Internationalization and Language Assessment in Brazil: Exploring the Interface between Language Proficiency and Rankings at UFES

The study is based on the assumption that knowledge of English is an important factor in the internationalization of higher education and that the English without Borders program has positively affected the overall level of proficiency in English as well as internationalization levels in Brazil. So as to verify this hypothesis, the study analyzed the interface between English proficiency levels measured with the TOEFL ITP test and internationalization scores measured in terms of rankings in a federal university in Brazil. Results of the study suggest that during the period analyzed these two variables remained stable, thus confirming the hypothesis raised.

Language policies can be analyzed through governmental actions and programs such as the Brazilian government-funded Science without Borders (SwB) and Languages without Borders (LwB) internationalization programs (Finardi & Archanjo, 2015;Finardi, Leã o, & Pinheiro, 2016). Indeed, one of the theses that this study aims to advance is that the English without Borders (EwB) program, part of the LwB, is perhaps one of the most important language policies proposed in Brazil having affected the overall level of proficiency in English as well as internationalization levels in that country. So as to verify this hypothesis, the present study analyzed the interface between one of the most important actions of the EwB, namely, the TOEFL ITP test, in relation to internationalization scores in a federal university in Brazil.
Though the context of this study is very particular (a federal university in Brazil), it is possible to see how results and reflections made in this paper relate to other contexts such as Turkey, according to Taquini, Finardi and Amorim (2017). Considering the fact that many countries are facing the pressure to internationalize their higher education institutions, we think that the discussion offered here regarding the interface between foreign languages in general and English in particular in the process of internationalization of higher education may serve other contexts where English is not the official language.

