Practices of cooperative training in the Amhara Regional State TVET colleges and companies, Ethiopia

Abstract The purpose of this study was to examine the practice of CT, companies’ involvement in CT, TVET Colleges’ and companies’ trainers’ engagement in CT from trainers’ and trainees’ perspectives and to determine whether their perspectives are significantly different. The current study also examined the influence of the involvement or engagement level of company/enterprise, company supervisors and TVET College trainers in CT program implementation. For this purpose, the convergent parallel mixed-methods design was employed. A sample of 207 trainers and 211 trainees were involved in this study. Questionnaire and semi-structured interviews were employed to collect quantitative and qualitative data, respectively. Cronbach’s alpha value for the reliability of the instrument ranged between .74 and .94. The independent-sample t-test, Pearson’s correlation, and regression methods were used to analyze quantitative data. The thematic analysis method was used to analyze qualitative data using research questions as core themes. In this study, we found that the CT practice, company, and TVET trainers’ engagement in CT-related tasks were at a medium level, but the companies’ involvement in CT implementation was low. The TVET trainers and trainees had significantly different views regarding the level of CT practice, companies’ involvement, supervisors’ and trainers’ engagement in CT activities. In conclusion, the companies’ involvement level and TVET trainers’ engagement level showed large influences on CT practice. Based on these findings, we recommended the need for devising binding rules and regulations in a way that enforces accountability amongst all stakeholders of the CT program in the TVET system.


PUBLIC INTEREST STATEMENT
This study attempted to invigorate the TVET colleges' trainers' and trainees' perceptions about cooperative training (CT) practice, companies' level of involvement in CT, and trainers' engagement in CT activities as well as the contributions of companies' level of involvement and trainers' level of engagement to effective CT practice in TVET colleges and companies. The CT program has become a crucial strategy to transform the TVET system and produce workforces having adequate occupational skills. By recognizing this fact, the Ethiopian government has introduced CT as a core learning and teaching approach in the TVET system. The CT system has acknowledged the important role of stakeholders' active engagement in accomplishing their duties and responsibilities. This study has shown the great influence of companies' and trainers' level of engagement on the overall CT practice in TVET colleges and companies. As a result, the government should devise directives that oblige stakeholders to actively participate in CT practice.
implementation was low. The TVET trainers and trainees had significantly different views regarding the level of CT practice, companies' involvement, supervisors' and trainers' engagement in CT activities. In conclusion, the companies' involvement level and TVET trainers' engagement level showed large influences on CT practice. Based on these findings, we recommended the need for devising binding rules and regulations in a way that enforces accountability amongst all stakeholders of the CT program in the TVET system.

