1. INTRODUCTION

The magnitude of resources and investments adressed to research requires having measures and indicators that calculate their impact results. Bibliometric indicators, associated with other economic and social indicators, have been used for decades for planning and implementing every aspect related to research, since they can characterise very accurately its development and, therefore, support decision-making on scientific politics (Cunningham, 1997; González de Déu et al., 1997).

These type of indicators allow identifying and defining the characteristics of the most productive authors and the cooperation between them, research generating centres, intra-institutional and inter-institutional cooperation, primary sources in which projects are published, their productivity, concentration or distraction, as well as their national and international dissemination and the repercussion and impact that they have on subsequent projects (Aleixandre Benavent i Porcel Torrens, 2000; López Piñero i Terrada, 1992a i 1992b).

Likewise, the combination of bibliometric indicators with socioeconomic indicators allow obtaining the so-called "mixed indicators", which stablishes the relation between scientific productivity and agents or factors related to it. In this way, absolute values of research results reflected in publications can be adjusted to the factors that have an impact on its development, such as, by way of example, the number of existing researchers, the amount of budget for research, the relation between scientific productivity and the number of inhabitants or the threshold wealth of a specific region, country or geographical area. These indicators are useful so as to determine if the instability that prompt to be produced in the resources distribution, with a regional concentration favourable to centres or communities of a higher demographic and economic weight, are also reflected in a higher scientific productivity in relative terms (Agulló Martínez i Aleixandre Benavent, 1999).

In parallel with the increase of scientific productivity in the majority of the thematic areas, reflected in the increase of the total number of documents, a significant increase of cooperative works has been confirmed. This is caused because cooperation is necessary for the advance of knowledge, since, increasingly, issues need multi-disciplinary approaches and it is essential so as to supplement shortcomings and reach the critical mass of knowledge for enabling the advance of knowledge (Klein, 1996; Bordons i Zulueta, 2002). In this context, the National Scientific Research, Development and Technological Innovation Plan has tried promoting multi-disciplinary research, creating theme collaborative research networks whose aim is contributing to scientifically substantiate programs and politics in priority areas of the I+D National Plan through the association of multi-disciplinal and multi-institutional centres and research groups. The aim of this network typology is the creation of more powerful scientific cooperation schemes that allow achieving the objectives that would be difficult to set out in a more restrained execution context.

The study of scientific cooperation through the use of bibliometric techniques and the identification of research groups through Social Network Analysis (ARS) allow stablished researchers to increase their circle of contacts, promoting more participation in discussion forums and the exchange of ideas and the increase of groups and work networks or the integration of new members on them (Van Raan, 1999; Barabásis et al., 2001; Newman, 2004; González de Déu i Aleixandre, 2007).

As explained in the DGT official website, road safety is a major social issue, since in Spain, during the decade of 2001-2010 there were 41,665 casualties and 1,380,572 injured people, with an immense economical cost and a terrible human drama. The implementation of viable politics that can reduce the possibility of accidents and injuries from scientific approaches requires knowing all existent information, including the curent state of  research in this field, which can be achieved from bibliometric approaches.

Objectives

General objective

The general objective is analysing through the use of bibliometric indicators and networks analysis, the Spanish scientific activity on traffic, road safety and mobility reflected in publications made by Spanish institutions and groups. Its aim is that the information obtained could be used as a tool for knowing the strengths and weaknesses of the area and could allow using data for making easier the informed decision-amking on investments policy.

Specific objectives

• Quantifying through bibliometric indicators the scientific productivity of researchers and Spanish institutions that publish works on traffic, road safety and mobility in national and foreign journals of the area or other knowledge fields.
• Identifying through Social Networks Analysis the research groups and cooperation networks of Spanish institutions and its international cooperation.
• Analysing the distribution of works in relation to journals source of publication.
• Studying the interaction of the theme field of traffic, road safety and mobility with other knowledge fields.
• Describing the evolution of scientific production through time and determining the amount of research on traffic, road safety and mobility in the national and European context, identifying the most effective research lines and determining the existence of gaps in the research.
• Determining the research role on Spanish traffic, road safety and mobility in terms of the impact of published works.

2. METHODOLOGY

   1. Selection of databases

The works needed for carrying out the study have been compiled in the following bibliographic databases:

• Thomson Reuters databases included in the Web of Science (WoS) (SCIE-Science Citation Index Expanded i SSCI-Social Science Citation Index).
• Scopus of Elsevier database.
• ISOC database of the High Council of Scientific Researches.
• IBECS databases (Spanish Bibliographic Index on Health Sciences).

