Balemen, N., & Özer Keskin, M. (2018). The effectiveness of Project-Based Learning on science education: A meta-analysis search. International Online Journal of Education and Teaching (IOJET), 5(4), 849-865. http://iojet.org/index.php/IOJET/article/view/452/297 Received: 12.06.2018 Received in revised form: 13.09.2018 Accepted: 18.09.2018 THE EFFECTIVENESS OF PROJECT-BASED LEARNING ON SCIENCE EDUCATION: A META-ANALYSIS SEARCH Review Article Nuri Balemen Ministry of National Education, Turkey nuribalemen@hotmail.com Melike Özer Keskin Gazi University, Turkey mozerkeskin@gazi.edu.tr Dr. Nuri Balemen is a Biology teacher in Ministry of National Education at Turkey. He studies about nanobiotechnology and science education. Dr. Melike Özer Keskin is currently an associate professor in the Department of Biology Education at Gazi University. Her research interest areas include argumentation based science inquiry and bioethics education. Copyright by Informascope. Material published and so copyrighted may not be published elsewhere without the written permission of IOJET. http://iojet.org/index.php/IOJET/article/view/452/297 mailto:nuribalemen@hotmail.com mailto:mozerkeskin@gazi.edu.tr International Online Journal of Education and Teaching (IOJET) 2018, 5(4), 849-865. 849 THE EFFECTIVENESS OF PROJECT-BASED LEARNING ON SCIENCE EDUCATION: A META-ANALYSIS SEARCH Nuri Balemen nuribalemen@hotmail.com Melike Özer Keskin mozerkeskin@gazi.edu.tr Abstract The present study aims to demonstrate, by means of a meta-analysis, the effectiveness of Project-Based Learning in the context of academic performance and various study characteristics. For this purpose, the relevant literature was reviewed to identify studies using Project-Based Learning in the fields of physics, chemistry, biology, and science. Following the literature review, the study characteristics and the criteria for their inclusion were determined. A total of 48 studies were included and, by means of the analyses conducted, the general effect size of Project-Based Learning in science education was found to be 1.063. This is quite a large effect size by Cohen’s criteria and it shows that Project-Based Learning is 86% more effective in science education compared to traditional learning approaches. Project-Based Learning was found to have a great effect size in different subjects (physics, chemistry, biology, and science), at different levels (primary, secondary, and tertiary), and with samples of various sizes (small, medium, and large). Project-Based Learning can thus be deemed more effective compared to traditional learning approaches. Keywords: Science education, Project-Based Learning, meta-analysis. 1. Introduction Technological advancement plays an important role in the economic development of countries and the improvement of their living standards (Ayas, 1995; Bayındır, 2007; Ünal, Coştu, & Karataş 2004). Technological advancement, however, requires societies to have science literacy in order to have a say in the developing world and to have an outstanding position amongst successful nations (Korkmaz & Kaptan, 2001). In order for such a society to emerge, there appears to be an acute need for individuals with the ability to use science process skills, to solve problems in the face of the unknown, to establish cause-and-effect links by observing the settings that they are in, and to use high-level cognitive skills (Korkmaz & Kaptan, 2001; Köseoğlu, Tümay, & Budak, 2008). An effective science education plays a crucial part in raising such individuals, as one of the functions of science education is to enable students to develop into science-literate individuals (Çakıcı, 2009; Emrahoğlu & Öztürk, 2010; Ünal et al., 2004). The realization of an effective science education in turn depends on a sound science curriculum and its proper implementation (Ayas, 1995). Many researchers have found that Project-Based Learning (PBL) is top of the methods that can be used in science classes to raise the quality of science education and to enable students to use their scientific knowledge and skills to solve problems in their daily lives and to become science-literate individuals (Barron et al., 1998; Dede & Yaman, 2003; Demirhan mailto:nuribalemen@hotmail.com mailto:mozerkeskin@gazi.edu.tr Balemen & Özer Keskin 850 & Demirel, 2003; Filippatou & Kaldi, 2010; Gillies & Ashman, 2000; Korkmaz & Kaptan, 2001; Liu & Hsiao, 2002; Şahin, 2009; Wolk, 1994) because PBL is a process that promotes individual learning, enables students to create links between school and life, supports lifelong learning, and encourages self-controlled learning (Dede & Yaman, 2003; Şahin, 2009). PBL is also thought to improve students’ knowledge and skills, raising their academic performance. The literature includes studies highlighting the positive effects of PBL on students’ academic performance in science education (Baran, 2007; Çeliker & Balım, 2012; Değirmenci, 2011; Ergül & Kargın, 2014; Güven, 2011; Hung, Hwang & Huang, 2012; İmer, 2008; Keskin, 2011; Nikbay, 2009; Özbek, 2010; Serttürk, 2008; Tortop, 2010; Wolk, 1994; Yurttepe, 2007). However, there are also studies which have found no change in students’ academic performance in science education (Ayan, 2012; Dilşeker, 2008; Ekiz, 2008; Gültekin, 2009; Özer & Özkan, 2011; Özer & Özkan, 2012; Özahioğlu, 2012; Toprak, 2007; Tuncer & Taşpınar, 2010). Furthermore, while PBL was found by Çil (2005) to have a positive effect on eighth-year students’ academic performance in science, it was shown to make no significant difference in seventh-year students’ performance. The literature contains many studies with conflicting views on the effects of PBL on student performance, but none on its effectiveness or on the consideration of its effects taking several variables into account. Disagreement on the effects of Project-Based Learning on student performance in science education has highlighted the need for studying the effectiveness of this approach. For this purpose, the current research was designed to find any effectiveness that Project-Based Learning might have in the context of students’ academic performance in science education. A meta-analysis was conducted to interpret the conclusions of the studies on PBL in the literature. 