Introduction
Cooperative training has played a key role in the effectiveness of the Technical and Vocational Education and Training (TVET, hereunder) program in producing quality workforces equipped with required competencies. Youth well-equipped with occupational knowledge and skills could have better employment opportunities compared to youth less-equipped with occupational knowledge and skills. However, unemployment has become a serious challenge across the world (Frick et al., 2013) as well as low productivity and skills shortages in several economic sectors have been holding back the economic growth of the countries (Kuczera et al., 2018). As Kuczera et al. (2018) documented, cooperative training has become a crucial delivery approach of TVET to produce a quality and more competent workforce that fits into the ever-changing world of work in a way that minimizes the unemployment rate. Likewise, countries that were suffered from high unemployment rates had been looking for ways to improve the transition between school and employment and emphasized the dual vocational training system (Frick et al., 2013). Accordingly, as Frick and colleagues have explained, combining theoretical reflection in vocational schools with practical in-company experience through a dual system of vocational training will facilitate the business world to have skilled workers with real-life training as well as youth's transition into the labor market. In this regard, global experience has shown that Germany's dual vocational training system appeared to be a suitable model to adapt based on the context of a country (Frick et al., 2013). This seems to indicate that the dual system of TVET is well-developed especially in Germany in a way that integrates work-based and schoolbased learning to prepare youth for a successful transition to full-time employment. From this perspective, a major strength of Germany's dual system vocational training is the high degree of engagement and ownership on the part of employers and other social partners (Frick et al., 2013). Recognizing this strength, the Ethiopia's TVET system is derived from Germany's dualtraining system which acknowledges the great value of cooperative training (MoE, 2008(MoE, , 2010. In Ethiopia, the TVET system has envisioned a strong partnership between the public sectors and medium and small enterprises (MSE, hereunder) to deliver a cooperative and in-company training which takes place in the training institution and at the place of work to improve the relevance of the training offered to employers' needs (MoE, 2010(MoE, , 2015. The overriding goal of the TVET program in Ethiopia was: "to produce a lower-and middle-level, competent, motivated, adaptable and innovative workforce, which can contribute to poverty reduction and social and economic development through facilitating demand-driven, quality TVET and transfer of demanded technology" (MoE, 2015, p. 94). Similarly, the national TVET strategy of Ethiopia (MoE, 2008) indicated that the main objective of the training was to create a committed, motivated, adaptable and innovative workforce contributing to poverty reduction as well as social and economic development through facilitating demand-driven, high-quality TVET relevant to all sectors of the economy, at all levels and to all people.
To realize the abovementioned goals and objectives, cooperative training is used as a core TVET delivery approach (MoE, 2008). It is almost "three years since cooperative training is conceived as a major training modality in [Ethiopia's] TVET system, and thousands of trainees and numbers of enterprises and TVET institutions have already been participated and benefited from it" (MoE, 2010, p. 3). To maintain quality cooperative training, the documents of the TVET system in Ethiopia have emphasized great involvement of MSEs as well as a strong linkage between TVET institutions and MSEs as a prominent strategy. Similarly, the TVET strategy of Ethiopia has overvalued the involvement of industries, companies, medium and small enterprises, and other private sectors in cooperative training, which are presumed as employers (MoE, 2008). In this regard, any enterprise capable of providing training facilities such as "working place to trainees, machines for training, tools, materials, and assigning a training coordinator/supervisor" can be a cooperative training partner as long as the enterprise's activities are similar with TVET institution's occupational training (MoE, 2010, p. 6). Hence, the effective practice of cooperative training is beneficial to enterprises, TVET institutions, and trainees. To benefit from cooperative training, all stakeholders (TVET institution, enterprise, trainee, trainer, and supervisor) should effectively undertake their predetermined responsibilities (MoE, 2010). Based on this evidence, it appeared to be important to invigorate the practice of cooperative training in TVET institutions and partner companies from the trainers' and trainees' points of view.
Previous studies regarding cooperative training practice have focused on the involvement of companies and their partnership with TVET colleges (Abebe, 2011;Desalegn Endashaw, 2014;Haileselassie, 2014;Mengistu, 2017;Tolla, 2016). These domestic studies have indicated negligible involvement of companies or enterprises in cooperative training. The limitation of all these studies was that they aggregated the trainers' and trainees' data altogether to show the involvement level of companies or enterprises in cooperative training. The studies did not address whether trainers and trainees had different views regarding companies' level of involvement in cooperative training and companies' accomplishment of responsibilities documented in the national TVET strategy (MoE, 2008). Moreover, the aforementioned researchers did not examine the company supervisors' and TVET trainers' engagement in cooperative training, though they were key stakeholders as documented in Ethiopia's TVET strategy. Nevertheless, evidence on the practice of cooperative training has suggested that the practice of cooperative training in both TVET institutions and companies as well as the engagement level of TVET trainers and companies' supervisors in cooperative training was not adequately investigated from trainers' and trainees' points of view. In addition, the stakeholders' level of involvement in cooperative training could have influence on the effectiveness of cooperative training practice in TVET colleges and companies. Despite the importance of cooperative training program to foster employability of TVET graduates, empirical evidence has indicated that cooperative training program in Ethiopia has suffered from low involvement and reluctance of responsible stakeholders (e.g., Desalegn Endashaw, 2014;Mengistu, 2017;Tolla, 2016). The Ethiopia's government has acknowledged the importance of stakeholders' participation to assure effective cooperative training practice in TVET system since 2008 (MoE, 2010). However, empirical evidence regarding the impact of stakeholders' level of involvement on cooperative training practice has been yet deficient. Accordingly, we examined the influence of the involvement level of companies, TVET trainers and supervisors from companies on the overall cooperative training practice. In current study, we uniquely focused on invigorating the TVET trainer's and trainee's views of the cooperative training practice, companies' involvement in cooperative training, and the trainer's engagement in cooperative training. In addition, the researchers intended to examine the influences of companies' level of involvement in cooperative training and TVET trainers' and company supervisors' level of engagement in undertaking their duties and responsibilities on the overall cooperative training practice in Amhara Regional State. The leading research questions were: (1) To what extent have TVET colleges implemented cooperative training?

