What Does Research Tell Us about Trends in Dissertations on PBL?

The aim of this study is to investigate the research trends in dissertations on PBL from 2002 to 2015 in Turkey. For this purpose, the master’s and doctorate dissertations in the National Thesis Database of Council of Higher Education (CoHE) were selected for rigorous content analysis. The analysis was utilized to classify the type of study, the learning domains, research methods, research designs, research domains, subject groups, group sizes, length of treatments and data collection instruments. For analysis purposes, frequencies and percentages were calculated for each category. Additionally, the results obtained in these dissertations were examined and reported separately with regards to cognitive, affective and conative domains. According to analysis results; science is the most favorable learning domain, quantitative methodology is more utilized as compared to qualitative and quantitative methods, cognition and affect followed by cognition only are the mostly utilized research domains, subject groups are predominantly at elementary education level and secondly from teacher education, studies preferred to have a group size of 41-80, the treatments on PBL research last mostly four to six weeks, and finally majority of the studies use achievement tests and attitude scales as instruments of measurement.


Introduction
The Turkish education system, which is under the supervision and control of the Ministry of National Education (MoNE), consists of three levels which make up 12 years of compulsory education: The elementary school education (grades 1-4), the middle school education (grades [5][6][7][8], and the high school education (grades [9][10][11][12]. Students who graduate from high schools have the right to continue to higher education, which is governed by the Council of Higher Education (CoHE) and includes universities, institutes of technology, vocational schools, and other higher education institutions such as military academies and colleges.
Starting in the 2005-2006 Academic Year, constructivist teaching and learning was introduced into elementary and secondary education and curricula were changed accordingly. The main aim was to assure that each pupil's social, psychological and moral development reach to the desired level and to give students a chance to apply what they have learned, to urge them use higher order thinking skills more and to train students as lifelong learners (MoNE, www.meb.gov.tr). With the onset of constructivism in Turkish education system, research started to focus on methodologies such as project-based learning, inquiry-based learning, case-based learning, etc., which embody the principles of constructivist teaching and learning [1].
Problem-based learning (PBL) is one of those constructivist teaching and learning methodologies implemented in schools and investigated in current research in Turkey. Literature review on PBL studies suggests that the majority of PBL research and practice is in higher education, specifically medical education, but recently more research has been published in different disciplines and levels [2,3]. Outside Turkey, many meta-analyses and reviews [3,4,5,6,7,8,9,10,11,12] have been conducted to analyze the effectiveness of PBL. The only meta-analysis on PBL research in Turkey, which could be accessed by the author as of October 2014, belongs to Üstün [13]. It is quite obvious that more review studies are needed to analyze PBL's appropriateness, its application and practices, its strengths and weaknesses, and its potential influence on today's education [14], specifically in Turkey. Besides, as more disciplines and educational settings (e.g. elementary education, higher education) began experimenting with PBL, it became necessary to review the outcomes reported in these Universal Journal of Educational Research 5(6): 972-988, 2017 973 studies. Therefore, reviewing the research trends in PBL studies may help the researchers in related fields to identify their research interests and designs.
In this respect, this study intends to investigate the research trends in dissertations on PBL from 2002 to 2015 in Turkey. For this particular purpose, the master's and doctorate dissertations in the National Thesis Database of CoHE were selected for analysis. The research questions addressed by this paper include: 1. What type of studies, learning domains, research methods, research designs, research domains, subject groups, group sizes, length of treatments and data collection instruments were used in the selected theses from 2002 to 2015? 2. What were the findings of these studies regarding cognitive, affective and conative domains?
Apart from the research questions above, the author aims to present a critical analysis of the methodologies used and provide implications for further practice and research on PBL in the discussion part.

Problem-Based Learning (PBL)
PBL is a student-centered, inquiry-based approach, which has its roots in constructivist epistemology [15] and begins with an ill-structured problem with possible multiple solutions. Teachers are facilitators, whereas learners actively construct knowledge by defining learning goals, seeking information to build upon prior knowledge, reflecting on the learning process, and participating actively in cooperative/collaborative learning groups [16,17,18]. Barrows [19] identifies the following objectives of problem-based learning:  Structuring of knowledge for use in clinical contexts,  The development of an effective clinical reasoning process,  The development of effective self-directed learning (SDL) skills,  Increased motivation for learning (p. 481-482).