2. Method 2.1. Research Model A meta-analysis was used in this study to determine whether PBL was effective in science education. The meta-analysis method involves literature review that combines and re- interprets the conclusions of similar individual studies in a given field (Hunter & Schmidt, 1990). Although many literature review methods exist, meta-analysis differs from them in that it is based upon statistical techniques and numerical data (Baran, 2007). Meta-analysis is used more and more often especially in many domains of social psychology, playing a crucial role in understanding social policies (Durlak & Lipsey, 1991). Meta-analysis has many different types but the present study used ‘Study Effect Meta-Analysis’. 2.2. Data Collection First, in line with the research objective and methodology, the inclusion criteria were determined by the researcher. The following criteria were used to identify the studies to be included in the meta-analysis:  The studies were to use a pre-test / post-test control group model,  They were to focus on the effects of PBL on students’ academic performance,  They were to report the sample size (n), arithmetic mean ( X ), and standard deviation (sd) values for the experiment and control groups, which would make the calculation of the effect size possible, or include data by which these values could be calculated. After the study criteria for the meta-analysis were determined, a literature review was carried out and the selection of the studies to be included was initiated. The literature review International Online Journal of Education and Teaching (IOJET) 2018, 5(4), 849-865. 851 began primarily by national studies. To this end, the Ulakbim and YÖK (Higher Education Council) databases as well as journals from faculties of education and congress and symposium papers were examined. In order to reach international studies, the EBSCO, ERIC and Proquest databases were searched online with the intention to increase the number of studies to be included in the meta-analysis. In this context, these databases were searched with the keyword ‘Project- Based Learning’. 1,059 studies showed up on the Proquest database and 418 on the ERIC database. After this literature review, the studies were all checked one by one for suitability for the study criteria. In this process, those studies on science education with a pre-test and post-test control group experimental pattern, academic performance measurement, and standard deviation and arithmetic mean were included in the meta-analysis. Studies considered to be unsuitable by the inclusion criteria were excluded from the meta-analysis. The remaining studies were excluded from the meta-analysis because some of them were not on science education, some did not measure academic performance, and some did not include the necessary statistical data (standard deviation and arithmetic mean). The reasons for the exclusion of 82 dissertations are presented in Table 1. Table 1. Reasons for excluding studies from the meta-analysis Reason Number of studies Computer 5 Life science 4 Social science 17 Mathematics 10 Geometry 3 English 4 Geography 1 Statistic 1 Electronic 2 Religion 2 Material development 2 Visual art 5 Planning and evaluation 1 Balemen & Özer Keskin 852 Studies not measuring academic achievement About attitude and risk taking 7 About literacy 5 About awareness 2 About self-efficacy 2 About epistemological belief 2 About motivation 1 Missing statistical information Missing data about control group 5 Studies with different experimental design One group experimental design 1 Total 82 The literature review yielded 113 dissertations of which only 31 were considered to be suitable by the meta-analysis inclusion criteria. 29 of these were master and 2 were doctoral theses. Of the 82 theses excluded, 57 were in different fields and 25 failed to either measure academic performance or contain sufficient statistical data (Table 1). Likewise, the literature review yielded 79 papers. After each of them was checked for suitability by the inclusion criteria, most were excluded. Only 12 papers were deemed to be suitable for inclusion in the meta-analysis. Apart from these, congress booklets, faculty of education journals, and social sciences journals were regularly screened. Of the studies found, some were registered on multiple databases, and some were non-experimental. These non-experimental studies were excluded from the meta-analysis. Some of the master’s and doctoral dissertations on Project-Based Learning were found to have been published also as papers. Such dissertations were included in the meta-analysis as papers. In the end, a total of 48 studies were found to meet the inclusion criteria. Studies with different domains and different sample groups within the same research were taken as separate studies. 