less than five, without subcontracting, less than five workers, no H&S plan on site, no compliance and finally with wrong previsions not according with conditions of site.
Finally, we want to conclude that our study reinforce the idea that risk on site is affected by a complex and long list of characteristics that influence by itself and crossing their effects. We can not study independently some of these characteristics as we have observed than they are intertwined. That means, for example, that in some sites can be more influential the level of resources from contactor or promoter than the level of intrinsic complexity (size, configuration). In other works intrinsic level of risk on site can be moderated or controlled by level of organisational resources. Normally we can not modify intrinsic complexity, just control it, but in our hand (all agents interviewing in construct ruction process) is improving the level of resources that has an important influence of risk.
- Limitations and future challenges
Present study provides empirical evidence of levels of risk on site by different characteristics and we consider that has an intrinsic value for this reason, taking in account the lack of empirical evidence of this kind of studies. Despite this, it may be appropriate try to analyse this data with a methodology of cluster analysis in order to study globally the influence of all characteristics.
This research was partially funded by La Caixa Foundation.
Cambraia, F. B., Saurin, T. A., & Formoso, C. T. (2010). Identification, analysis and dissemination of information on near misses: A case study in the construction industry. Safety Science, 48(1), 91–99. https://doi.org/10.1016/j.ssci.2009.06.006
Camino López, M. A., Fontaneda, I., González Alcántara, O. J., & Ritzel, D. O. (2011). The special severity of occupational accidents in the afternoon: ‘The lunch effect’. Accident Analysis & Prevention, 43(3), 1104–1116. https://doi.org/10.1016/j.aap.2010.12.019
Camino López, M. A., Ritzel, D. O., Fontaneda González, I., & González Alcántara, O. J. (2011). Occupational accidents with ladders in Spain: Risk factors. Journal of Safety Research, 42(5), 391–398. https://doi.org/10.1016/j.jsr.2011.08.003
Camino López, M. A., Ritzel, D. O., Fontaneda, I., & González Alcantara, O. J. (2008). Construction industry accidents in Spain. Journal of Safety Research, 39(5), 497–507. https://doi.org/10.1016/j.jsr.2008.07.006
Cheng, C.-W., Leu, S.-S., Lin, C.-C., & Fan, C. (2010). Characteristic analysis of occupational accidents at small construction enterprises. Safety Science, 48(6), 698–707. https://doi.org/10.1016/j.ssci.2010.02.001
European Commission. (2002). European Statistics on Accidents at Work (ESAW). Retrieved 29 September 2016, from http://ec.europa.eu/eurostat/ramon/statmanuals/files/ESAW_2001_EN.pdf
EUROSTAT. (2014). Standardised incidence rate of accidents at work by economic activity and employment status. Retrieved from http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=hsw_aw_inaws&lang=en
Fang, D. P., Huang, X. Y., & Hinze, J. (2004). Benchmarking Studies on Construction Safety Management in China. Journal of Construction Engineering & Management, 130(3), 424–432. https://doi.org/10.1061/(ASCE)0733-9364(2004)130:3(424)
Forteza, F. J., Carretero-Gómez, J. M., & Sesé, A. (2017). Effects of organizational complexity and resources on construction site risk. Journal of Safety Research, 62(Supplement C), 185–198. https://doi.org/10.1016/j.jsr.2017.06.015
Forteza, F. J., Sesé, A., & Carretero-Gómez, J. M. (2016). CONSRAT. Construction sites risk assessment tool. Safety Science, 89, 338–354. https://doi.org/10.1016/j.ssci.2016.07.012
Hinze, J., Thurman, S., & Wehle, A. (2013). Leading indicators of construction safety performance. Safety Science, 51(1), 23–28. https://doi.org/10.1016/j.ssci.2012.05.016
Instituto Nacional de Seguridad e Higiene en el Trabajo (INSHT). (2015). Encuesta nacional de condiciones de trabajo. Retrieved from http://www.insht.es/InshtWeb/Contenidos/Documentacion/FICHAS%20DE%20PUBLICACIONES/EN%20CATALOGO/GENERALIDAD/ENCT%202015.pdf
Liao, C.-W., & Perng, Y.-H. (2008). Data mining for occupational injuries in the Taiwan construction industry. Safety Science, 46(7), 1091–1102. https://doi.org/10.1016/j.ssci.2007.04.007