Among the educational objectives and goals listed above, value is placed on PBL's promotion of learning outcomes in the cognitive, affective, and conative domains. Huitt and Cain [20] define;  Cognition as a reference to the process of coming to know and understand; of encoding, perceiving, storing, processing, and retrieving information. It is generally associated with the question of "what" (e.g., what happened, what is going on now, what is the meaning of that information).  Affect as a reference to the emotional interpretation of perceptions, information, or knowledge. It is generally associated with one's attachment (positive or negative) to people, objects, ideas, etc. and is associated with the question "How do I feel about this knowledge or information?".
 Conation as a reference to the connection of knowledge and affect to behavior and is associated with the issue of "why." It is the personal, intentional, planned, deliberate, goal-oriented, or striving component of motivation, the proactive (as opposed to reactive or habitual) aspect of behavior.
The references to cognitive, affective, and conative domains of learning above presents PBL's potential to promote not only what is learned, but also put emphasis on feelings on what is learned and the learning experience itself, together with willingness and desire to learn. Investigating the dissertations with regards to the findings related to these domains would help practitioners and researchers in their studies.

Review of PBL Research
Review studies on PBL are predominantly in medical education; however, the adoption of this particular method has extended into elementary schools, secondary schools, universities of different disciplines, and professional schools. In his review, Savery [2] gives a list of several studies, which illustrates the multiple contexts where PBL instructional approach is being utilized.
Albanese and Mitchell [4] categorized and listed the qualitative results of studies in medical education from 1972 to 1993. The main results of this review were that students held more positive views on PBL than traditional instruction and PBL graduates performed as well and sometimes better than traditional graduates on clinical examinations and faculty evaluations. However, PBL students scored lower on basic science examinations in comparison with their conventionally trained counterparts.
More recently, Dochy, Segers, Van den Bossche and Gijbels [7] reviewed studies on PBL beyond the domain of medical education. The results of their meta-analysis suggested that PBL had statistically and practically significant positive effects on students' knowledge application. The findings on knowledge base of students led the authors to conclude that students in PBL performed at least as well as students in conventional learning environments. The last remarkable finding related to the retention period was that students in PBL gained slightly less knowledge, but remembered more of the acquired knowledge.
In their meta-analysis, Gijbels, Dochy, Van den Bossche and Segers [8] investigated the assessment of three levels of the knowledge structure as main independent variable in PBL settings: (a) understanding of concepts, (b) understanding of the principles that link concepts and (c) linking of concepts and principles to conditions and procedures for application. They found that PBL had the most positive effects when the focal constructs being assessed were at the level of understanding principles that link concepts. The results suggested that the implications of assessment must be considered in examining the effects of PBL and probably in all comparative education research.
In their synthesis Vernon and Blake [12], analyzed all available research from 1970 to 1992. In their comparison of PBL with more traditional methods of medical education, they found that PBL was superior with respect to students' attitudes and opinions about their programs and with respect to measures of students' clinical performance. They also reported that contrary to the previous review findings, the scores of PBL students on various tests of factual and clinical knowledge were not significantly different from the scores of conventionally taught students. However, the conventionally taught students performed significantly better than their PBL counterparts on the National Board of Medical Examiners (NBME) Step 1, which focuses on concepts of basic science that are important to the practice of medicine, with special emphasis on principles and mechanisms underlying health, disease and methods of therapy.
Berkson [5] also searched for evidence of the effectiveness of PBL in medical education. The results showed no distinction between graduates of PBL and traditional instruction, considering problem solving, imparting knowledge, motivation to learn medical science, promoting self-directed learning skills, and student and faculty satisfaction. However, PBL was found to be stressful for both students and faculty.
In their meta-analytic study, Kalaian, Mullan and Kasim [9] indicated that PBL curricula showed higher performance on standardized profession-wide clinical science outcome measures, but lower on basic science outcome measures. Additionally, their findings suggested that schools which had a past history in PBL utilization, which had developed teaching expertise and learning materials for PBL instruction were able to score on basic science outcome measures as high as those schools which followed more traditional methods of instruction.