Theoretical and conceptual frameworks
The Bandura's social cognitive theory (Bandura, 1988) and Vygotsky's social constructivist theory (Shabani , 2016) have provided theoretical grounding for conceptualizing cooperative training in the TVET system. According to Eun (2018), social cognitive and sociocultural learning theories have been more valid learning theories in explaining and predicting human development and change. These theories provide foundations for understanding why cooperative training leads to youth or trainee development and guidelines for translating trainee's learning into improved performance of industries or other employer organizations.
The organizational application of social cognitive theory explicates psychosocial functioning in terms of triadic reciprocal causation where "behavior, cognitive and other personal factors and environmental events all operate as interacting determinants that influence each other bidirectionally" (Bandura, 1988, p. 275). In this regard, the perspective of social cognitive theory has emphasized the contribution of personal factors to alter interaction among the three determinants to improve the level of organizational functioning (Bandura, 1988). In connection to this notion of social cognitive theory, cooperative training has been a work-based learning model that provides an opportunity for individuals to learn valuable trade skills in the workplace, which in turn will improve the performance and development of enterprises (workplaces). Work-based learning system has constituted the theoretical courses delivered in TVET institutions and the practical courses delivered in enterprises that are related to a particular trade or profession and result in certification in a skilled trade. In addition to cooperative training, work-based learning encompasses an array of work experiences at the workplace that ranges from job shadowing for a few hours or days, internships, and school-based simulated learning to youth apprenticeships that can last for more than a year and are seen as more rigorous (Alfeld et al., 2013;Hamilton & Hamilton, 1997;Stone & Lewis, 2012).
Cooperative training as a work-based learning system serves several purposes including engaging and motivating students by linking academic work to practical real-world experiences, developing workplace skills and competencies, increasing achievement and school (TVET Colleges) completion rates, enabling students to explore career options, meeting employer needs in the labor market and transitioning students into adulthood or career or higher education (Bailey et al., 2004;Darche et al., 2009;Halpern, 2009;Hamilton & Hamilton, 1997;OECD, 2010;Stone & Lewis, 2012;Taylor & Watt-Malcolm, 2007). In the same vein, the cooperative training system in Ethiopia has emanated from Germany's dual-training system and has become a "mode of technical and vocational education and training delivery that combines training in enterprise and institution based on a training plan collaboratively designed and implemented by industries and respective TVET institutions" in a way that benefits all involved influence (MoE, 2010, p. 6). Accordingly, the principle of cooperative training in Ethiopia is that the industry, the trainees, and TVET institutions will gain immediate and long-lasting mutual benefits from cooperative training (MoE, 2010). Bandura (1988) pinpointed that the three relevant aspects of social cognitive theory in organizational functioning: "developing competencies through mastery modellintrengthening people's beliefs in their capabilities so they make better use of their talents and enhancing self-motivation through goal systems" (p. 276). According to social cognitive theory, learning occurs enactively by doing and vicariously by observing, reading, and listening (Schunk, 2012). Schunk further explained that learning takes place "enactively through actual performances and vicariously by observing models, by listening to instructions, and by engaging with print or electronic materials" (Schunk, 2012, p. 159).
Based on the aforementioned theoretical perspective, the trainees' of cooperative TVET training should acquire the required technical vocational competence, carefully perform the tasks entrusted to them as part of their training program at the workplace, and follow the instructions given to them by enterprises, instructors, or any other persons entitled to give them such instructions (MoE, 2010). In cooperative training, as social cognitive theory claims, trainees can enactively learn the best possible job qualifications from the company through practical training (MoE, 2010). Likewise, trainees are expected to learn an adequate level of specific, general, and occupationrelated basic competency from the TVET institutions vicariously by observing instructors, by listening to instructions given by the instructor, and by engaging with modules or supportive materials in their profession (MoE, 2010).
In general, the social cognitive theory provides "explicit guidelines on how to equip people with the competencies, self-regulatory capabilities, and a resilient sense of efficacy that enables them to enhance their psychological well-being and personal accomplishments" (Bandura, 1988, p. 299). In connection with this theory, in cooperative training, trainees acquire knowledge by observing TVET trainers or company supervisors and peer models at the workplace. From a social cognitive theory perspective, the reciprocal interaction between trainees' personal and TVET institutions' or enterprises' environmental factors can be seen clearly with social and self-variables. "Social (environmental) factors can affect many self (personal) variables, such as learners' goals, self-efficacy, outcome expectations, attributions, self-evaluations of learning progress, and self-regulatory processes" (Schunk, 2012, p. 148). In line with this notion, Taylor and Watt-Malcolm (2007) noted that the learning environment is important for overall learning outcomes as the different learning environments-school and work-provide a broad range of experiences for apprenticeships.
