The origins of CLIL can be traced back to the immersion programs of the 50 s in Canada. In that decade, thousands of Canadian families moved from the English-speaking area to the French-speaking region of Quebec and Montreal. Many studies which explored these types of bilingual programs (Pérez-Vidal, 2011) showed their success from the linguistic and content point of view. In fact, Pérez-Vidal (2011) underscored the influence of these programs in the implementation of CLIL in Europe and in many ways can be considered the precursors of CLIL education. Nowadays, CLIL is implemented in and out of Europe and is experiencing success globally, specifically in Canada and more recently in United States due to immigration. Both CLIL and immersion, together with other models of bilingual education seem to evidence that L2 instruction integrated with content matter is more effective than L2 instruction in isolation (Badertscher and Bieri, 2009).

Nonetheless, it is important to highlight that the European context cannot be directly compared to the North American one despite their similarities and the influence of the latter over the former. Consequently, not all content-based or immersion principles can be transferred to the CLIL approach. As Gallardo del Puerto et al. (2009) pointed out, “most of the immersion conditions bear little resemblance to the study of English through CLIL programs in Europe, particularly in terms of the sociolinguistic and sociocultural context in which the L2 is learned and the authenticity of the input” (p. 65). In fact, the L2 adopts a secondary role in content-based teaching contexts even though it is the language of instruction. On the contrary, CLIL pursues a balance between the L2 and content in terms of their weight in instruction. Thus, language and content are intertwined without a preference of one over the other (Coyle, 2007).

The CLIL approach aims at the creation of a learning context where the L2 acquisition takes place in a naturalized way (Juan-Garau and Salazar-Noguera, 2015). In addition to the integration of language and content, Mehisto et al. (2008) adds that CLIL must also pursue the development of learning skills, as they are essential to reach the linguistic and content matter aims. CLIL is characterized by its flexibility, dynamism, and versatility (Papaja, 2014). Yet, the 4C framework suggested by Coyle (2006) constitutes a reference for researchers and teachers. These four Cs stand for content, communication, cognition, and culture. As for content, it refers to non-linguistic content or subject matter. Any subject may adopt the CLIL approach. Meyer et al. (2015) affirm that CLIL is not only a question of acquiring knowledge, but that the learner is expected to take ownership of his/her own knowledge. In this line, content is closely related to cognition and learn to learn, which encompasses a cognitive challenge. The concept of culture is also relevant for understanding CLIL. The student must develop intercultural awareness, which is at the core of CLIL (Coyle, 2006). CLIL provides an opportunity for students to understand and tolerate different perspectives and ways to conceive reality.

From the standpoint of communication, language in CLIL plays a complex role. Authors such as Coyle et al. (2010) and Llinares et al. (2012) distinguish between language of learning, language for learning and language through learning. Language of learning entails the words, phrases and expressions which are key to access the subject content. The language for learning is needed by students to carry out classroom tasks such as debating or presenting information. The third role of language in CLIL is related to the new language that emerges from the cognitive process in which students are immersed.

The 4C framework introduced has evolved and some scholars have recently considered adding a fifth C to Coyle’s proposal. For instance, Agolli (2013) suggests the C for context should also be considered as, that which determines instruction and at the same time enriches the framework. Others such as Renau and Mas (2019) suggest that the new C should refer to competence: “the can-do statements […] students are able to make after the lesson, either about the lesson content or about the language that is being learnt” (p. 1113).

An important portion of the CLIL research hinges on exploring this approach in terms of motivation, analyzing the attitudes of teachers, students and even families. Most of the studies reveal that there is a positive attitude toward CLIL, although there is a certain degree of concern about the challenges involved in its proper implementation (Dafouz, 2009; Alcaraz-Mármol, 2018). For instance, Issabekova et al. (2020) remarks on the success of the implementation of CLIL programs in Kazakhstan, although the authors warn that there is a need for both teachers and students to rethink some traditional language and subject concepts in the teaching and learning process. Likewise, Nurzhan and Ashirimbetova (2017) observed that the CLIL approach increased students’ motivation in subjects such as Math. In this sense, Pladevall-Ballester (2016) states that we should take into account the subject CLIL will be implemented for, requiring willing, prepared and involved teachers. These teachers are in charge of choosing appropriate materials and deal with parents’ expectations and concerns.