In his review Colliver [6] examined the educational superiority of PBL in comparison to traditional approaches. He concluded that there was no convincing evidence that PBL improved students' knowledge base or clinical performance. Nevertheless, he stated that PBL might provide a more challenging, motivating, and enjoyable approach to medical education.
In her study, Leary [10] conducted a meta-analysis across all disciplines to examine the extent to which problem-based learning engendered self-directed learning compared to a lecture-based approach. From her synthesis, she showed that PBL promoted conative and affective skills in self-directed learning.
In their review of empirical studies that examined Self-Directed Learning (SDL) and Self-Regulated Learning (SRL) in PBL environments, Loyens, Magda and Rikers [11] gave evidence on PBL's positive influence on the development of SDL and SRL, though the studies showed mixed results. They also concluded that more research was needed to better understand how, when, and why PBL fostered development of SDL and SRL.
In his meta-analysis of Turkish studies, Üstün [13] investigated the effectiveness of PBL on student achievement and motivation in science, together with its influence on attitudes towards science and skills in elementary, secondary and higher educational levels. Additionally he examined the effects of some moderator variables including publication type, research design, teacher effect, researcher effect, country, subject matter, school level, PBL mode, length of treatment, group size, type of questions and assessment instrument on the effectiveness of PBL. The results of 147 effect sizes he calculated from 88 primary studies included in his research clearly showed PBL's superiority over traditional teaching methods on different outcomes. More specifically, PBL had a huge impact with a large effect size of 0.820 on students' achievement in science subjects in different levels and revealed medium effect sizes of 0.566, 0.616, and 0.565 for students' attitude towards science, motivation in science and different kinds of skills respectively. Moreover, he reported that moderator analyses indicated noteworthy impact of publication type, country, subject area, school level and length of treatment on the effectiveness of PBL.
There is also a body of evidence derived from interviews and observations in related research on PBL. In summary, it was suggested that PBL granted students the chance to work on real life problems in their cooperative learning groups and supported student development by enhancing sharing of knowledge; PBL was well received and found useful both by the instructors and students; PBL enhanced the development of self-regulated learning skills of students better than their counterparts receiving traditional instruction; the interactions among students in PBL tutorials contributed substantially to group processes and performance outcomes; the ultimate gains of PBL classes were interaction and access to more information as a result of increased research skills; PBL students' reliance on their instructors' guidance decreased and they focused more on their peers' opinions and contributions; and PBL was influential in the development of group interactions, student self-reliance, research skill development and active participation during the learning process [21,22,23,24,25,26,27,28,29]. On the other hand, the workload of PBL tutorials was a common complaint among students and students' statements of uncertainty (what to learn, where to find, how to start etc.) at the beginning of PBL tutorials were some of the challenges noticed [30,26].
The prior meta-analyses and reviews summarized above clearly suggest that PBL studies mainly focused on cognitive and affective outcomes and there is lack of review on conative outcomes of PBL instruction. Nevertheless, in these analyses, PBL students seemed to have gained better performance than those students with traditional training in knowledge principles and application, problem solving, and self-directed learning, whereas traditional students' performance in basic knowledge were better than PBL students.

Method
This paper examined the dissertations on PBL in the National Thesis Database of CoHE in Turkey. For this purpose, the master's and doctorate dissertations were searched by using "problem-based learning" as a keyword with no time restrictions. A total of 129 dissertations (82 master's, 47 doctorate) were found (as of December 2016), 20 (12 master's, 8 doctorate) of which were not accessible due to restrictions requested by the authors themselves. In line with pre-determined criteria of inclusion, out of the remaining 109 dissertations (70 master's, 39 doctorate), 8 dissertations (6 master's, 2 doctorate) that failed to provide information for the majority of the selected analysis criteria were eliminated by the author and raters unanimously. In the end, 101 dissertations (63 master's, 38 doctorate), were included in the study (the list of dissertations is in Appendix A). They were all downloaded from the database and undergone rigorous content analysis.