López-Alonso, M., Ibarrondo-Dávila, M. P., Rubio-Gámez, M. C., & Munoz, T. G. (2013). The impact of health and safety investment on construction company costs. Safety Science, 60, 151–159. https://doi.org/10.1016/j.ssci.2013.06.013
López Arquillos, A., Rubio Romero, J. C., & Gibb, A. (2012). Analysis of construction accidents in Spain, 2003-2008. Journal of Safety Research, 43(5–6), 381–388. https://doi.org/10.1016/j.jsr.2012.07.005
López-Arquillos, A., Rubio-Romero, J. C., & Gibb, A. (2015). Accident data study of concrete construction companies’ similarities and differences between qualified and non-qualified workers in Spain. International Journal of Occupational Safety and Ergonomics, 21(4), 486–492. https://doi.org/10.1080/10803548.2015.1085750
Manu, P., Ankrah, N., Proverbs, D., & Suresh, S. (2010). An approach for determining the extent of contribution of construction project features to accident causation. Safety Science, 48(6), 687–692. https://doi.org/10.1016/j.ssci.2010.03.001
Manu, P., Ankrah, N., Proverbs, D., & Suresh, S. (2013). Mitigating the health and safety influence of subcontracting in construction: The approach of main contractors. International Journal of Project Management, 31(7), 1017–1026. https://doi.org/10.1016/j.ijproman.2012.11.011
Memarian, B., & Mitropoulos, P. (2013). Accidents in masonry construction: The contribution of production activities to accidents, and the effect on different worker groups. Safety Science, 59, 179–186. https://doi.org/10.1016/j.ssci.2013.05.013
Ministerio de Fomento. (n.d.). Retrieved 29 September 2017, from http://www.fomento.gob.es/BE/?nivel=2&orden=09000000
Mohamed, S. (1999). Empirical investigation of construction safety management activities and performance in Australia. Safety Science, 33(3), 129–142. https://doi.org/10.1016/S0925-7535(99)00028-4
Pérez-Alonso, J., Carreño-Ortega, Á., Callejón-Ferre, Á. J., & Vázquez-Cabrera, F. J. (2011). Preventive activity in the greenhouse-construction industry of south-eastern Spain. Safety Science, 49(2), 345–354. https://doi.org/10.1016/j.ssci.2010.09.013
Ros, A., Ortiz-Marcos, I., Uruburu, A., & Palomo, J. G. (2013). A proposal for improving safety in construction projects by strengthening coordinators’ competencies in health and safety issues. Safety Science, 54, 92–103. https://doi.org/10.1016/j.ssci.2012.12.004
Rubio-Romero, J. C., Suárez-Cebador, M., & Abad, J. (2014). Modeling injury rates as a function of industrialized versus on-site construction techniques. Accident Analysis and Prevention, 66, 8–14. https://doi.org/10.1016/j.aap.2014.01.005
Salguero-Caparros, F., Suarez-Cebador, M., & Rubio-Romero, J. C. (2015). Analysis of investigation reports on occupational accidents. Safety Science, 72, 329–336. https://doi.org/10.1016/j.ssci.2014.10.005
Sexta Encuesta europea sobre las condiciones de trabajo: 2015 | Eurofound. (2015). Retrieved 29 September 2017, from https://www.eurofound.europa.eu/es/surveys/european-working-conditions-surveys/sixth-european-working-conditions-survey-2015
Swuste, P., Frijters, A., & Guldenmund, F. (2012). Is it possible to influence safety in the building sector?: A literature review extending from 1980 until the present. Safety Science. https://doi.org/10.1016/j.ssci.2011.12.036
Teo, E., & Ling, F. (2006). Developing a model to measure the effectiveness of safety management systems of construction sites. Building and Environment, 41(11), 1584–1592. https://doi.org/10.1016/j.buildenv.2005.06.005
Wu, X., Liu, Q., Zhang, L., Skibniewski, M. J., & Wang, Y. (2015). Prospective safety performance evaluation on construction sites. Accident Analysis & Prevention, 78, 58–72. https://doi.org/10.1016/j.aap.2015.02.003
Yung, P. (2009). Institutional arrangements and construction safety in China: an empirical examination. Construction Management & Economics, 27(5), 439–450. https://doi.org/10.1080/01446190902855633
Zhou, Z., Goh, Y. M., & Li, Q. (2015). Overview and analysis of safety management studies in the construction industry. Safety Science, 72(0), 337–350. https://doi.org/10.1016/j.ssci.2014.10.006
Papers relacionados