Another research line for CLIL concerns the development of the L2 communicative competence in general, and each of the communicative skills and linguistic aspects in particular. On this basis, research strongly supports the benefits of CLIL in the mastery of a L2 in comparison to non-CLIL contexts (Lorenzo, 2010; Várkuti, 2010; Navés, 2011; Goris et al., 2013; Yang, 2015; Coral et al., 2018). Comparisons of L2 vocabulary acquisition in CLIL vs. non-CLIL contexts is one of the aspects where more literature can be found in this respect. Agustín-Llach (2016) carried out a longitudinal study where she measures and compares the receptive vocabulary size and lexical growth of one group of CLIL learners and another group of traditional EFL learners. Her results show significantly higher vocabulary size in the CLIL learners, which increased in the last years of the study. A significant effect of CLIL on L2 vocabulary knowledge was also observed in Xanthou (2011), and Sylven and Ohlander (2015) when the CLIL group outpaced the non-CLIL one. When considering morphosyntactic elements of a lesson, Ruíz de Zarobe (2008) and Lázaro (2012) indicate a clear advantage of CLIL learners over traditional EFL students.

Focusing on skills, Nieto (2016, p. 83) concludes that “although there is a growing body of research that shows the many benefits of the CLIL approach, there are still some aspects that require further investigation, such as the most and least benefited skills in the target language” (p. 83), and deems further research is necessary to confirm those benefits. For instance, Hüttner and Smit (2014) and Pérez-Cañado and Lancaster (2017) argue that it is oral communicative skills, particularly speaking, which are most benefited by CLIL in the medium and long term as communicative interaction in the L2 is clearly promoted. Likewise, Gené-Gil et al. (2015) found that writing skills were better among CLIL peers than among EFL students, the former outperforming the latter in all aspects except for lexical complexity.

Nonetheless, Dalton-Puffer (2008) affirms that listening and reading are the most positively affected by CLIL, even though they have usually received less attention than production. Aguilar and Muñoz (2014) showed that there was a significant improvement in students’ listening skills within a CLIL context, in comparison with EFL contexts, especially regarding less proficient students. In terms of reading, Ruíz de Zarobe and Zenotz (2018) observed how reading strategies and awareness were enhanced in the CLIL classroom, imparting a positive impact in L2 reading comprehension. In this line, Prieto-Arranz et al. (2015) highlighted a remarkable development in reading comprehension in secondary education learners immersed in CLIL programs.

CLIL has adopted several models, as it is a versatile approach which can be implemented from elementary to higher education (Coyle et al., 2010). One of the educational approaches involves team teaching. No matter which model is used, CLIL aligns with team teaching in which the content and language teacher collaborate (Coyle et al., 2010; Chocholatá and Gahurová, 2018). During a lesson two teachers are in the class (a language teacher and a subject content teacher) and these teachers cooperate not only during planning the lesson but also during the lesson. This approach demands methodological preparation of both teachers (Hanušová and Vojtková, 2011). Content teachers usually keep the balance between teaching the subject content and helping language understanding (Coyle et al., 2009).

Team teaching provides each teacher with his/her own role in the lesson and be a part of the whole message transmitted to students. In fact, it is important to consider not only the content but also the language structure used for this message which, in some cases, would not be analyzed when the topic is covered in the mother tongue (Elorza and Guinda, 2015). The teacher of the content subject sets out subject aims and subsequently on their basis, set out the language aims in cooperation with a language teacher (Hanušová and Vojtková, 2011).

In the CLIL context, the first debatable issue is “what” to assess. Should teachers assess language or content or both (Coyle et al., 2010). Teachers should bear in mind that when assessing their students, the focus should first be on the content, then on the language (Coyle et al., 2010).