2.3. Data Analysis The statistical data from different studies needs to be converted into a common unit of measurement – the effect size – in order to be interpreted as a whole (Şahin, 2005). The effect size is a standard measurement value used to determine the force and direction of a correlation in a given study. In this research, the effect size for each study and the combined effect size were calculated by using the Metawin 1.00 program (Rosenberg, Adams & Gurevitch, 1997). The Metawin program calculates the experiment and control group averages and the combined standard deviation values by using ‘Hedges’ d’ for the average effect size and the general effect size of each study (Özdemirli, 2011). Furthermore, the random effects model was selected on the Metawin program to calculate the average effect size in this study. The random effects model calculates the effect size by considering both the International Online Journal of Education and Teaching (IOJET) 2018, 5(4), 849-865. 853 inherent variance of the studies and the variance between studies (Okursoy, 2009). For all the calculations and graphs in data analysis, the Microsoft Excel 2007 program and the SPSS 18.0 program were used. The level of significance for all statistical calculations was accepted as .05 in the study. Developed in 1977 by Cohen, the effect size is a standard measurement value used to determine the force and direction of a correlation in a given study and, for whatever is studied, it answers the question ‘How effective is it?’ (Okursoy, 2009). Found by dividing the difference between the experiment group and the control group by the combined standard deviation value of the two groups, the effect size value is between -∞ and ∞. Negative (-) values indicate a higher score for the control group while positive (+) values indicate a higher score for the experiment group (Özdemirli, 2011). The greater the difference between the two groups and the smaller the standard deviation is, the larger the value of the effect size becomes. The effect size of a study offers clues on the significance of the conclusions of that study. For instance, in a study with a statistically significant difference between the experiment and control groups, there may be no significant effect size between the groups (Ergene, 1999). Even though the classification of effect sizes may vary, the effect sizes of the studies analyzed in this research were interpreted according to Cohen, who interprets effect size values as follows (Yıldız, 2009);  If the effect size value is between 0.20 and 0.50, there is a small-scale effect,  If the effect size value is between 0.50 and 0.80, there is a medium-scale effect,  If the effect size value is greater than 0.80, there is a large-scale effect. The present study aims to reveal the effectiveness of Project-Based Learning in the context of student performance in science education. By means of a meta-analysis, the conclusions of studies in the PBL literature were combined and interpreted. Through the meta-analysis, this paper seeks to answer the following questions: 1) Does Project-Based Learning have a positive effect on students’ academic performance in science education? 2) Is there a significant difference between the effect sizes of the publications in the Project-Based Learning literature depending on the publication type? 3) Is there a significant difference between the effect sizes of the publications in the Project-Based Learning literature depending on the selected subject (physics, chemistry, biology, and science)? 4) Is there a significant difference between the effect sizes of the publications in the Project-Based Learning literature depending on students’ education levels (primary, secondary, and tertiary)? 5) Is there a significant difference between the effect sizes of the publications in the Project-Based Learning literature depending on the sample size (small, medium, and large)? 6) Is there a significant difference between the effect sizes of the publications in the Project-Based Learning literature depending on the technique used (studies using PBL only vs studies using PBL in addition to another method)? Balemen & Özer Keskin 854 7) Is there a significant difference between the effect sizes of the publications in the Project-Based Learning literature depending on the publication status (published / unpublished)? 3. Findings 3.1. General Effect Size of the Effectiveness of PBL The general effect size of the studies included in the research was found to be d=1.063 (Figure 1). According to Cohen (1977), this is quite a great general effect. For this general effect size, the corresponding value in the z table was found to be 86.6%. In other words, students receiving PBL education performed 86.6% better than those receiving education with traditional methods. Many studies on PBL in science education conclude that students receiving PBL education perform better than those receiving education with traditional methods (Baran, 2007; Çeliker & Balım, 2012; Değirmenci, 2011; Ergül & Kargın, 2014; Güven, 2011; Hung, Hwang & Huang, 2012; İmer, 2008; Keskin, 2011; Nikbay, 2009; Özbek, 2010; Serttürk, 2008; Tortop, 2010; Yurttepe, 2007). Therefore, it can be claimed that that the conclusion of this meta-analysis is in line with the conclusions of other studies using PBL. Figure 1. General effect size of the studies included in the meta-analysis 3.2. Effectiveness of PBL Depending on Publication Type In this meta-analysis, 32 articles, 14 master’s theses, and 2 doctoral theses that met the inclusion criteria were analyzed. The effect sizes in all publication types were found to be large (Table 2). According to the findings, the largest effect size was found in the doctoral dissertations (d=1.8957). The effect size was calculated as d=1.0497 for the master’s theses and as d=0.8830 for the articles. The analysis results suggest that there are no statistically International Online Journal of Education and Teaching (IOJET) 2018, 5(4), 849-865. 