These databases are the ones usually used in bibliometric studies, since they include the most important and high-quality journals nationally and internationally, and have been used widely by many researchers and by our group.

Searches and downloads of information in these databases will be done through terminals of the Libraries and Documentation Service of the Universitat de València.

2. Strategy for the recovery of documents

The identification of documents under study has been conducted in several successive phases:

1. Initial research and selection in the mentioned databases of all works on traffic, road safety and mobility, in which had participated, at least, one Spanish institution.
2. Conceptual analysis of the records by the research team so as to assess the relevance and quality of the research. This procedure has been conducted through the reading of summaries and keywords of the articles obtained. This reading has allowed obtaining a widen list that compiles all terminology used in the works on traffic, road safety and mobility.
3. Carrying out a second research including all selected terms.

3. Download of information of recovered bibliographic records and elimination of overlaps

The results obtained in databases were the followings:

1. Download records from WoS and Scopus: 2,974 of WoS and 4,210 of Scopus.

Overlpas between WoS and Scopus: 4,654. Overlpads of WoS (2,327) are eliminated, keeping, therefore, Scopus records, but some data of the WoS, such as WoS Category, ID (record indetifier) and the number of citations of the articles were inserted in the resulting database. After eliminating the overlaps, the total number of records were of 4,846 (4,199 of Scopus and 647 of WoS).

2. Downloaded records from ISOC and from IBECS: 7,867 (6,090 of ISOC and 1,777 of IBECS).

It did not exist overlapped records between ISOC and IBECS, by owing to the fact of technique limitations of these databases, documents that were not journal articles were downloaded (such as letters, comments, biographies and editorials) and the same record several times. For this reason, once added bibliographic records in the relational database this type of duplicated records were eliminated. After eliminating records that were not journal articles, the total number of records was of 6,138 (4,782 of ISOC and 1,356 of IBECS).

In order to manage the records information, these were exported to a database in Access of Microsoft using an own developing program named Bibliometrics. Finally, overlapped bibliographic records between Scopus-WoS and ISOC-IBECS were integrated in a single bibliographic record. The number of records prepared to analyse were of 7,249.

4. Content analysis by the research team in order to choose the relevant works and reject unnecessary documents

Once the records are downloaded to the work database in Access, titles were read for obtaining a list of terms useful for making reviews so as to identify the relevant and irrrelevant works.

Two columns were created in the records table, 'relevant' column and 'delete' column. with the terms obtained as relevant and irrelevant, reviews in Access were made in the fields title, key words, abstract and descriptive words, and the corresponding indications were made in each column.

With the terms used, 1,12 records were not identified as relevant or irrelevant, so another group of records were examined that were not chose so as to obtain nre relevant and irrelevant terms. this process was repeated fime times.

Once all the records were marked, the ones that were in the 'relevant' and in the 'delete' field were examined manually.

Hereunder some examples of relevant and irrelevant terms are shown:

Examples of important terms used for marking records to be deleted from the database:

VANET or Vehicular Ad hoc Network; trafficking; fish; airspace; IEEE 802.3; IP network; Matlab Simulink model.

Examples of important terms used for marking records as relevant in the database:

smart cit* with traffic, vehicle or driv*, among others; electric vehicle; tramway (since it is related with mobility), roadkill, whiplash injury (when it is related with car and car accident, among others); automobile accident; brake assist system; driving licence; pedestrian death; road traffic; Matlab Simulink model.

The methodology used for determining the relevance or exclusion of documents of the database from the IBECS-ISOC is described below.

1st PHASE

A research by descriptive words was made in order to see the most repeated ones. From that point they were chosen and it was decided to choose 47 descriptive words considered relevant. Those were automatically ascribed as valid if they appeared in Title, Abstract or Descriptive words. Those descriptive words together with the number of times they appeared are shown below.

2nd PHASE

A research of other general terms was done, and those that had not been marked by the previous research were revised in order to see its relevance or irrelevance for the study. The process followed and a explanatory table are indicated below.

• Total: indicates the total of records found with the research.
• Relevant to the 1st phase: all of them that have been chosen as valid. In "accidente", for example, all of them that had “accidentes de tráfico” were included as valid.
• Relevant to the 1st phase and manual: are included the ones relevant in the 1st phase and the manual relevant.
• Delete: all of them that have not been valid for our study.

3rd PHASE

On this phase other terms were looked, since its generic nature could be linked to other scientific disciplines and were not chose in the previous phases of the combination of some of these terms with the ones in the 1st phase. Based on this data, a manual selection was done in the same way than in phase 2. The terms and results obtained are shown in the following table.