Experiences of CLIL with science subjects have been widely documented due, perhaps, to the prominent role of English in the scientific community. In non-English speaking countries, various types of English communication skills are being incorporated in higher education to help students reach a reasonable English level after graduating from universities. As a result, STEM and English education are becoming closer to each other in these countries (Capone et al., 2017; Mutseekwa, 2021). Nonetheless, as far as we know, despite the dual focus of CLIL, most of the reported results pay attention only to language learning, or to content (Capone et al., 2017) no matter the stage of education (Beas Catena, 2020). One study from the literature, developed in a multicultural context, reports that content results improve after CLIL biology lessons. It is this author’s opinion that this is the weakest conclusion for the study considering the evidence provided from the research as it exhibits a weak experimental design that lacks a control group (Schietroma, 2019). Among the rest of literature reviewed, the focus is on language. Thus, for instance, in the work of Pérez-Mirón (2016) for Primary, Compulsory Secondary Education and High School students in Spain, by using macro-activities like races and external contests in chemistry, physics, biology and mathematics through cooperative learning, the author reports that students feel more confident regarding science subjects, and improve their communication skills in English, but no results about content learning are presented.

Also, in preuniversity levels, in the study of Tagnin and Ní Ríordáin (2021), biology is taught to upper secondary education students through CLIL with the use of inquiry. They proved that the use of inquiry can help students to develop the argumentation and discourse creation in English. These authors recommended to Science, Technology, Engineering and Mathematics, STEM, teachers not to oversimplify the discipline language aiming to a better understanding of contents and language structure.

When considering Higher Education, there are few studies about STEM subjects taught through CLIL at university level and also there are few that report results regarding content knowledge. For instance, Aiba and Izumi (2019) analyzed how effective CLIL is in teaching STEM subjects to STEM major students when using input or output skills by focusing on cause-effect language structure, highly present in scientific contexts, but without paying attention to any content. They find that students learn no matter which skills are involved or promoted, provided that neither the input nor the output skills can exist in an isolated manner in any teaching activity, and both are combined in teaching practices in the reported CLIL environment. The perceptions of teachers are captured by Block and Moncada-Comas (2019), who studied science teachers using English-medium instruction, and find that they are acting as English as foreign language teachers, without their realization.

Therefore, the reviewed literature provides evidence on the gap that there are few studies that use a quasi-experimental design and assess content learning of CLIL. This study is aiming to fill this gap to any extent.

CLIL has gained importance over the last decade, especially in some Asian countries. The government of some countries are seriously considering this approach and take action for its promotion (LaPrairie, 2014). Ministry of Education and Science of the Republic of Kazakhstan [MoES] (2014) states that Kazakhstan is one of the first Central Asian countries, and it is among the first post-Soviet countries, to adopt CLIL as a pedagogical approach through which content subjects are taught in three languages.

Different training programs have been designed since 2016 with the aim to prepare STEM teachers so that they can be able to deliver instruction in English. Thus, extensive teacher training programs in Kazakhstan have incorporated CLIL among their contents (Goodman and Karabassova, 2018). Through interviews and observations Karabassova (2018) found that many of the teachers were not familiar with the educational goals behind CLIL and conceptualized CLIL simply as teaching in a new language. More recently Nuranova (2020) used interviews with teachers teaching science subjects and discovered that science teachers experienced numerous challenges when teaching with CLIL, specifically, increased class preparation, curriculum design, and low competence in English on the part of teachers and students. However, teachers participating in the study who exhibited greater experience felt that it only influenced their daily teaching routine.

In the studies conducted in Kazakhstan, it has been emphasized the difficulty of assessment by content teachers in biology lessons taught through CLIL (Abzhanova, 2020). This author realized that balancing content and language was a real challenge for teachers in CLIL environments. In fact, despite familiarity of biology teachers with regards to assessment tools and strategies, they were reluctant to adopt criteria-based assessment, as they felt insecure when evaluating students within a CLIL context; the teachers found difficulty in their use and implementation of L1 and L2 evaluation process as they focused on content goals instead of language goals; they tended to give more relevance to content over language, given their limited knowledge about language assessment.