In the workplace, students have the opportunity to learn through hands-on applications of theory and to develop other employability skills such as critical thinking, creativity, teamwork, and interpersonal skills that are not readily learned at school. Taylor and Watt-Malcolm (2007) studied the opportunities and limits of learning among apprentices in Canada and indicated that the students believed that the support and resources available to them were helpful in their development. Furthermore, the dynamics of the workplace-exposed students to everyday work life and helped them to understand issues such as organizational culture, power relations at work, and discrimination. Trainees' exposure to the work environment has helped them develop critical skills such as teamwork, critical and higher-order thinking, time management, and communication and interpersonal skills, all of which employers cited as essential skills for success (Halpern, 2009;Woods, 2012). Although evidence has indicated that the workplaces have provided trainees with opportunities for growth and development through fostering their participation in multiple communities of practice and possibilities of feedback, some apprenticeships programs offered a narrow range of training that may not be relevant to the students' knowledge bases, thereby limiting the progress of the apprentices (Onstenk et al., 2007;Taylor & Freeman, 2011). Based on this evidence, we argued that the enterprises' (as future workplaces) level of involvement and duty accomplishment could have a significant influence on the effective practice of cooperative training as a TVET delivery approach. Like Bandura's social cognitive theory, Vygotsky's social constructivist theory also focuses on the social context of human learning. As with the social cognitive theory, the sociocultural theory posits a reciprocal causation and bidirectional influences between individuals and their environment (Eun, 2018). According to Eun (2018), both theories acknowledge the importance of the dynamic and changing relationship between the person and his/her environment. The main difference between the two theories lies in the mechanism of development they provide and the focus on specific processes within that developmental mechanism, indicating their complementary role in understanding human learning and development in the social context (Eun, 2018). In this regard, the current study was also guided by the assumptions of social constructivist theory in human learning and development. According to Stevens (2008), the previous studies on cooperative learning have used Lev Vygotsky's sociocultural theory as a foundation for the theoretical framework to explain the academic effects of cooperation. This theory proponents claimed that human development and learning occur when "individuals internalize new information and skills within the proximal zone of development" through a scaffolding system (Stevens, 2008, p. 189). Vygotsky's sociocultural theory emphasizes social interaction as a fundamental aspect of successful cognitive and intellectual growth (Eun, 2018;Pritchard & Woollard, 2010;Shabani & Ewing, 2016). The proponents of this theory have argued that the individual's cognitive development and learning potential depends on the transition across the zone of proximal development (ZPD, hereunder;Pritchard & Woollard, 2010). In Vygotsky's view, the ZPD is "the distance between the actual developmental level where independent problem-solving occurs and the potential developmental level where problem-solving can occur with the guidance of an adult or more knowledgeable peer" (Larkin, 2008, p. 863). As Pritchard and Woollard (2010) point out, the ZPD is a notional area of understanding or cognitive development that is close to but just beyond a learner's current level of understanding. The key to the ZPD is social interaction and collaborative problem-solving (Larkin, 2008) as the ZPD is the level of potential development settled "through problem-solving under adult guidance or in collaboration with more capable peers" (Pritchard & Woollard, 2010).
In particular, Vygotsky's theory states that for complex cognitive tasks, learners benefit from interactions with more competent peers, like those interactions in cooperative learning. The theory suggests that interaction facilitates the internalization of newly learned skills. This type of interaction has also been called a cognitive apprenticeship, where learning occurs while engaging in academic interactions with a more competent peer or adult. The theory offers an understanding of the broader socio-cultural context of cooperative learning (Stevens, 2008). From this perspective, cooperative TVET training may have great benefits to trainees. In cooperative training, the trainees can get experience in competencies, values, and attitudes of real-world environment (MoE, 2010).
The instructional approaches in cooperative training may include large group presentations and discussions, microteaching, workshops, seminars, demonstrations, simulations, etc. Training, which is generally employed as large group-based activities, has the advantage of reaching a large number of participants in a single session as well as being cost-effective (Eun, 2018;Shabani & Ewing, 2016). From a social cognitive theory perspective, as Eun (2018) pointed out, training has allowed trainees with enactive mastery experiences (e.g., microteaching) as well as vicarious experiences (e.g., demonstrations) that were important sources of trainees' professional selfefficacy. The opportunity to engage in guided mastery experiences such as microteaching has allowed teachers to feel efficacious in their teaching through facilitating successful performance (Eun, 2018). As Eun (2018) and Shabani and Ewing (2016) indicated, observing successful models has served as the vicarious source of efficacy enhancement by allowing teachers to possess selfefficacy belief to perform successfully what the models have performed. As a result, the impact on self-efficacy has become greater when individuals perceived their models to be similar in terms of teaching capabilities (Shabani & Ewing, 2016;Stevens, 2008). Moreover, models have served as powerful sources of observational learning in which demonstrations of effective instructional strategies were communicated to participating teachers (Shabani & Ewing, 2016).
On the other hand, the sociocultural theory emphasizes the social interaction that occurs during the training sessions as the main mechanism for trainee development (Eun, 2018;Shabani & Ewing, 2016). Furthermore, the sociocultural theory argues for the use of various mediators to guide the internalization process. The more competent (e.g., the trainer or supervisor) should mediate the development of the less competent (i.e., the trainer) via technical tools and other symbolic means. According to Eun (2018), mere observation of a successful performance did not lead to development and learning unless it is accompanied by guided interaction between the trainer or supervisor and the trainee. In cooperative training, as sociocultural theory advocates suggested, TVET trainers as a scaffolder have the responsibility to give awareness or orientation to the trainees, prepare a training plan, follow-up and record the results of trainees, organize access to workplace, and establish guidelines and standards for the CT program (MoE, 2010). As scaffolders, they are not only trainers but also facilitators of enterprise-based training having the responsibility of linking training plan with the work, assuring that a trainee acquired the required competence, and preparing performance or training achievement report to the industry and the TVET institutions (MoE, 2010). These ideas of cooperative training coincide with sociocultural theory's concept of the zone of proximal development. Taken together, the proponents of the aforementioned theories have suggested a strong social interaction between responsible parties and high involvement and accomplishment of their duties and responsibilities to assure effective practice of cooperative training in a way that benefits all participants.