Some examples of these strategies about the theme of the documents and if they were chosen or deleted are indicated below.

• Búsqueda de automóvil

They have been marked as deleted all those works that deal with subjects such as:

• The automotive industry from an automotive industry demand perspective.
• After-sales services that cars need and how this have affected to manufactures prices policy and its distribution networks.
• Descriptive panoramic of mergers, acquisitions and alliances launched by automotive manufacturer business globally.
• Study of physiotherapy service in the night shift of an automative business.
• The automative industry has recorded excellent results in the year of 2020, both, in the internal market and in its external side.

• Autopista
  • Some of 'autopistas' that cover road insfrastructures in the creation of motorways, the payment of tolls, tourist signalling, have been included.
  • On the other hand, they have been downloaded others that although mentioning 'autopistas' are about wastewaters and its trade through motorways, such as "la utilización de los ríos naturales para desagües, empeora la situación así como la construcción de la autopista", others of internet about the global information motorway; and others about the study of cases such as the design of a motorway which construction is the cheapest as possible.
• Autovía
  • Some documents that cover motorways funding, infrastructures, noise, environmental awareness related to dual carriageway and improvements of dual carriageways are included.
  • The documents dealing with the problems for an archeological site caused by the passage of a dual carriageway have been deleted.
  • Records where the terms appearing are circumstantial to the study itself have been deleted. For example, the document named: “Políticas de comunicación de las instituciones públicas ante proyectos que originan una gran contestación social. Caso Zabalgardi”, in this case, the dual carriageway Madrid-Valencia is one of these projects, but there are many others, so this article was deleted. Another case was the article named “El ciclo completo de policía en Brasil”, which deals with citizen security by the police but only stablishes that polices can act in the field of dual carriageways for avoiding crimes.
• Bicicleta
  • They have been included the works that deal with the environmental impact by the use of bicycles, or injuries produced by bycicle accidents.
  • Articles about the health benefits by using the bicycle, or about a mental and motor improvement have been deleted.
• Camión
  • This term has produced some issues since not noly does "camión" include truck or trucks, but also terms such as "acercamiento" or others similar to it.
  • Records such as, for example, the one that covers “la muerte de un anarquista en un camión cisterna nazi” are deleted.
• Velocidad
  • Some related variables have been included, such as anticipatory speed as a preventive variable in youth road accidents.
  • All those related to speed of nerve conduction, speed of disease progression, speed runs of athletes, effects of speed of speech in languages have been removed.
  • Documents dealing with high-speed rail without affecting roads or traffic have been deleted.
• Vehículos de motor
  • Some documents dealing with staff working in factories constructing motor vehicles have been deleted.
• Transporte escolar
  • Some documents dealing with the improvement of social services such as school transport, TIC rooms, etc., were deleted, as well as some dealing with school transport as an economic cost in families and the lack of resources.
• Transporte de mercancías peligrosas
  • Some dealing with the transport of these goods by railway were deleted.
• Transporte de mercancía
  • Documents dealing with European trade have been removed, citing freight transport as something tangential that may have affected this trade, as well as the analysis of cost efficiency in freight transport companies.
• Seguro de responsabilidad civil
  • Some dealing with this insurance but referred to the environment have been deleted.
• Preveción de accidentes
  • Those dealing with the prevention of work accidents, of children, in the education, but not of traffic, have been deleted.
• Multa

  • Documents dealing with fines for drug abuse, or related to drug trafficking have been deleted.

  • Those that had the term "simultaneidad" have been deleted, since it is a completely different field from traffic..