Content analysis was chosen as a qualitative method to investigate the research trends in research. Content analysis is defined by Krippendorff [31] as "a research technique for making replicable and valid inferences from texts (or other meaningful matter) to the contexts of their use" (p. 18). Additionally, Cohen, Manion and Morrison [32] define it as "the process of summarizing and reporting written data -the main contents of data and their messages" (p. 475). They also add that "it defines a strict and systematic set of procedures for the rigorous analysis, examination and verification of the contents of written data." As Krippendorff states it, content analysis is learnable and divorceable from the personal authority of the researcher and it "provides new insights, increases a researcher's understanding of particular phenomena, or informs practical actions" (p. 18). Therefore, the author preferred this scientific tool to investigate research trends in dissertations in order to inform practitioners and the faculty about PBL's practical actions and considerations.

Data coding and analysis
This study used 101 dissertations derived from the Council of Higher Education National Thesis Database in Turkey that were completed from 2002 to 2015 to examine the research trends in the PBL field. Throughout the process of content analysis, three colleagues with doctorate degrees in educational sciences helped the author in reviewing the dissertations and doing the analyses. The agreement among the raters was calculated as 90.2.
As one of the major purposes of this study, a dissertation classification form (DCF) was developed by the author and assisting colleagues to classify the categories to help identify the research trends. This form was inspired from a previously developed "paper classification form" of Sözbilir and Kutu [33]. After examined by colleagues and experts, the form was implemented on randomly selected 15 dissertations for reliability. The results of this initial trial were discussed to alleviate any disagreements on the items. According to discussions and negotiations, the form was revised again and its final version (see Appendix B) was agreed upon to be used for comprehensive content analysis.
In line with the second research objective, the results obtained in these dissertations were also examined and reported separately with regards to cognitive, affective and conative domains.
For the analysis purposes, by means of the dissertation classification form, all data were recorded on an electronic database and later transferred to SPSS 20. Frequencies and percentages were calculated for each category. These data were later commented and discussed.

Research Question 1
Among the 101 studies examined in this research, 63 were master's studies and 38 were doctoral studies ( Table 1)  Analysis show that PBL research was most frequently utilized in the learning domain of science (86%), followed by social studies domain (10%), whereas there were only 2 studies in the domain of arts and 3 studies in the domain of language arts ( Table 2).
The findings on research method revealed that quantitative methodology was more prevalent among the dissertations. 51.52% of the studies used quantitative research methodology, whereas 48.5% utilized quantitative and qualitative research methods (Table 3).
It was found that researchers favored experimental design (90.1%) more than the others ( Table 4).
The analysis on research domain focused in studies revealed some interesting facts. The most frequently focused research domain was cognition and affect (56.4%), followed by cognition only (27.7%) ( Table 5).   The findings showed that the most frequently used subject group in the experimental studies were elementary education students (42.6%), followed by teacher education students (31.7%) ( Table 6). Considering group sizes, the studies on PBL preferred to have group sizes of 41-80 (60.4%), whereas the percentages of studies which used group sizes of 1-40, 81-120, and 120 and above were 17.8%, 11.9% and 12.5%, respectively (see Table 7). Taking the length of treatments into account, the treatments on PBL research lasted mostly four weeks (17.8%), five weeks (11.9%) and six weeks (16.8%) ( Table 8). The findings of the content analysis showed that there were 37 different types and in total 320 instruments used to collect data for research purposes. It is quite obvious that majority of the studies used achievement tests (30.9%) and attitude scales (18.1%) as instruments of measurement (see Table 9). The instruments categorized under the title "attitude scale" were the ones used to assess students' dispositions towards the course subject in which PBL was implemented (attitude towards chemistry, math, science and technology, etc.). The cross-reference of the findings depicted in Tables 1-9  learning domain, whereas studies completed in teacher education chose physics (N=7) and chemistry (N=7). Considering the length of treatments, 10 of the 18 treatments that lasted 10+ weeks were completed in teacher education, four of the remaining research was held in higher education institutions. As stated before the treatments in master's research mostly lasted 4-6 weeks (in 25 out of 43 studies).