855 significant differences between the average effect sizes depending on different publication types (QB=3.6372; p>0.05). Table 2. Effect sizes of the studies depending on publication type Variable Homogeneity between groups n The Overall mean effect size d d for %95 CI Homogeneity within groups (QB) Lower Upper (QWI) Type of publication 3,6372 Article 14 0,8830 0,5191 1,2469 60,1970 Master thesis 32 1,0497 0,8026 1,2969 238,1288 PhD dissertation 2 1,8957 0,9141 2,8773 0,0000 3.3. Effectiveness of PBL Depending on Selected Subject In order to examine the effect of different subjects on the effect size, the studies were grouped into four categories: those conducted in physics, chemistry, biology, and science. Looking at the average effect sizes of these four categories, it can be argued that PBL has a great effect in all subjects. The largest effect size (d=1.3638) was found in biology (Table 3). The analysis results suggest that there are no statistically significant differences between the average effect sizes depending on different subjects (QB=1.3627; p>0.05). Table 3. Effect sizes of the studies depending on selected subject Variable Homogeneity between groups n The Overall mean effect size d d for %95 CI Homogeneity within groups (QB) Lower Upper (QWI) Subject area 1,3627 Science 35 0,9738 0,7306 1,2169 236,4182 Chemistry Physics Biology 2 6 5 1,0101 1,1476 1,3638 -0,0047 0,5587 0,7116 2,0248 1,7365 2,0160 0,0084 33,9053 36,2566 Balemen & Özer Keskin 856 3.4. Effectiveness of PBL Depending on Education Level In order to examine the effect of the education level on the effect size, the studies were grouped into three categories: those conducted in primary, secondary, and tertiary education. PBL was found to be highly effective in secondary education (d=1.7677), primary education(d=0.9638) and tertiary education (d=0.9089) (Table 4). The analysis results suggest that there are no statistically significant differences between the average effect sizes depending on different education levels (QB=6.0919; p>0.05). The effectiveness of Project- Based Learning in all levels demonstrates that it can be used more frequently in all levels. Table 4. Effect sizes of the studies depending on students’ education level Variable Homogeneity between groups n The Overall mean effect size d d for %95 CI Homogeneity within groups (QB) Lower Upper (QWI) Education level 6,0919 Primary school 33 0,9638 0,7275 1,2001 213,2403 Tertiary Secondary school 10 5 0,9089 1,7677 0,4820 1,1502 1,3359 2,3852 59,8492 12,8727 3.5. Effectiveness of PBL Depending on Sample Size While including the studies in the meta-analysis, differences were noticed between their sample sizes. Therefore, sample size was added to the study characteristics. In this context, the studies included in the meta-analysis were grouped into three categories depending on their sample size: small (n ≤ 50), medium (51< n ≤ 100), and large (n>100). A great majority of these studies had a medium-scale sample size (n=27). Project-Based Learning was found to be effective mostly in studies with a small-scale sample size (d=1.2314) (Table 5). Studies with medium- and large-scale sample sizes were found to have effect sizes of d=0.8984 and d=1.2013, respectively. The analysis results suggest that there are no statistically significant differences between the average effect sizes depending on different sample sizes (QB=2.4148; p>0.05). International Online Journal of Education and Teaching (IOJET) 2018, 5(4), 849-865. 857 Table 5. Effect sizes of the studies depending on sample size Variable Homogeneity between groups n The Overall mean effect size d d for %95 CI Homogeneity within groups (QB) Lower Upper (QWI) Sample size 2,4148 Small 18 1,2314 0,8855 1,5773 76,0539 Medium Large 27 3 0,8984 1,2013 0,6305 0,4277 1,1663 1,9750 211,4180 18,9237 3.6. Effectiveness of PBL Depending on Techniques Used Of the studies included in the meta-analysis, 40 used Project-Based Learning only whereas 8 complemented PBL with other techniques. The average effect sizes were then calculated for PBL and PBL plus other techniques (Table 6). While the average effect size for the group with PBL only was dPBL=1.0205, it was found to be dPBL+other=1.1086 for the group with PBL and another technique. A high level of effectiveness was detected in both techniques. The results suggest that there are no statistically significant differences between the average effect sizes depending on the use of PBL only or PBL and other techniques (QB=0.1018; p>0.05). Table 6. Effect sizes of the studies depending on techniques used Variable Homogeneity between groups n The Overall mean effect size d d for %95 CI Homogeneity within groups (QB) Alt Üst (QWI) Study design 0,1018 PBL 40 1,0205 0,7950 1,2461 278,6000 PBL and other techniques 8 1,1086 0,6168 1,6004 40,2038 Balemen & Özer Keskin 858 3.7. Effectiveness of PBL Depending on Publication Status As for the publication status of the studies, 14 published and 34 unpublished studies were included in the meta-analysis. The average effect sizes for published and unpublished studies were found to have quite high values (dpublished=0.8831 and dunpublished=1.1020) (Table 7). No significant difference was detected between the average effect sizes of published and unpublished studies (QB=0.9228, p>0.05). In other words, there is no