Therefore, taking into account all the settled arguments and the gaps detected, this study is aimed at analyzing whether both content and language learning are promoted through CLIL by using team teaching at university level into a biology-based science subject. The dependent and independent variables of this study are achievement and CLIL-T, respectively.

The purpose of this study is to explore the effectiveness of team-teaching enhanced CLIL on student achievement. The following research question led to this study. Can engaging university students in CLIL increase their achievement in Biology and English language?

The study was conducted at the end of the spring semester of 2020–2021 academic year. The sample of this study consisted of trainee teaching students enrolled in introductory Biology course in a university in Kazakhstan. Convenience sampling method was employed to select students. There were 12 and 13 students in the control and experimental groups, respectively. Almost all the participants were female; one student from control group and two from experimental group were male. They all voluntarily participated in the study. Because of students’ voluntariness, they were not arbitrarily assigned to treatment conditions. Instead, each was allowed to select which treatment they would receive.

The university was situated in an economically moderate suburb but was well resourced. After a successful completion of the undergraduate study program, the students are awarded a bachelor’s degree in biology education. The classes on teaching and learning biology, are typically limited to around 25 students. The classes are subjected to teacher education certification program. The 4-year education comprises four curricular elements: biology content, models of teaching and learning, English and Turkish languages, and practical teaching experience. One of the 2 weekly biology class periods concentrates mainly on the study of conventional biology content. The other weekly class session centers on problem solving and theoretical foundations of learning biology. At least once a month, students take part in the laboratory section of the program. The last year of the program is mainly devoted to pedagogical practice both at the university and at schools. The pedagogical practices provide students with useful skills both in biology and in techniques of teaching and learning (Özdemir et al., 2019).

The biology achievement test consists of 25 multiple-choice questions to assess students’ four different knowledge areas in human digestive system. These areas are introduction to digestive system, steps in digestion, organs of digestive system, digestive secretions, and diseases of digestive system.

There are two criteria for evaluating a standardized test: validity and reliability. For validity, experts’ validation was used, while for reliability Kuder-Richardson reliability index (KR20) and Ferguson’s delta (δ) for reliability and discriminatory power of tests were used in this study. KR20 reliability index is a measure of internal consistency of a whole test when test items are dichotomous (i.e., correct or incorrect answers). Ferguson’s delta is a measure of the discriminatory power of a test. It considers how broadly students’ total scores are distributed over the possible range (Hankins, 2008). The KR-20 values for the biology post-test were found to be 0.881 for experimental group and 0.835 for control group which are very reasonable values for statistical purposes (Maloney et al., 2001). Similarly, the Ferguson’s delta values for the biology post-test was found to be 0.81 for experimental group and 0.89 for the control group.

The English language achievement test consisted of 17 multiple-choice questions and two constructed response items (for assessing writing skill) to assess students’ four different knowledge areas in language. These areas are vocabulary (4 items), comprehension (5 items), grammar (8 items) and writing (2 items). The KR-20 values of the English language post-test were found to be 0.744 for the experimental group and 0.742 for the control group. Similarly, the Ferguson’s delta values for the English language post-test were found to be 0.80 for the experimental group and 0.85 for the control group.

The study was started with a pilot implementation of CLIL-T. The aim of the pilot study was to improve the quality and efficiency of the main study. In addition, it was conducted in order to increase the researchers’ experience with the CLIL-T. A total of 37 students participated in the pilot study who later voluntarily divided into control and experimental groups where 12 students do not accept to participate in the main study. So, the students in this study had already experienced CLIL teaching in “respiratory system” unit.

The main study consisted of a week-long series of meetings (see Table 1) wherein students and the researchers came together to discuss the teaching of digestive system. Overall, six 2-h (including 2 h pilot study) online meetings were held with the participation of freshman year biology students from a university. For meetings the online platform Webex¹ was used. During each meeting, students discussed important concepts in digestive system, solved problems, sing songs related to digestive system, “repeat after me” vocabulary exercises, fill in the blanks, spelling biology related terminology, etc. All meetings were led by the first author and an English language teacher.