• Motocicleta
  • A record has been chosen for dealing with "Quiste pericárdico post-traumático de localización atípica" since it includes the keyword "motocicletas" understanding that it is an injury that can be caused by an accident or by the use of a motorcycle.
• Ciclista
  • Some were removed that dealt with physical skills in cyclists but not in relation to traffic or road, as well as sports performance in cyclists for nutrient utilisation and similars.
• Ciclismo
  • They were deleted those records dealing with injuries in cyclists that were not caused by falls caused by traffic accidents, but for common knee injuries.
• Carretera
  • They were deleted all documents dealing with archaelogy that have been found in deposists, necropolis where a road has been affected.
  • All those dealing with compulsory purchase issues have been removed as they were affected by road construction.
• Tráfico
  • In the individual search for trafficking, internet, human trafficking, trafficking in women, commercial, maritime, real state, organs, immigrants, influence peddling, etc. have been eliminated.
  • By searching tráfico so as to avoid the accent, appear terms such as "trófico", or "catastrófico", "neurotrófico", etc. These records have been deleted.
• Conductor
  • The vast majority of documents reviewed manually use the expression “hilo conductor”, “eje conductor”, and for that reason they have been deleted. As well as “elementos conductores” or “el líder como conductor del grupo”.
• Conducción
  • Documents dealing with “conducción el agua”, “conducción eléctrica”, “conducción cardiaca”, “conducción nerviosa”, “estudios de conducción motora”, and other expressions that are not related with traffic have been deleted. However, thet were kept dealing with reckless driving.
• Vial
  • Apart from “seguridad vial”, the term “vial” has been looked for in order to verify if it gave relevant results. Some of those dealing with road safety education or road safety training have been marked as good, while others dealing with road safety, trivia, etc. have been discarded.
• Vehículo
  • Were deleted all of those talking about a vehicle as a source for obtaining something: “vehículo para la divulgación”, “vehículo trasmisor”, “usado como vehículo”, “vehículo de transmisión”, etc.
• Cirulación
  • Documents dealing with the movement of data, persons and goods, or related to the trafic of blood were deleted..
• ​​​​​Transporte
  • All documents dealing with “transporte urbano”, “transporte interurbano”, “transporte metropolitano”, “transporte terrestre”, “transporte por carretera”, “transporte público”, since they are included in the section of mobility. Then, the 1,332 documents that only contain the word  “transporte”, were revised. There were 1,082 valid records for using one of the compound terms mentioned before..
• Tránsito
  • Some documents containing terms such as “transitoria”, were deleted, for example, the diagnosis of transient neonatal myasthenia gravis.
  • Other deleted records were the ones dealing with “tránsito digestivo”, “tránsito gastrointestinal”, “tránsito político, económico”, “tránsito entre épocas”, etc.

4ª FASE

Once all the previous combinations wer done, there were still 1,4000 records that were reviewed manually. In some of them appeared the term "motriz" or "motor", making reference to motor aspects of the body or to mobility or motor skills. For example, the document named “Tratamiento de la punción dural accidental en la analgesia epidural del trabajo de parto”,since it includes the term “accidental”and in the abstract it appears the term “bloqueo motor”.

After processing all documents in the previous phases, there were still 4,909 records considered valid.

5. Standardisation of authors names and institutional affiliations

In this phase, a process of standardisation of the different variants by which the same author or institution is identified in the case of works signed by at least one Spanish institution was carried out, since the bibliographic data in databases such as those used often present quality problems, especially important when authorship is analysed, either because the authors themselves do not always sign their works in the same way or due to erros at the time of processing the information.

In order to minimise this problem, a thorough manual filtering for unifying the different variations of names of an author or institutions has been carried out. When it comes to authors, the procedure that has been followed with two or more variations of a same and/or surnames consisted in verifying the coincidence in the work places of the different variations. As far as institutions are concerned, apart from the existence of variations in its names, there is a problem when it comes to some bibliographic records are likely to include in an institutional affiliation two or more institutions (specially in the case of research institutites linked to universities). In this case, apart from standardising different variations, the firms of those records in which appear more than an institutional affiliation willl be differentiated. In this way, the number of signatures for each bibliographic record should be as many as the number of institutions that can be identified. Some examples are shown below:

Variations of the firm of Francisco Javier Álvarez:

Álvarez, F. J
Alvarez, F. Javier
Alvarez, F.J.
Álvarez, F.J.
Alvarez, F.Javier
ALVAREZ, FJ
González, F.J.Á.

Variations of the firm of Miguel Ángel Sotelo:

Angel Sotelo, Miguel
Sotelo, M.A
Sotelo, M.A.
Sotelo, M.-A.
Sotelo, M.Á.

Variations of the institutional firm of the Universidad Complutense de Madrid:

Complutense University Madrid
Univ. Complutense de Madrid
Universidad Complutense, Madrid
University Complutense of Madrid
Univ. C., Madrid,
Complutense University
UNIV COMPLUTENSE MADRID
UCM

Variations of the institutional firm of the Universitat Politècnica de Catalunya:

Univ. Politécnica de Cataluña
UNIV. POLITECNICA DE CATALUNYA
UPC BarcelonaTech
Technical University of Catalonia
Tech Univ Catalonia
Univ Politecn Cataluna
Polytechnic Univ in Catalonia
UPC

6. Indicators calculation

Different measures or indicators have been calculated organised in three groups: indicators of scientific productivity, indicators of collaboration and indicators of impact or visibility. Its name, the calculation procedure and the type of information they provide are included below.