publication bias. Table 7. Analysis results on effect size differences between the studies depending on publication status Variable Homogeneity between groups n The Overall mean effect size d d for %95 CI Homogeneity within groups (QB) Alt Üst (QWI) Publishing status 0,9228 Published 14 0,8831 0,5098 1,2563 60,1970 Unpublished 34 1,1020 0,8565 1,3475 260,2372 4. Results and Discussion The general effect size of the studies included in the research was found to be d=1.063. According to Cohen (1977), this is quite a great general effect. For this general effect size, the corresponding value in the z table was found to be 86.6%. In other words, students receiving PBL education performed 86.6% better than those receiving education with traditional methods. Many studies on PBL in science education conclude that students receiving PBL education perform better than those receiving education with traditional methods (Baran, 2007; Çeliker & Balım, 2012; Değirmenci, 2011; Ergül & Kargın, 2014; Güven, 2011; Hung, Hwang & Huang, 2012; İmer, 2008; Keskin, 2011; Nikbay, 2009; Özbek, 2010; Serttürk, 2008; Tortop, 2010; Yurttepe, 2007). Therefore, it can be claimed that that the conclusion of this meta-analysis is in line with the conclusions of other studies using PBL. In other words, it can be argued that PBL raises students’ academic performance. According to the findings from the three groups compared, the largest effect size was found in the doctoral dissertations (d=1.8957) and the smallest effect size was found in the articles (d=0.8830). No statistically significant difference was detected between the average effect sizes of articles, master’s theses, and doctoral theses (QB=3.6372; p>0.05). When the effect sizes were compared according to the subjects selected, the largest average effect size was found in biology (d=1.3638) and the smallest average effect size was found in science (d=0.9738). In other words, PBL was mostly effective in biology but still highly effective in physics, chemistry, and science. No statistically significant difference was detected between the average effect sizes depending on different subjects (QB=1.3627; p>0.05). International Online Journal of Education and Teaching (IOJET) 2018, 5(4), 849-865. 859 A great majority of the studies included in the meta-analysis were in primary education (n=33), followed by tertiary (n=10), and secondary education (n=5). Although only 5 studies were in secondary education, PBL was found to be most effective at secondary education level (d=1.7677). It was also quite effective in primary and tertiary education, with effect sizes of d=0.9638 and d=0.9089, respectively. However, no statistically significant difference was detected between the average effect sizes depending on different education levels (QB=6.0919; p>0.05). Project-Based Learning was found to be effective mostly in studies with a small-scale sample size (d=1.2314), and still quite effective in studies with medium- and large-scale sample sizes. Studies with medium- and large-scale sample sizes were found to have effect sizes of d=0.8984 and d=1.2013, respectively. The analysis results suggest that there are no statistically significant differences between the average effect sizes depending on different sample sizes (QB=2.4148; p>0.05). Of the studies included in the meta-analysis, 40 used Project-Based Learning only whereas 8 complemented PBL with an additional technique. While the average effect size for the group with PBL only was d=1.0205, it was found to be d=1.1086 for the group with PBL and another technique. A high level of effectiveness was detected in both techniques. The results suggest that there are no statistically significant differences between the average effect sizes depending on the use of PBL only or PBL and other techniques. Used alone or in combination with other techniques, PBL affects students’ academic performance in a positive way. Since studies with statistically significant findings are more likely to be published, published works are expected to have a greater average effect (Rosenthal, 1991; Cooper & Hedges, 1994). Publication status can be found by looking at the difference between the effect sizes of published and unpublished works (Ergene, 1999). The present meta-analysis included 14 published and 34 unpublished studies which met the criteria. The average effect sizes for both published and unpublished studies using PBL were found to have quite high values (dpublished=0.8831 and dunpublished=1.1020). No significant difference was detected between the average effect sizes of published and unpublished studies. In other words, there is no publication bias. In addition to combining the findings of scientific studies, meta-analyses also give information on tendencies on the research topic. In this context, the tendencies revealed by the studies on PBL conducted between 2002 and 2014 can be summarized as follows:  A great majority of the studies on PBL in science education were master’s and doctoral dissertations and a large number of these were not turned into scientific articles. The fact that most of these studies (66.6%) were master’s theses may partly explain why these theses were not turned into articles.  A large majority of the studies on PBL in science education were conducted in Science (72.91%), which accounts for the prevalence of studies at primary education level over those at secondary and tertiary levels.  The number of studies on PBL in science education peaked between 2007 and 2010, followed by the period between 2011 and 2014.  