Design of the study
To answer the research questions, the researchers used a mixed methods approach (Teddlie & Tashakkori, 2009), which is a procedure for collecting, analyzing, and integrating both quantitative and qualitative data at some stage of the research process within a single study (Creswell, 2012). The rationale for mixing both types of data is that neither quantitative nor qualitative methods are sufficient by themselves to capture the complex issue of cooperative training practice, companies involvement, and trainers' engagement in cooperative training in diverse contexts of TVET colleges.
As this study followed the mixed methods research approach that is rooted in the paradigm of pragmatism, the researchers used a convergent parallel mixed research design. The rationale behind choosing this research design was because the researchers intended to concurrently collect quantitative data using survey questionnaire and qualitative data using semi-structured interviews, merge the data, and use the results (Creswell, 2012) to understand the practice of cooperative training in the Amhara Regional State TVET colleges and companies. Accordingly, the rationale for choosing this design is that "one data collection form supplies strengths to offset the weaknesses of the other form, and that a more complete understanding of a research problem results from collecting both quantitative and qualitative data" (Creswell, 2012, p. 540).

Participants of the study
The sampling procedure consisted of multiple steps. In the first step, the researchers randomly selected four Technical and Vocational Education Training (TVET) Polytechnic Colleges from the total of twelve Polytechnic TVET Colleges in Amhara National Regional State. Secondly, we randomly selected departments delivering Level II and above TVET programs using a lottery system. Thirdly, a total sample of 418 trainees and trainers had involved in the current study. From this total sample size, 211 were trainees learning in level II and above programs, whereas 207 were trainers working in those four TVET Colleges. Furthermore, we collected qualitative data from purposively selected three TVET College Deans, three TVET College Training Unit Coordinators or Department heads, three Company/Enterprise managers, one Directorate Director, and one Trainees Development Expert working in Amhara Regional State Technical Vocational and Development Bureau as they were core responsible bodies in cooperative training. To protect participants' rights and address ethical concerns, the researchers informed that their participation is voluntary and anonymous, and affirmed that their response would be used only for academic purposes.

Questionnaire
In the current study, a structured questionnaire was used to collect quantitative data. The overall instruments adapted from four previous domestic studies (Abebe, 2011;Demessew, 2012;Desalegn Endashaw, 2014;Hailemariam, 2018) were used to collect quantitative data, because these studies had established the reliability and validity of the questionnaire items through rigorous procedures. Specifically, the researchers adapted 15 items from Demessew (2012), 17 items from Desalegn Endashaw (2014), 13 items from Hailemariam (2018), and 11 items from Abebe (2011). Next to the adaptation of the questionnaire from the above sources, the researchers asked one professional who was knowledgeable and a researcher in the Ethiopian TVET system particularly cooperative training to review whether the questionnaire items were clear, unbiased, concise, and provided the same meaning for all readers. In addition, we conducted a pilot test of the questionnaire on 30 TVET trainers and 35 trainees selected randomly from TVET colleges who did not involve in the final study. Based on pilot test evidence, we corrected the typographical errors within the questionnaire.
In this study, cooperative training (CT, hereunder) practice was measured with 18 five-point Likert-scale items having scores that ranged from 1 (strongly disagree) to 5 (strongly agree). The items, for example, included: "Our TVET College and industries in the area closely work together in cooperative apprenticeship training; Our College has established a platform for cooperative apprenticeship training." The Cronbach's alpha (α) coefficient was used to assess the reliability of subscales. In this regard, Cronbach's alpha value for CT practice measurement subscale was .87 showing high reliability or internal consistency among items measured CT practice.
In the current study, we hypothesized that the involvement level of a company would influence CT implementation. Accordingly, the involvement level of Companies in CT was measured using 24 fivepoint Likert-type scale items with scores ranging from 1(very low) to 5 (very high). Sample items used to measure the Companies' level of involvement included: "Willingness of companies to receive trainees and providing cooperative-apprenticeship training; Companies' evaluation of trainees based on the occupational standard and unit of competency." The alpha value for scale reliability was .94 indicating excellent reliability and high internal consistency among items. Likewise, in the current study, we assumed that the company trainers' level of engagement in CT would affect the effective implementation of the CT program. This variable was measured using six 5-point Likert scale items with scores ranged from 1(strongly disagree) to 5 (strongly agree). The items, for example, included: "Company trainers plan, organize, and implement the apprenticeship training in the enterprise. Company trainers ensure that the apprentices work and learn in the occupation as specified in the training contract . . .." The reliability coefficient for this scale was the Cronbach's alpha value of .84 showing great internal consistency among items. Moreover, TVET trainers' level of engagement was presumed to influence CT program implementation. This variable was assessed using eight 5-point Likert scale items with scores ranged between 1(strongly disagree) and 5 (strongly agree). The sample items used to measure TVET trainers' engagement in CT included: "TVET Trainers have regularly monitored the learning progress and skills development of apprentices, and . . .. ensured that apprentices understand the theory and key concepts required for the trade." The Cronbach's alpha value for scale reliability was .87 indicating great internal consistency among items. As shown in Table 1, itemtotal correlation coefficients were used to determine whether each item had measured what it was intended to measure.

Semi-structured interview
In addition to the close-ended questionnaire, the researchers held the personal interview to collect qualitative data by reading semi-structured interview questions to the respondent in a face-to-face setting and recording the answers. All interviewees were involved in the current study voluntarily.

Quantitative data analysis
The participants' background characteristics information was analyzed using frequency and percentage. T-test, Pearson's correlation, and regression analysis methods were used to analyze quantitative data collected to answer the research questions.

Qualitative data analysis
The researchers transcribed audio data into texts and coded them using interview guiding questions as core themes. As a result, we used the thematic analysis method to analyze qualitative data using research questions as core themes. Finally, we triangulated quantitative results with qualitative, though quantitative results had more emphasis. Qualitative data had a supportive role in the current study.