1. Indicators of scientific production

These type of indicators offer a quantitative dimension of the scientific production, in an absolute or relative way, when analysing in absolute values other variables .

2. Collaboration and network indicators

These type of indicators offer a quantitative dimension of the concurrence and cooperation degree of the different scientific agents when carrying out the researches. The development of disciplines and of the scientific knowledge requires, more and more, sharing resources and knowledge, existing a positive correlation between collaboration of scientific works allows identifying the role played by each of the agents in the scientific system, enhancing the integration between them and detecting possible lacks or gaps. These networks were built from the quantification ot signatures in works.

3. Indicators of impact or visibility of works

The impact of visibility of works measures the effect that works have in other subsequent studies from the citations they have received. Despite its limitations and controversial nature, it is the most widely used measure to evaluate Spanish scientific activity, as the journals are included in the Thomson Reuters lists are those indicated by the National Commission for the Evaluation of Research Activity (CNEAI) as a reference for the evaluation processes. The following indicators have been determined: number of citations in the Web of Science and in Scopus, average number of citations per work, impact factor of journals and quartile in Journal Citation Reports.

4. Identification, analysis and evolution of research lines

In order to identify the research lines, it has been used a combination of bibliometric indicators based on the frequency of keywords and of the thematic areas of the journals in which works are published, and visualization techniques based on the Social Network Analysis, SNA. In this way, the main scientific fields in the area have been identified, delimited and visualised, as well as possible gaps and future trends. On the other hand, the evolution of the main lines of research from their creation to the present day has been followed, showing two perspectives of each scientific field: static and dynamic.

The visualisation of research lines and its descriptive words or keywords was developed combining several techniques from the SNA. On one hand, the Pajek software with the Pathfidner pruning algorithm to reduce the dimensional space and the Kamada-Kawai as visualisation algorithm, setting the necessary frequency thresholds to achieve a sharp visualisation in the generation of the maps, tipically drawing networks composed of less than 100 nodes. On the other hand, the VOSviewer software, which combine visualisation and clustering techniques. Among the different representation systems, labelled maps in which each node corresponds to a descriptor were chose.

5. Identification, analysis and evolution of research lines

In order to determine the position of Spanish research of traffic, mobility and road safety in Spain, a search was carried out in the Web of Science, the records retrieved were downloaded and the country of origin of the papers was identified. If in a work took part several institutions of the same country, this country was only counted once, and if took part several institutions of different countries, a work has been allocated to each of the participating countries. In the bibliometric analysis the following types of documents have been taken into consideration: articles, reviews, letters and editorials. The search was carried out on 11 February 2016 and the year 2015 was stablished as a limit date.

3. RESULTS

   1. Indicators of scientific production

      1. Annual evolution and productivity and impact by journals

The total number of articles published was of 4,909 and has increased progressively from the period of 1969-1974, in which 13 works were published, until the five-year period of 2010-2014, in which 1,876 works were published (figure 1); this last five-year period almost doubled the previous production (2005-2009).

Works have been published in 1,382 different journals. The table 1 shows 23 journals that have been published more than 25 articles. The journals that include a greatest number of articles have been "Carreteras" (n=178), followed by La Ley. Revista Jurídica Española de Doctrina, Jurisprudencia y Legislación (n=85) and Revista de Responsabilidad Civil, Circulación y Seguro (n=85). Of these journals, 16 are edited in Spain and 7 in the United Kingdom. The foreign journal with a greater number of works has been Accident Analysis and Prevention (n=79).

The increase in the number of articles published has been bigger in some journals, such as Carreteras, Accidents, Analysis and Prevention, IEEE Transactions on Intelligent Transportation Systems, Transportation Research Record y Construction and Building Materials. The comprehensive list of journals is shown in the annex 1.

The table 2 shows some impact indicators of the most productive journals. It has been used the impact factor of the Journal Citation Reports of the year 2014. The journal with a greater impact factor was the American IEEE Transactions on Intelligent Transportation Systems (FI=3,064), followed by Construction and Building Materials (FI=2,818) and Transportation Research Part A-Policy and Practice (FI=2,725), both edited in the United Kingdom. The journal with impact factor in which a greater number of articles have been published is Accident Analysis and Prevention (FI=1,964). Of the 7 journals in the table that have an impact factor, 6 are found in the first quartile of its respective categories in the JCR.