In general, PBL was found to be used on its own in the studies examined (83.33%). Studies in which it was used in combination with other methods were in minority.  In PBL studies, the use of tests developed by the researcher is more common (93.75%). Balemen & Özer Keskin 860  PBL studies in science education used medium-scale samples (51< n ≤ 100).  The time spans over which PBL studies were conducted varied in length. No clear-cut tendencies were observed. Meta-analyses are also used to combine the findings of studies conducted independently of one another. Such findings can be summarized as follows;  Project-Based Learning affects students’ academic performance in a positive way.  Publication status makes no difference on the studies meeting the research criteria.  No statistically significant difference is detected between the average effect sizes of articles, master’s theses, and doctoral theses. Publication types have large effect sizes.  PBL practices are an effective method in the fields of science, physics, chemistry, and biology.  PBL has a high effect value whether used on its own or in combination with other techniques.  PBL studies yield positive results at primary, secondary, and tertiary education levels.  The effectiveness of PBL practices are independent of the sample size. The approach is effective with small, medium, and large-scale samples. 5. Suggestions 1. Most PBL studies seem to be conducted at primary education level even though data suggests that the highest effect value is obtained in studies conducted at secondary level. This could justify using PBL more frequently at secondary and tertiary levels. 2. Most Project-Based Learning studies seem to involve an experiment group where only teacher-centred traditional teaching methods are employed. No comparisons seem to be drawn between Project-Based Learning and other student-centred approaches such as Argumentation-Based Learning and Problem-Based Learning. Future studies could focus on comparisons between innovative and student-centred teaching approaches in both experiment and control groups. 3. Great difficulties were encountered in looking into the general characteristics of the studies included in the meta-analysis. A great majority of the studies contained no sufficient information about their researchers. There was a lack of information especially about the education level and professional experience of the implementer, and whether they were the same person as the researcher. In many studies where implementation was not carried out by the researcher, whether the implementers were given any informative training was not mentioned, either. In some studies, the duration of implementation was not given clearly, if at all. Since the duration of implementation as well as the professional qualities of the implementer such as their professional experience and pedagogical field knowledge are crucial in determining the outcome of a study, such detailed information must be given in all studies. 4. The effect size calculated for this study (dgeneral=1,063) is a criterion for researchers in Project-Based Learning. In future PBL studies, researchers could compare the effectiveness of their research to the effect size value of the present study for an idea. 5. The effect size obtained in this study shows that, in science education, PBL is an effective method on quite a large scale similar to laboratory-supported teaching (d=1,063>0,80). In our country, apart from the present study, only two meta-analysis International Online Journal of Education and Teaching (IOJET) 2018, 5(4), 849-865. 861 studies were found in Problem-Based Learning and Laboratory-Based Learning. Therefore, measuring the effectiveness in science education of other student-centred teaching methods such as Argumentation-Based Teaching, Collaborative Teaching, Computer-Aided Teaching, and Brain-Based Teaching, and comparing the findings with those of the present study will make a difference in raising the quality of science education. 6. Again national studies were found to include no meta-analysis on the effectiveness of Project-Based Learning in other subjects (mathematics, English, art, music, etc.). For this reason, meta-analysis studies in different fields will be useful in revealing where Project-Based Learning is most effective. 7. All studies included in the current meta-analysis on PBL were found to use only multiple-choice tests in assessing academic performance. In addition to the product, however, the process should also be evaluated in Project-Based Learning and must be included in student assessment. It could therefore be suggested that researchers use the progress files and rubrics in evaluating the process in their PBL studies. 8. In conclusion, PBL is found to be 86% more effective in science education than traditional teaching approaches. Therefore, PBL should be frequently used in science education in primary as well as secondary and tertiary education. For this purpose, this finding should be taken into consideration in curriculum development and more PBL activities should be included in course books. Balemen & Özer Keskin 862 References Ayan, M. (2012). Proje tabanlı öğrenme yaklaşımının ilköğretim öğrencilerinin fen bilgisi dersi akademik başarı düzeyine etkisi. Türk Eğitim Bilimleri Dergisi, 10(1), 167-183. Retrieved from http://www.tebd.gazi.edu.tr/index.php/tebd/article/view/104/91. Ayas, A. (1995). Fen bilimlerinde program geliştirme ve uygulama teknikleri üzerine bir çalışma: iki çağdaş yaklaşımın değerlendirilmesi. Hacettepe Üniversitesi Eğitim Fakültesi Dergisi, 11, 149-155. Retrieved from http://www.efdergi.hacettepe.edu.tr/yonetim/icerik/makaleler/1231-published.pdf. Baran, M. (2007). Proje tabanlı öğrenme modelinin fizik öğretmenliği ikinci sınıf öğrencilerinin elektrostatik konusu başarısına ve fiziğe yönelik tutumlarına etkisi üzerine bir araştırma. Yüksek Lisans Tezi, Dicle Üniversitesi Fen Bilimleri Enstitüsü, Diyarbakır. Barron, B., Schwartz, D., Vye, N., Moore, A., Petrosino, L., & Bransford, J. (1998). Doing with understanding: lessons from research on problem and project-based learning. The Journal of the Learning Science. 