In the control group biology lessons on digestive system were provided in English language. No team teaching and CLIL were employed in this group.

After the pilot study, some revisions such as increasing the interaction in English exercises, adding some new activities during presenting the topic, include cultural attributes in the presentations to carry out more attractive lessons, and include exercises into the songs presented during the lesson were included on the lesson plans of CLIL-T. Similarly, some revisions such as increasing the interaction between students and teachers, allowing the Biology and English teacher to interject when needed, having students make presentations, and include videos and animations into the lesson were added to the structure of the study.

During the CLIL-T, cultural issues were brought to the agenda and students were invited to attend the class. Below are some of the dialogues that took place during the meetings that exemplify students’ participation during this program. These excerpts are the transcriptions of video recordings during the meetings.

Similarly, discussions took place about unclear points. Among the many topics taken up in the discussions was the most productive way of stating important concepts like “physical digestion”:

As decided after the pilot study, the English language teacher interjected when language related issues raised:

Along with many language activities “fill in the blanks” was a primary activity. Following is an example of activity led by both teachers.

The data gathered for this study highlight student performance, and in particular assess student knowledge in the digestive system. The assessment of student performance involves pre-tests and post-tests of the basic concepts in the digestive system and their vocabulary, comprehension, grammar and writing skills in English language. The tests were administered to all groups of students online before and after the implementation of the CLIL strategy. Before the testing, the researcher explained the purpose and the importance of the study. The testing was not anonymous. To motivate the students to participate in the study and to take the test seriously, we made three compensations; add incentives to their final grades, to reteach one more biology topic in CLIL method and have a dinner at a restaurant the end of the study.

For the analysis, Jamovi² was used. The t-test was carried out for locating group differences. The Mann-Whitney U-test was used to compare differences among two separate groups when the data is ordinal or when the assumptions of the t-test are not met (MacFarland and Yates, 2016). Finally, a non-parametric ANCOVA was carried out because parametric test conditions were not met for Biology post test scores. To know the extent of the group differences, the effect size through d-Cohen coefficient was calculated.

During CLIL implementation pre and post tests were carried out for both Biology content and English language skills. Table 3 indicates Independent Samples t-test results for both subjects’ pretest sores. The pretest score for control and experimental groups were compared to see if the groups are initially equal in terms of academic achievement both in Biology and in English language.

Table 2 indicates that pretest scores for control and experimental groups in Biology are statistically significantly different while for English language it is not. Since the pretest scores in Biology subject was statistically significantly different, ANCOVA analysis was conducted to adjust students’ initial academic achievements. However, since post test scores in Biology are not normally distributed (see Table 2) a non-parametric ANCOVA was carried out (Cangür et al., 2018). Similarly, post test scores in English language are not normally distributed. Thus, a Mann Whitney U-test was carried out to analyze scores from English language test. The results are shown in Table 4.

The Mann-Whitney U-test showed that there was a significant difference (p < 0.05) between the control group compared to the group receiving the CLIL treatment. The median was 11.5 for the control group compared to 15 (see Table 4) for those receiving the CLIL treatment in learning English language suggesting that the CLIL treatment is more effective than traditional teaching methods.

As seen in Table 5 the pretest means of the control and experimental groups varied significantly? (Mcon = 10.00, Mexp = 13.31) while adjusted means are very close to each other (Mcon = 8.05, Mexp = 7.09). On the other hand, the mean difference between unadjusted posttest scores is 3.31 and that of adjusted scores is only 0.96. There is a substantial effect for the co-variate which suggests that including it has changed the apparent group difference on the dependent variable (post test scores). The significance of the group difference after the adjustment is checked by the non-parametric ANCOVA (see Table 6).

Non-parametric ANCOVA analysis results (using Quade Method) showed a significant difference among groups, i.e., teaching status, in terms of pretest-adjusted posttest values (p = 0.028). According to pretest-adjusted posttest values, control group scores were significantly lower than that of the experimental group. In summary, CLIL teaching was significantly effective in increasing students’ scores in Biology.