Review of Literature
The federal university where the present study was carried out (Note 1) was selected for this study because of results of previous studies carried out there and which concluded that one of the main challenges for its internationalization was the lack of overall English proficiency of its academic community Finardi & Ortiz, 2015;Finardi, Santos, & Guimarã es, 2016, Kawachi, Amorim, & Finardi, 2017.
Still regarding the role of English in the internationalization process, but not specifically in the same university, Finardi and Franç a (2016) examined the production and circulation of the Brazilian academic publication in the area of Letters in English and in Portuguese. According to these authors, although Brazil had the 13th largest academic production in the world, it had little international impact because its publications were made mainly in Portuguese.
In Brazil and as reported by Finardi and Guimarã es (2017) and according to Gama and Klagsbrunn www.scholink.org/ojs/index.php/selt Studies in English Language Teaching Vol. 6, No. 2, 2018 141 Published by SCHOLINK INC.
(2014), many educational agents and policy makers claim that it is possible to measure an institution by quality and quantity indicators, including rankings. These authors also claim that it is by no chance that the Brazilian Law of Education (Lei de Diretrizes e Bases da Educaç ão, LDB) determines the minimum standard of quality of education offered by the State (BRAZIL, 1996). This standard also applies to higher education where the main indicator of quality is the GIC-General Index of Courses (INEP, 2011), which considers the quality of undergraduate and postgraduate courses. Another index of undergraduate quality measure is the Preliminary Courses Concepts (PCC) used to assess the performance of students in the National Student Performance Exam (ENADE). The Higher Education Personnel Improvement Coordination-CAPES-is responsible for assessing (with a score that goes from 3 to 7) the quality of graduate courses for each area. Another institutional evaluation measure that has gained popularity with the increasing internationalization of higher education and which has provoked much discussion and controversy regarding its use is university rankings, also reviewed in Finardi and Guimarã es (2017) and partially reported here.
Lourenç o (2014) claims that rankings are embedded in public evaluation policies, especially regarding large scale assessments as, for example, the former National Course Examination (ENC) and even the National Student Performance Examination (ENADE) though the author warns against a number of criticisms that have been raised against rankings claiming that they are not good quality indicators and that their assessment may be unfair since it may be based on erroneous data.
Yet, Lourenç o (2014) recognizes other authors who are in favor of rankings because they understand that they may confer rights and provide useful information as such they may constitute powerful tools to strengthen the rights of consumers. Laus and Magro (2013) suggest that university rankings have been used as benchmarking for self-development of institutions, providing quality certification and exchange of good practice for the internationalization of Higher Education Institutions (HEI). Another effect of rankings is the international visibility and competition between HEIs to be recognized as world-class universities.
Among the most respected international rankings are the Academic Ranking of World Universities According to the 2014 Latin American ARWU, USP lost the first place in the region for the Catholic University of Chile whose scientific production in collaboration with foreign scientists was greater than that of USP. According to Gama and Klagsbrunn (2014) one of the reasons why USP is not in the list of the top 200 universities is the lack of English Medium Instruction (EMI) there. Regarding the use of EMI in Brazil, Martinez (2016) states that very little evidence can be found for it before 2010.
According to the same survey reported in Finardi and Guimarã es (2017) and in Gama and Klagsbrunn (2014), Brazilian universities have ten universities among the top twenty in Latin America: USP, UNICAMP, UFRJ, UNESP, UFMG, UFRGS, PUC-RJ, UNIFESP, UNB, UFSCAR and of these, only USP, UNESP and PUC-RJ offer EMI courses at undergraduate level. Still in relation to the role of English and regarding the ARWU assessment of Brazilian universities, Gama and Klagsbrunn (2014) claim that Brazilian universities have good results in number of papers per faculty member, but not in citations which shows that Brazilians are doing better in terms of quantity than on quality and as suggested by Finardi and Franç a (2016), the quality assessment may be affected by the language of publication since the Brazilian academic production is still published mainly in Portuguese.
Regarding Brazilian rankings and still as reported in Finardi and Guimarã es (2017) should also take into account regional characteristics of HEIs recognizing that quality concepts may not be shared by different institutions. Regarding methodology, the IREG recommends the use of audited and verifiable data whenever possible, including those collected in accordance with recognized procedures for data collection, enhancing the credibility of classification, such as using advisory boards.
Regarding the presentation of results, the IREG recommends the provision of clear information regarding all the factors used to develop a classification original data being organized or published in a format that allows the compiler to make corrections, if needed, so that users are aware of it. The principles of Berlin, as they came to be known, have been adopted in a multi-dimensional classification model, which considers the roles that universities play in the society and in the education system in which they operate. We can conclude that rankings require a substantial investment of resources from HEIs and as suggested by Finardi, Santos and Guimarã es (2016) and by Finardi and Guimarã es (2017), they seem to have been designed by and for English-speaking universities or by those that adopted English medium instruction (EMI) curricula.
Regarding the role of English in EMI and in the internationalization process, Taquini, Finardi and Amorim (2017) compared the offer of EMI courses in Turkey and in Brazil, two countries that were strategically attempting to increase the number of courses offered in English, especially in graduate courses, so as to be accepted in the European Community (in the case of Turkey) and so as to internationalize Brazilian institutions, in the case of Brazil. The study concluded that the offer of EMI course in those countries is still timid and subject to a number of challenges, among which the greatest is the lack of proficiency in English of university students and faculty.
As suggested by Laus and Magro (2013) and Finardi and Guimarã es (2017) and Japan (5th), but also among major developing countries such as China in second position and India (10th) both in the BRICS group with Russia in 15th position. Still according to Laus and Magro (2013) and as reported in Finardi and Guimarã es (2017) Similarly to the ARWU and the THES, the Scimago also has regional rankings such as the Ibero-American Ranking of International Scientific Production whose indicators include the scientific production, measured by number of publications, international collaboration, as measured by joint publications with other countries, the average scientific quality, measured by the scientific impact of an institution, and the percentage of papers published in the most influential journals in the world, measured by the amount and origin of the citations they receive. According to this ranking, Spain and Brazil lead the scientific production in Latin America and Spanish and Brazilian universities appear among the ten most productive ones, with five Spanish, four Brazilian and one Mexican institution in the list.
With regards to international cooperation and still as reported in Finardi and Guimarã es (2017) Another variable that this study aims to analyze in relation to the assessment of internationalization of higher education is related to language. According to Gimenez (2013), Brazilians want to speak English fluently but only about 5% of them do so. Assuming that English expands access to information (Finardi, Prebianca, & Momm, 2013) and education online (Finardi & Tyler, 2015), and given the fact that very few Brazilians speak English fluently, Finardi (2014) suggests that English should be taught as an international language in Brazilian schools so as to fight against the social gap created by the offer of private language courses available to only a small (and privileged) part of the population.
Finardi (2014)  The present study focuses on one of the three main actions offered by the EwB, namely, the TOEFL ITP test insomuch as it offers an x-ray of the language proficiency level of a given institution. So as to review the context in which the test was thought of by the program, the next section reviews a study carried out by the EwB's president and vice-president in relation to the TOEFL ITP test. universities which adopted EMI courses.

Still according to Abreu-e-Lima and Moraes Filho (2016), another reason for the administration of
English tests in Brazil was to have a diagnostic evaluation of the overall level of proficiency in English of Brazilian students in public universities. The authors go on to explain that so as to guarantee that the diagnosis was internationally endorsed, it was crucial that a verified and recognized evaluation instrument was used, that is, a proficiency exam which was well accepted in the international community. The EwB board concluded that for a test to be administered across regions in Brazil it was necessary to consider regional and specific characteristics of universities in Brazil and so paper-based exams were selected because the only infrastructure required for their administration are classrooms, trained proctors and audio resources. Another factor taken into consideration was that the test selected had to be academic oriented and of easy correction, without involving production skills, which would complicate logistics and add additional costs to the correction of exams. Also, the exam had to be an internationally recognized exam so as to be accepted by most foreign universities. Given all these  (2016) suggest that many local coordinators estimate there is a large percentage of students in A1 level, but those students have not registered to take the test because they may feel they do not have the minimal requirements to "pass" the test. As it is, Abreu-e-Lima and Moraes Filho (2016) believe that the tests were taken by those who believed to have some knowledge of English.