7(3), 271-311. Retrieved from http://web.mit.edu/monicaru/Public/old%20stuff/For%20Dava/Grad%20Library.Data/P DF/Brigid_1998DoingwithUnderstanding-LessonsfromResearchonProblem- andProject-BasedLearning-1896588801/Brigid_1998DoingwithUnderstanding- LessonsfromResearchonProblem-andProject-BasedLearning.pdf . Bayındır, M. (2007). Nanoteknoloji hayatımızda. Bilim ve Ütopya Dergisi, 152, 12-20. Cohen, J. (1977). Statistical power analysis for the behavioral sciences. Routledge. Cooper, H., & Hedges, L. (1994). The handbook of research synthesis. New York: Russell Sage Foundation Çakıcı, Y. (2009). Fen eğitiminde bir önkoşul: Bilimin doğasını anlama. Marmara Üniversitesi Atatürk Eğitim Fakültesi Eğitim Bilimleri Dergisi, 29, 57-74. Retrieved from http://e-dergi.marmara.edu.tr/maruaebd/article/view/1012001052/1012000903 . Çeliker, H. D., & Balım, A. G. (2012). Effects of Project Based Learning of the “Solar System and Beyond: Space Puzzle” Unit on Student Achievement. Kuramsal Eğitimbilim Dergisi, 5(3), 254-277. Retrieved from http://www.keg.aku.edu.tr/index.php/anasayfa/article/view/215/104 . Çil, A. (2005). Kimya eğitiminde proje tabanlı öğrenmenin incelenmesi ve öneriler. Yüksek Lisans Tezi, Hacettepe Üniversitesi Fen Bilimleri Enstitüsü, Ankara. Dede, Y., & Yaman, S. (2003). Fen ve matematik eğitiminde proje çalışmalarının yeri, önemi ve değerlendirilmesi. Gazi Üniversitesi Gazi Eğitim Fakültesi Dergisi, 23(1), 117-132. Retrieved from http://gefad.gazi.edu.tr/article/view/5000078861/5000073078. Değirmenci, Ş. (2011). Fen ve teknoloji dersinde “Canlılar ve Enerji İlişkileri” ünitesinin öğretilmesinde proje tabanlı öğrenmenin öğrenci başarısına etkisi. Yüksek Lisans Tezi, Selçuk Üniversitesi Eğitim Bilimleri Enstitüsü, Konya. Demirhan, C., & Demirel, Ö. (2003). Program geliştirmede proje tabanlı öğrenme yaklaşımı. Abant İzzet Baysal Üniversitesi Eğitim Fakültesi Dergisi, 3(1), 48-61. Retrieved from http://www.efdergi.ibu.edu.tr/index.php/efdergi/article/view/1100/2042 . Dilşeker, Z. (2008). Fen ve Teknoloji dersinde proje tabanlı öğrenme yöntemi kullanımının ilköğretim 5. sınıf öğrencilerinin fen ve teknoloji dersine yönelik tutumlarına, ders http://www.tebd.gazi.edu.tr/index.php/tebd/article/view/104/91 http://www.efdergi.hacettepe.edu.tr/yonetim/icerik/makaleler/1231-published.pdf http://web.mit.edu/monicaru/Public/old%20stuff/For%20Dava/Grad%20Library.Data/PDF/Brigid_1998DoingwithUnderstanding-LessonsfromResearchonProblem-andProject-BasedLearning-1896588801/Brigid_1998DoingwithUnderstanding-LessonsfromResearchonProblem-andProject-BasedLearning.pdf http://web.mit.edu/monicaru/Public/old%20stuff/For%20Dava/Grad%20Library.Data/PDF/Brigid_1998DoingwithUnderstanding-LessonsfromResearchonProblem-andProject-BasedLearning-1896588801/Brigid_1998DoingwithUnderstanding-LessonsfromResearchonProblem-andProject-BasedLearning.pdf http://web.mit.edu/monicaru/Public/old%20stuff/For%20Dava/Grad%20Library.Data/PDF/Brigid_1998DoingwithUnderstanding-LessonsfromResearchonProblem-andProject-BasedLearning-1896588801/Brigid_1998DoingwithUnderstanding-LessonsfromResearchonProblem-andProject-BasedLearning.pdf http://web.mit.edu/monicaru/Public/old%20stuff/For%20Dava/Grad%20Library.Data/PDF/Brigid_1998DoingwithUnderstanding-LessonsfromResearchonProblem-andProject-BasedLearning-1896588801/Brigid_1998DoingwithUnderstanding-LessonsfromResearchonProblem-andProject-BasedLearning.pdf http://e-dergi.marmara.edu.tr/maruaebd/article/view/1012001052/1012000903 http://www.keg.aku.edu.tr/index.php/anasayfa/article/view/215/104 http://www.efdergi.ibu.edu.tr/index.php/efdergi/article/view/1100/2042 International Online Journal of Education and Teaching (IOJET) 2018, 5(4), 849-865. 863 başarısına ve kavram yanılgılarının giderilmesine etkisi. Yüksek Lisans Tezi, Dokuz Eylül Üniversitesi Eğitim Bilimleri Enstitüsü, İzmir. Durlak, J. A., & Lipsey, M. W. (1991). A practitioner’s guide to meta-analysis. American Journal of Community Physcology, 19(3), 291-332. Retrieved from http://link.springer.com/article/10.1007/BF00938026 . Emrahoğlu, N., & Öztürk, A. (2010). Fen bilgisi öğretmen adayların akademik başarılarına bilişsel farkındalığın etkisi: bir nedensel karşılaştırma araştırması. Çukurova Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 19(2), 19-30. Retrieved from http://dergipark.ulakbim.gov.tr/cusosbil/article/view/5000001418/5000002109 . Ekiz, S. O. (2008). Fen ve teknoloji laboratuvarının proje tabanlı öğrenme yaklaşımı ile desteklenerek öğretiminin öğrenci başarısına, hatırda tutma seviyesine ve duyuşsal özelliklerine etkisinin araştırılması. Yüksek Lisans Tezi, Muğla Üniversitesi Fen Bilimleri Enstitüsü, Muğla. Ergene, T. (1999). Effectiveness of test anxiety reduction programs: a meta-analysis review. Unpublished PhD Thesis, Ohio University, Ohio. Ergül, N. R., & Kargın, E. K. (2014). The effect of project based learning on students’ science success. Procedia – Social and Behavioral Science, 136, 537-541. Retrieved from http://ac.els-cdn.com/S1877042814038531/1-s2.0-S1877042814038531- main.pdf?