Assumption testing
Before conducting regression analysis, we tested the assumptions of multicollinearity, linearity, normality, and homoscedasticity. Correlation coefficients among variables, variance inflation factor (VIF) values, and tolerance values were used to assess multicollinearity assumption. In connection to this, we used Brown's (2015) criteria that suggested the correlation coefficient above .85, the VIF value greater than 10, and tolerance value less than .10 as indicators of multicollinearity concern. As a result, the correlation coefficients were between .477 and .679 (<.85), the VIF values were between 1.464 and 2.101 (<10), and the tolerance values were between .476 and .683 (>.10), which indicated the absence of multicollinearity concern. Likewise, the assumption of normality of the distribution was examined using Histogram, which showed normal distribution of the regression standardized residuals (see, Figure A1). The Normal P-P Plot of regression standardized residuals also depicted the linear relationships establishing linearity assumption (see, Figure A2). Moreover, the scatterplot showed that the assumption of homoscedasticity was established (see, Figure A3). This evidence indicated that the assumptions of multiple linear regressions were not violated.

Results for CT practice
The one major purpose of this study was to examine the practice of CT, companies' involvement in CT, TVET College, and company trainers' engagement in CT from trainers' and trainees' perspectives and to determine whether their perspectives were significantly different. In this regard, both trainers (M = 3.29) and trainees (M = 3.46) reported better practice of cooperative training above the expected mean value (M = 3), which resulted in the higher overall average practice of CT (M = 3.38). On the other hand, an independent t-test was carried out comparing the differences in the perceptions of CT implementation between TVET trainers and trainees. Levene's test for equality of variances showed a nonsignificant F value which suggested equal variances assumption (F(416) = 1.492, p = .223). Accordingly, results showed that the TVET trainees reported relatively better CT practice (M = 3.46, SD = .66) compared to trainers (M = 3.29, SD = .68). The mean difference (0.17) was statistically significant, t(416) = 2.639, p = .009. Likewise, the Cohen's d value (.254) suggested a small to a medium effect size of the aforementioned mean difference, based on Cohen's (1988) criteria. Results of each CT practice item revealed that both trainers and trainees reported better practice of TVET College regarding closely working together with industries or companies, contributing to the design of CT programs for selected occupations, supporting the partner enterprises in establishing effective planning, and following the overall objectives of CT, establishing a platform for cooperative training, and supporting the partner industries/enterprises in planning CT. Other good practices reported by both trainers and trainees include: the TVET College training unit coordinators (department heads) are working in interaction and coordination with the external partners, trainers have possessed many roles in improving CT management, there is an appropriate guideline for CT in TVET College, trainees are always provided with cooperative training guideline, enough time is allotted for CT, trainers always supervise the occupational training at the enterprises, trainees readiness and interest to go out for CT, and assigning trainees on the right work position during CT. On the other hand, both trainers and trainees reported relatively low practice of TVET College heads on frequently visiting industries/ enterprises and delivering professional guidance and counseling services during CT. However, the trainees and trainers had different views on some aspects of CT practice. In this regard, the trainees reported better practice of TVET colleges in giving recognition to companies that host trainees and objective evaluation and assessment of trainees by the industries compared to trainers' responses. In general, the trainees' mean response for each item ranged from 2.94 to 4.02, whereas the trainers' mean response ranged between 2.86 and 3.62. In addition to quantitative data, the Polytechnic College deans and vice deans who shouldered more responsibility for CT practice were asked to describe their experience regarding the status quo of CT practice in TVET colleges and companies. As a result, the qualitative data revealed that the implementation of CT was not so effective because most enterprises were not voluntarily admitting the trainees due to fear of trainees breaking of the machines. This participant also reported that even TVET colleges themselves assign trainees after they learn more than two units of competencies because of the lack of companies or enterprises in their vicinity. According to Vice Dean from one Polytechnic College: The CT implementation is different according to the occupation. We plan to implement but, in reality, we failed to implant CT effectively due to different problems. The strategy gives direction to implement the cooperative training 70% in the company and 30% of the theory given by Colleges but the reality is the reverse.

Results for companies' involvement in CT
Regarding companies' level of involvement in CT, results revealed that TVET trainees had high average perception (M = 3.01, SD = .78) compared to trainers' mean perception (M = 2.80, SD = .83). The mean difference (0.21) was statistically significant, t(416) = 2.604, p = .010. As Cohen's d value (.261) indicated, the mean difference had a small to a medium effect size. The trainees reported moderate involvement of companies in CT while trainers reported low involvement, which resulted in overall low involvement of companies in CT (M = 2.90) compared to the expected mean value (3).
In addition to quantitative data, Trainees Development Expert was asked to describe his experience regarding the companies and industries involvement in CT implementation. As a result, Trainees Development Expert reported that the most of Industries and Enterprise were not voluntary to work with TVET colleges for the effectiveness of CT implementation. According to this participant, industries and enterprises complained that the assigned trainees will damage the quality of our production and break their production machines.