The qualitative findings derived from interviews and observations were too fragmented to include in the present study, hence should be reported in detail somewhere else. However, a surface analysis of those qualitative findings indicated that PBL was reported to increase students' engagement, taking responsibility of their own learning, cooperation and interaction among group members and their self-efficacy in various subjects. The reports also showed that students learned more by doing or working on real life problems, they used higher-order thinking skills more frequently, they searched and shared more sources, hence their research skills developed substantially. On the other hand; difficulty in having access to information, noise in the classroom, confusion in the beginning, having difficulty in adapting to active participation instead of passive learning were some of the challenges noted.

Research Question 2
The findings of content analysis showed that PBL students did better in achievement tests than their conventional counterparts, except for a few studies (N=12), where no difference was observed between two groups of students. It is important to note that seven of those non-significant findings were obtained in studies at elementary education level, which lasted 3-5 weeks and focused on students' achievements only in one unit (e.g. human and environment, algebraic expositions and equations, concept of area and perimeter). On the other hand, in other research which also lasted 3-5 weeks (N=31), students in PBL classes showed significant progress in their achievements. Nevertheless, the findings of the present study regarding cognitive outcomes of PBL instruction revealed that PBL students were superior to traditional students both in factual and conceptual knowledge and specifically in skills of knowledge application. These findings seem to be in line with those of previous meta-analyses and reviews [7, 8,12,13], where PBL students were found to be doing better or equally well as their conventional counterparts.
Regarding the findings on attitudes on course subjects (e.g. chemistry, biology, etc.), in 35 studies out of 58 (60.3%), there was positive influence of PBL instruction on student attitudes, whereas in the remaining studies no significant difference was observed. Among those 23 studies with PBL treatment's no significant influence on attitudes, 16 were master's and 7 were doctoral research; 13 studies were conducted at elementary education level, and science teaching (N=7) and math (N=7) were the learning domains. Overall, as reported in prior reviews on PBL instruction [4,5], PBL students' attitudes and dispositions towards learning after PBL treatment and student and faculty satisfaction with the instruction were found positive. In those studies examining students' dispositions towards the PBL itself (N=13, PhD=7 and Master's=6), student reports were highly favorable and positive.
The findings regarding other measures revealed mixed results. Considering the more frequently utilized measures, the PBL instruction was found to have significant effect on scientific process skills in 12 out of 17 studies (70.6%), on creative thinking skills in 6 out of 11 studies (54.5%), and on problem solving skills in 8 out of 12 studies (66.7%).
The only studies that investigated conative outcomes of PBL instruction focused on such elements like self-direction/self-regulation, motivational styles and strategies, locus of control, and learning approaches. In 5 out of 8 studies focusing on motivational elements and in two studies examining learning approaches in PBL contexts, the findings demonstrated that PBL instruction had significant influence. While the studies investigating PBL's influence on self-directed learning readiness (one 4-week and one 15-week treatment) showed significant differences, the studies on PBL's effect on students' self-regulatory skills and locus of control reported mixed results. Students' self-regulation levels and their locus of control in PBL contexts were found to be similar to those of students in traditional teaching and learning settings after treatment periods of 12, 14 and 15 weeks; however in a research which lasted 19 weeks at elementary education level, significant positive gains were recorded on students' self-regulated learning scores. The inconclusive and inconsistent results did not provide enough evidence on PBL's positive influence on such elements. Having no concordance with the results of studies by Leary [10] and Loyens, Magda and Rikers [11], these findings confirm the necessity to have more research to get a better picture of PBL's influence on conative elements like self-direction or self-regulation.

Discussion
This study aimed to investigate the research trends in studies of PBL from 2002 to 2015 in Turkey. The results of the findings are discussed and commented below.
The content analysis of master's and doctoral dissertations showed that the number of studies on PBL was the highest during the period of 2007-2013. However, this increase in the number of dissertations is considered unexceptional since constructivist approaches were introduced into Turkish education system starting in 2005-2006 Academic Year.
Considering the analyses results of categories under focus, this study indicated that; science was the most favorable learning domain (see Table 2), quantitative method was slightly more utilized as compared to qualitative and quantitative (mixed) methodology (see Table 3), experimental was the dominant design in the dissertations (see Table 4), cognition and affect followed by cognition only were the mostly utilized research domains (see Table 5), subject groups were predominantly at elementary education level and secondly from teacher education (see Table 6), studies preferred to have the subject group size of 41-80 (see Table 7), the treatments on PBL research lasted mostly four and six weeks (see Table 8), and finally majority of the studies used achievement tests and attitude scales (see Table  9) as instruments of measurement.