Method
So as to verify whether the English without Borders (EwB) program in general and the TOEFL ITP test in particular have positively affected the overall level of proficiency in English as well as internationalization levels at a given institution, a study was carried out at a federal university (UFES) where previous studies had suggested that the level of English proficiency of its academic community impacted its internationalization levels , Finardi & Ortiz, 2015Finardi, Santos, & Guimarã es, 2016, Kawachi, Amorim, & Finardi, 2017. With that aim, the internationalization levels of that institution, assessed in terms of rankings, as well as the English proficiency levels, assessed in terms of the TOEFL ITP scores, were analyzed. In what follows, the context where the data were collected will be described.

Context for Data Collection
Data was collected at the Federal University of Espí rito Santo (UFES the Coimbra Group of Brazilian Universities (GCUB).

UFES also participated in the Science without Borders program until its extinction and in the English
Without Borders program since 2013. The EwB at UFES offers 540 seats for face-to-face English for Specific Purposes (ESP) and English for Academic Purposes (EAP) courses for the academic community at UFES. According to Kennedy (2012), projects related to teaching ESP have been expanded because of globalization, language policies and the need to use English as an international language.

Results and Discussion
Regarding UFES evaluation in terms of rakings, it appears in only four of them, namely: 1) the Intelligence Unit (Note 9), 2) the Folha (Note 10), 3) the Top Universities (Note 11) and 4) the Scimago (Note 12). A search in these databases for the position of UFES since 2013 (when the TOEFL Test started to be administered nationwide in Brazil and at UFES) in these rankings shows that it varies depending on the ranking used. Table 3 shows the ranking position of UFES between 2013 and 2016: As we can see in Table 1, UFES' ranking is somewhat variable depending on the ranking analyzed. It is important to mention that the rankings for 2016 must be taken as speculative since the year had not finished (as of the date of writing this manuscript) and thus the assessment may be partial and incomplete. If we look at the Top Universities ranking, the only one whose ranking for UFES was available between 2013 and 2016, we can see that the evaluation of this university has somewhat improved over the years and so it is expected that the final ranking for UFES in 2016 is better than the one reported in Table 1 (129). In what follows the results of the TOEFL ITP results for UFES will be analyzed so as to see whether it is possible to see a pattern of improvement of scores of this test, once according to Finardi and Ortiz (2015), Amorim and Finardi (2017), Kawachi, Amorim and Finardi (2017) and Finardi, Santos and Guimarã es (2016), the overall (low) level of proficiency in English of the academic community at UFES is one of the obstacles for its internationalization.   As we can see from the TOEFL ITP results at UFES between 2013 and 2016, it is safe to say that the overall level of English proficiency there has not improved but rather maintained an average of 50% level A2, 30% of level B1, 15% of level B2 and 2% of level C1 as can be seen in Tables 9 and 10 and in Figure 1 below.    If we triangulate the levels of English proficiency and internationalization levels at UFES, we can say that these levels have remained stable in the period analyzed, which might corroborate the hypothesis raised in this study, namely, that internationalization levels should increase as the overall English proficiency level of the academic community of UFES increases. Amorim and Finardi (2017) seem to agree with this interpretation when they analyzed the level of engagement of the academic community in the referred university, concluding that despite their enthusiasm towards internationalization actions, the academic community of UFES still lacks engagement with the TOEFL ITP test and other internationalization actions offered by the EwB program at UFES.

Conclusion
The present study analyzed the interface between English proficiency levels measured in terms of the TOEFL ITP test and internationalization scores measured in terms of rankings in a federal university in Brazil. The results of the study confirm that during the period analyzed these two variables, namely, the English proficiency levels and internationalization rankings at UFES, remained stable, therefore confirming the hypothesis raised.
It is possible to notice that, although the university community recognizes the importance of the English language (as reported by , their proficiency level is still below expected. As it was verified in Kawachi, Amorim and Finardi's (2017) study, proficiency levels remained stable despite the efforts of promoting courses and workshops designed to meet the community's needs. One possible explanation for this unexpected result is the low level of engagement of the academic community investigated with the internationalization actions  and the short period of time of this investigation (2013-2016) which perhaps was not sufficient to show considerable results in proficiency levels at UFES. Regardless of the reason for these results, the study calls the attention to the fact that it is fundamental to raise the awareness of the importance of assessment variables such as proficiency levels and internationalization rankings in HEs though we must be very careful to design and select these variables, as suggested by Finardi and Guimarã es (2017). Moreover, the study concludes with the suggestion that these variables should be discussed with the academic community and integrated in curricula design in order to guarantee more positive, effective and long-lasting results.