Results for company and TVET college trainers' engagement
The current study also assessed whether TVET trainees and trainers had different views regarding company supervisors' engagement in CT. Thus, the trainees reported relatively high average  Cohen's d value (.23) showed that the effect size of the mean difference ranged from small to medium effect.
In addition to quantitative data, Trainees Development Expert asked to describe his experience regarding the status quo of company supervisors' professional engagement in CT implementation. Hence, the qualitative data from Trainees Development Expert indicated: The duties of company supervisors are mentoring, coaching, and deliver training with good methodology. But most of them do not have pedagogical skills to mentor and coach as well as to deliver the training, [though] some have adequate skills to accomplish their roles in CT implementation. In addition, they are discouraged that they need some incentives. Their education status is not as reported to the Bureau. They are less qualified and are not certified. They are traditional practitioners.
This highlights that the company supervisors have had the limitations not only in engaging on CT implementation but also they have had professional competency gaps to train up to the standard.
Moreover, results revealed that the trainees reported relatively low average engagement of TVET trainers' in CT (M = 3.14, SD = .85) compared to trainers' mean response (M = 3.45, SD = .85), which resulted in the overall moderate and above mean engagement of TVET trainers' in CT (M = 3.29) compared with the expected value (3). The mean difference (−0.31) was statistically significant at p < .001, t(416) = −3.749. Furthermore, the Cohen's d value was .365 indicating a small to medium size of the mean difference. In addition, qualitative results from the semi-structured interview also revealed that most TVET college trainers do not accomplish their duties. Therefore, the involvement of College trainers in the implementation of CT is not more satisfactory, even though few trainers work according to CT guidelines.

Influences of companies' involvement and trainers' engagement on CT practice
As shown in Table 3, the practice of CT program had significantly positive and high associations with companies' level of involvement (r = .651), company trainers' level of engagement in undertaking CT activities (r = .546), and TVET College trainers' level of engagement in undertaking CT activities (r = .547) at p < .001. The proportions of variance (r 2 ) revealed that 42.38%, 29.81%, and 29.92% of the variances in CT implementation were, respectively, attributable to Company Involvement, Company Supervisors Engagement, and TVET College Trainers Engagement, which indicated large effect size in linear correlations based on Cohen's (1988) criteria. Similarly, the involvement level of a Company in CT had a highly positive and significant correlation with Company Trainers Engagement (r = .679) at p < .001. Thus, the 46.10% proportion of variance (r 2 = .4610) indicated a large correlational effect size. Similarly, the Company Trainers Engagement highly and significantly correlated to the TVET College Trainers Engagement in CT undertakings (r = .544, p < .001). The 29.59% proportion of variance (r 2 = .2959) indicated a large correlational effect size, as suggested in Cohen (1988). Moreover, the Company Involvement in CT had a positive and significant correlation with the TVET College Trainers Engagement in CT undertakings (r = .477, p < .001). The 22.75% proportion of variance (r 2 = .2275) indicated a medium to large correlational effect size.
The current study also examined the influence of the involvement or engagement level of company/enterprise, company supervisors, and TVET College trainers in CT program implementation. In connection to this study purpose, standard multiple regression analysis was conducted in SPSS 23. As a result, the squared multiple correlations (R 2 ) of .495 indicated that the three explanatory variables, Companies Involvement level, the engagement level of Company Trainers, and TVET Trainers, together explained 49.5% of the variance in the implementation level of the CT program. Likewise, Cohen's f 2 value of 0.98 shows a large effect size as suggested in Cohen (1988) and Keith (2019). The ANOVA table also showed that the overall regression equation was statistically significant, F(3, 414) = 137.119, p < .001. The unstandardized regression coefficients showed that for each additional involvement or engagement of the Company, Company Supervisor, and TVET Trainer, CT program implementation might improve by .384, .059, and .222, respectively (see, Table 3). The standardized regression coefficients showed that the involvement level of Company or Enterprise (β = .463) had a relatively stronger influence in CT program implementation compared to the engagement level of TVET Trainers (β = .222) and Company Supervisors' engagement (β = .077). In other words, each standard deviation increase in Company involvement will lead to a .463 standard deviation betterment in CT implementation, whereas a standard deviation increase in the engagement level of Company Supervisors and TVET Trainers will result in .077 and .222 standard deviations in the betterment of CT implementation, respectively. Based on Keith's (2019) rules of thumb for the magnitude of effects, the Companies' involvement level and TVET Trainers' engagement level showed large influences in CT program implementation, whereas the engagement level of Company supervisors exhibited moderate effect in CT practice.