These findings with regard to the first research question revealed evidence on several research trends. First, studies generally focused on PBL's effect on cognition and affect, hence they preferred to use achievement tests and attitude scales of various sorts. Second, there was an increase in the utilization of both quantitative and qualitative methods together, although it was slightly lower than quantitative method only. Third, studies at elementary education and teacher education were on the rise. Fourth, it is obvious that science was and potentially will continue being the dominant learning domain in PBL research. Finally, the researchers' preferences for length of treatment in majority ranged from 3 to 9 weeks due to curricular and time constraints.
The findings regarding PBL's influence on cognitive and affective outcomes seemed to be in concordance with the findings obtained in previous meta-analyses and reviews. Supporting the reports of Albanese and Mitchell [4], Dochy et al.
[7], Gijbels et al. [8] and Üstün [13], in the present study there is evidence on PBL's significant positive effect on knowledge application, however contrary to previous reviews [5,6,9,12], PBL's meaningful influence on basic knowledge is observed as well. Also, similar to the findings in other reviews [13,12], this study revealed some affective outcomes like positive attitudes and views towards learning of subjects and specifically on the implementation of PBL.
In comparison with previous meta-analyses and reviews, it is clear that PBL research in Turkish dissertations was mainly concerned about cognitive outcomes and secondly affective outcomes. There were just a few studies which focused on conative elements of learning like self-direction, self-regulation, locus of control, and motivational strategies in learning. Leary [10] pointed to the same fact and suggested that more studies should be done in order to get a better picture of PBL's influence on SDL and SRL. The results in the present study regarding the elements of this domain were found inconclusive. The same mixed results on whether PBL fosters self-direction or self-regulation were also reviewed in the study of Loyens, Magda and Rikers [11]. Erdogan and Senemoglu [34] addressed to this issue in their research and stated that different study results on PBL's effect on conative elements such as self-direction/self-regulation could be explained by students already being equipped with such skills and the short period of application reserved to see the changes PBL would have on them. Similarly, in their study investigating students' progress on self-regulation through undergraduate levels of 1-4, Van der Hurk, Wolfhagen, Dolmans and Van der Vleuten [35] found that real development occurred only in the 3rd and 4th grades. It is viable to say that a better and healthier observation of self-directed/self-regulated skills development or the improvement of other conative elements could be made in a study that would last longer than a few months or a year.
The qualitative results of the dissertations revealed overall positive attitudes towards PBL instruction on part of both learners and teachers. In their reviews, Albanese and Mitchell [4] noted the same satisfaction of faculty and students with PBL instructional method. Certainly, an increase in qualitative assessments like interviews could provide faculty and teachers with valuable aspects of PBL use.
So far, the dissertations on PBL showed promising results. Because, instead of passive learning, in PBL students participate in cooperative learning groups where they work on real life problems, hence they have the chance to search for more information, access different sources, see the actual practical results of theoretical constructs and eventually take the responsibility of their own learning. However, more studies are required before conclusions are drawn about the effectiveness and usefulness of PBL in a variety of disciplines and levels. As in the studies found in literature, the real influence of PBL can be seen in activities based on application and in real life problems encountered during occupational experiences [36,15,37]. Therefore, studies on long-term effects of PBL use, which would also cover after school periods, could provide the faculty and teachers with better picture of PBL's real power.
Considering the methodologies used in the dissertations, it is important to note that except for a few medical schools, in none of the studies PBL was the common methodology followed and implemented through all levels and courses. In other words, PBL was implemented and the results were analyzed as part of a research, in which PBL as a constructive methodology was new to both instructors and teachers. Further, a few studies undertaken about Turkish education system has shown that neither teachers nor students were ready enough to accept or adopt the new constructivist approaches, both at theoretical and application level. Teacher training programs and in-service courses were seen insufficient; crowded classes, having difficulty at accessing sources, lack of infrastructure such as laboratories added up the problem; the teachers and students tended to continue using traditional teaching and assessment approaches more frequently [38,39]. Additionally; Gür, Dilci and Arseven [40] reported that prospective (pre-service) teachers found themselves adequate in terms of theoretical knowledge, but lacked practical knowledge. Student teachers believed that deficiency originated from lack of education including enough training on constructivist applications. Aygören [41] suggested that even school administrators were in need of in-service training on constructivism and constructivist learning environments.
It was found that specific care was given to put the instructional principles of PBL methodology into action. The courses were designed to begin with the introduction of an ill-structured problem, where teachers acted as facilitators and learners actively constructed knowledge by defining learning goals, seeking information to build upon prior knowledge, reflecting on the learning process, and participating actively in cooperative/ collaborative learning groups [16,17,18]. However, under the condition of insufficient infrastructural support (e.g. no tutorial rooms, limited access to information in and out of classes through the Internet, small crowded classes, etc.) and tutors not assigned to tutorial groups especially at elementary and secondary education levels challenged the instructors to develop their own version of PBL. For example, in one of the dissertations, in the absence of specific tutors to each cooperative group in a large class, the researcher described the instructor taking the role of a "floating tutor" [42, p. 40-41]. Barrows [19] also exemplified different versions of PBL methodologies that might vary according to different conditions, settings, learner groups, etc.
Regarding the findings of the present research and those in literature confirm the fact that PBL could be used as an instructional methodology in all disciplines and subject areas, through all levels (from elementary to higher education) as long as the core principles are practiced. However, it is not plausible to expect immediate results from a PBL implementation in the beginning. Also, it is also not viable to expect sudden changes in student roles. Students may need more time and experience as they have been taught passively in controlled and well-structured learning environments. It is deemed important that for the successful application of methods like PBL, infrastructure needs should be met and necessary changes must be incorporated into the curriculum, especially to content and timing. The undergraduate and in-service training of teachers as practitioners on constructivist teaching and learning methodologies including PBL should be granted priority. Finally, it is worth remembering that it is the responsibility of practitioners to ensure the implementation of PBL according to its principles, to provide assistance to students in their transition to more active learning, cooperation and taking responsibility of their own learning.

Conclusions
The present study examined the research trends in studies of problem-based learning from 2002 to 2015 in Turkey. For this purpose, 101 master's and doctoral dissertations underwent rigorous content analysis. In short, this study suggested some research trends in PBL dissertations through years. Besides these major considerations, this present research drew attention to analysis of the findings with regard to cognitive, affective, and conative outcomes.
The research has limitations though. First, the current study focused only on PBL dissertations completed between 2002 and 2015 and published in the dissertation database of CoHE in Turkey. Second, for analysis purposes, the study reviewed only the learning domains, research methods, research designs, research domains, subject groups, group sizes, length of treatments, and data collection instruments in these dissertations and analyzed the findings with regards to cognition, affect and conation. Third, the dissertations not accessed due to authors' restrictions and those providing not sufficient data for analysis were kept out of scope.
The findings of this analysis indicated that constructivist learning approaches such as PBL could be used successfully in other disciplines apart from medical education and could be utilized in different levels from primary to higher education. However, a comparative analysis of these findings in Turkey with those results of PBL studies in other countries or comparing the present findings with the results of studies investigating other constructivist methodologies such as project-based learning, case-based learning etc. would assist researchers in their prospective studies and at the same time create a better picture of strengths and challenges of using constructivist teaching and learning applications.
In conclusion; PBL as educational practice continues to have large impacts on all levels of education and across different disciplines. While the results of studies examining the effects of PBL seem to be conclusive regarding students' problem-solving abilities, attitudes and predispositions towards learning, and clinical or occupational skills in application, much more needs to be known about PBL's influence on the acquisition of basic knowledge and some conative elements like self-direction/self-regulation and locus of control. The analysis of an English preparatory program based on the needs assessment of problem-based learning and regular groups in a foreign language-medium institution

Appendix A List of Dissertations Used in the Analyses
The effects of individual and collaborative problem-based learning using an online synchronized learning tool on critical thinking abilities, academic achievements, and attitudes toward internet use The effect of know-want-sample-learn strategy, which is developed by using problem based learning and know-want-learn strategy, towards the 5th grade students' attitudes in science and technology lesson and towards to remove their misconceptions