Discussion
The first research question has investigated the extent to which TVET polytechnic colleges in Amhara national regional state, Ethiopia, have been implementing cooperative training. One sample t test results revealed there is above average CT practice but below the expected standard set by ministry of education.Put it briefly, results showed low practice of TVET polytechnic college heads on visiting industries/enterprises that engaged in CT and delivering professional guidance and counseling services for trainees during CT. Likewise, qualitative data showed that the effectiveness of CT practice in TVET polytechnic colleges is not to the expected level. CT practice has been reversed though the national TVET strategy documented to own 70% of cooperative training in the company and 30% of training (theoretical part) in TVET colleges. As qualitative results showed, the participants commonly mentioned involuntariness of enterprises to admit the trainees and lack of companies or enterprises in the surrounding as major reasons for ineffective practice of CT. Taken together, the findings in the current study have showed that there is a gap in CT practice. These study findings substantiate the results of previous studies (Abebe, 2011;Desalegn Endashaw, 2014;Mengistu, 2017;Tolla, 2016) which indicated low or moderate practice of cooperative TVET training. Abebe's (2011) study found a very limited relation between TVET colleges and industries to exchange best practices in cooperative training. This study further indicated that TVET colleges and industries did not plan and evaluate cooperative training practices together, and TVET colleges did not provide any service to industries involved in CT (Abebe, 2011). In his study, Desalegn Endashaw (2014) attested that both trainers and trainees have rated the overall issues related to cooperative training as moderate or low indicating that cooperative training was not carried out properly. Furthermore, Tolla's (2016) study revealed a medium practice of CT in TVET Colleges. This study also showed that the weak collaborative work culture among TVET colleges and that of industries has limited the opportunity of exploiting the potential of all stakeholders for the successful implementation of cooperative training.
The second research question aimed at uncovering the companies level of involvement in implementing cooperative training. Accordingly, the results of this research question revealed low involvement of companies and enterprises in CT implementation. Put it briefly, companies and enterprises refused to involve in CT implementation due to different reasons, especially fear of damage on production machines and materials that might be caused by the trainees. These findings are consistent with previous studies (Abebe, 2011;Desalegn Endashaw, 2014;Mengistu, 2017;Tolla, 2016) which found low involvement of industries or enterprises in CT implementation. Zelalem's study (Abebe, 2011) showed that the enterprises or industries are reluctant to take part more genuinely in the cooperative training. Likewise, Desalegn Endashaw (2014) found low involvement of companies or enterprise in planning, implementing, and evaluating the cooperative training. In addition, Melaku's study (Mengistu, 2017) revealed that the industries or enterprises' involvement in CT is very low in all TVET institutions, regardless of enterprises' availability in the area. Moreover, Tolla (2016) found out that most of the concerned stakeholders were not actively and strongly involved in supporting, implementing, and follow-up the cooperative training.
In the current study, the third research question aimed at uncovering the extent to which TVET and Company trainers have engaged in undertaking their duties and responsibilities of cooperative training. In this regard, results showed low engagement of TVET trainers and company supervisors in undertaking their duties and responsibilities of cooperative training to promote effective CT practice. This finding is in line with previous studies' findings (Abebe, 2011;Tolla, 2016) which indicated some limitations of TVET College trainers in supporting CT in industries or enterprises. Abebe (2011) found that there was a lack of follow-up from the TVET College trainers. This study showed that the cooperative training in the industries was not adequately supported by the TVET College trainers. Similarly, a study by Tolla (2016) showed that the cooperative training providing enterprises or companies were rarely visited by the TVET College trainers and TVET vocational counselors. Furthermore, this study result substantiates the previous study (Tolla, 2016) which indicated some limitations in company supervisors' engagement in CT because of their need for incentives during cooperative training time, which in turn impeded the implementation of cooperative training in the area.
Furthermore, the fourth research question investigated whether companies' level of involvement, TVET trainers' level of engagement, and company trainers' level of engagement have significant influence on the overall CT practice. According to the findings of the multiple regression analysis, it was indicated that the CT practice is highly influenced by the companies' involvement level and TVET Trainers' level of engagement in accomplishing their prescribed tasks. However, it was found that school facilities dimension did not have a remarkable effect on the EFL teachers' job satisfaction. However, as the findings revealed, CT practice is moderately influenced by the companies' trainers' level of engagement in undertaking their prescribed tasks. These finding support the notion of important role of the stakeholders' participation for effective practice of CT documented in national TVET strategy and cooperative training handbook (MoE, 2008(MoE, , 2010. The findings are also in line with the findings of Desalegn Endashaw's (2014) study which indicated the influences of the adequacy of time, the capability of supervisors and their engagement in visiting trainees, similarities of training instrument/ machines in TVET and companies on CT program implementation.

Conclusion
The findings of this study indicate that the TVET Colleges and Companies have had a moderate practice of cooperative training (CT). To elevate an effective practice of CT in TVET Colleges and companies, instituting a strong involvement of the companies, professional engagement of trainers and supervisors is found to be very crucial. A weak link between TVET Colleges and companies would impede the practice of cooperative training, which in turn may have a great impact on developing youth's employability skills. Boosting the trainers and supervisors professional engagement in cooperative training is also important to equip trainees with adequate and up-to-date professional competencies. However, the evidence in the current study has shown that the engagement level of the trainers and supervisors in cooperative training is below the expected level, which may result in less qualified TVET graduates. Moreover, the companies' involvement level and TVET trainers' engagement level have substantial influence on CT practice. Based on these findings, we recommend devising binding rules and regulations in a way that enforces accountability among all stakeholders of the CT program in the TVET system of Ethiopia in general. To this end, the current study has contributed to the knowledge of cooperative training literature by providing empirical evidence regarding the stakeholders' level of accomplishing their CT duties and its impact on the effective practice of CT in a dualapproach TVET system.