_tid=9b85ee7c-2751-11e5-b6f8- 00000aacb362&acdnat=1436566896_f914b7cbbd6680907cd4932e9f85400b . Filippatou, D., & Kaldi, S. (2010). The effectiveness of project-based learning difficulties regarding academic performance, group work and motivation. International Journal of Special Education, 25(1), 17-26. Retrieved from http://files.eric.ed.gov/fulltext/EJ890562.pdf. Gillies, R. M., & Ashman, A. F. (2000). The effects of cooperative learning on students with learning difficulties in the lower elementary school. The Journal of Special Education, 34(1), 19-27. Gültekin, Z. (2009). Fen eğitiminde proje tabanlı öğrenme uygulamalarının öğrencilerin bilimin doğasıyla ilgili görüşlerine, bilimsel süreç becerilerine ve tutumlarına etkisi. Yüksek Lisans Tezi, Marmara Üniversitesi Eğitim Bilimleri Enstitüsü, İstanbul. Güven, E. (2011). Çevre eğitiminde tahmin-gözlem-açıklama destekli proje tabanlı öğrenme yönteminin farklı değişkenler üzerine etkisi ve yönteme ilişkin öğrenci görüşleri. Doktora Tezi, Gazi Üniversitesi Eğitim Bilimleri Enstitüsü, Ankara. Hung, C. M., Hwang, G. J., & Huang, I. (2012). A Project-based digital storytelling approach for improving students learning motivation, problem solving competence and learning achievement. Educational Technology & Society, 15(4), 368-379. Retrieved from http://www.ifets.info/journals/15_4/31.pdf . Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis correcting error and bias in research findings. Newbury Park, California: Sage Publications. Retrieved from http://www.jstor.org/stable/2289738?seq=1#page_scan_tab_contents . İmer, N. (2008). İlköğretim fen ve teknoloji öğretiminde proje tabanlı öğrenme yaklaşımının öğrencilerin akademik başarı ve tutumuna etkisinin araştırılması.Yüksek Lisans Tezi, Gazi Üniversitesi Eğitim Bilimleri Enstitüsü, Ankara. http://link.springer.com/article/10.1007/BF00938026 http://dergipark.ulakbim.gov.tr/cusosbil/article/view/5000001418/5000002109 http://ac.els-cdn.com/S1877042814038531/1-s2.0-S1877042814038531-main.pdf?_tid=9b85ee7c-2751-11e5-b6f8-00000aacb362&acdnat=1436566896_f914b7cbbd6680907cd4932e9f85400b http://ac.els-cdn.com/S1877042814038531/1-s2.0-S1877042814038531-main.pdf?_tid=9b85ee7c-2751-11e5-b6f8-00000aacb362&acdnat=1436566896_f914b7cbbd6680907cd4932e9f85400b http://ac.els-cdn.com/S1877042814038531/1-s2.0-S1877042814038531-main.pdf?_tid=9b85ee7c-2751-11e5-b6f8-00000aacb362&acdnat=1436566896_f914b7cbbd6680907cd4932e9f85400b http://www.ifets.info/journals/15_4/31.pdf http://www.jstor.org/stable/2289738?seq=1#page_scan_tab_contents Balemen & Özer Keskin 864 Keskin, E. (2011). Proje tabanlı öğrenme yönteminin ilköğretim ikinci kademe öğrencilerinin başarı ve fen motivasyonlarına etkisinin incelenmesi. Yüksek Lisans Tezi, Uludağ Üniversitesi Eğitim Bilimleri Enstitüsü, Bursa. Korkmaz, H., & Kaptan, F. (2001). Fen eğitiminde proje tabanlı öğrenme yaklaşımı. Hacettepe Üniversitesi Eğitim Fakültesi Dergisi, 20, 193-200. Retrieved from http://www.efdergi.hacettepe.edu.tr/makale_goster.php?id=1005 . Köseoğlu, F, Tümay, H., & Budak, E. (2008). Bilimin doğası hakkında paradigma değişimleri ve öğretimi ile ilgili yeni anlayışlar. Gazi Üniversitesi Gazi Eğitim Fakültesi Dergisi, 28(2), 221-237. Retrieved from http://gefad.gazi.edu.tr/article/view/5000078593. Liu, M., & Hsiao, Y. (2002). Middle school students as multimedia designers: a project- based learning approach. Journal of Interactive Learning Research, 13(4), 311-337. Retrieved from http://search.proquest.com/openview/4fa03dae1dfd91552fbf53c210f8c910/1?pq- origsite=gscholar&cbl=2031153 . Nikbay, N. A. (2009). Anlatma yöntemi ile proje tabanlı öğrenme yönteminin öğrenci başarısı açısından verimliliğinin karşılaştırılması (besinlerin bozulma nedenleri ve bozulmayı engelleme yolları fermantasyon ünitesi).Yüksek Lisans Tezi, Gazi Üniversitesi Eğitim Bilimleri Enstitüsü, Ankara. Okursoy, G. F. (2009). Kavram haritaları öğretim stratejisinin öğrenci başarısına etkisi: Bir meta analiz çalışması. Marmara Üniversitesi, Eğitim Bilimleri Enstitüsü, İstanbul. Özahioğlu, B. (2012). İlköğretim fen ve teknoloji dersinde proje tabanlı öğrenmenin bilimsel süreç becerilerine, başarı ve tutum üzerine etkisi.Yüksek Lisans Tezi, Çanakkale Onsekiz Mart Üniversitesi Eğitim Bilimleri Enstitüsü, Çanakkale. Özbek, Ö. (2010). İlköğretim fen ve teknoloji dersinde küresel ısınma konusunun proje tabanlı öğretim modelinde incelenmesi. Yüksek Lisans Tezi, İnönü Üniversitesi Eğitim Bilimleri Enstitüsü, Malatya. Özdemirli, G. (2011). İşbirlikçi öğrenme yönteminin matematik başarısı ve matematiğe ilişkin tutumu üzerindeki etkililiği: Bir metaanaliz çalışması.Yüksek Lisans Tezi, Çukurova Üniversitesi Sosyal Bilimler Enstitüsü, Adana. Özer, D. Z., & Özkan, M. (2011). Proje tabanlı öğrenme yaklaşımının fen bilgisi öğretmenliği bölümü öğretmen adaylarının biyoloji konularındaki akademik başarılarına etkisi. Uludağ Üniversitesi Eğitim Fakültesi Dergisi, 24(1), 181-207. Retrieved from http://ucmaz.home.uludag.edu.tr/PDF/egitim/htmpdf/2011- 24(1)/M10.pdf. Özer, D. Z., & Özkan, M. (2012). Proje tabanlı öğretimin fen bilgisi öğretmen adaylarının biyoloji konuları ile ilgili akademik başarılarına etkisi. Uludağ Üniversitesi Eğitim Fakültesi Dergisi, 25(1), 87-102. Retrieved from http://ucmaz.home.uludag.edu.tr/PDF/egitim/htmpdf/2012-25(1)/M7.pdf . Rosenberg, M. S., Adams, D. C., & Gurevitch, J. (1997). MetaWin: Statistical software for meta-analysis with resampling tests. Version 1.0 Sunderland. MA: Sinauer Associates. Rosenthal, R. (1991). Meta-analytic procedures for social research. Sage publications: Newbury Park, California. Retrieved from http://www.amazon.com/Meta-Analytic- Procedures-Research-Applied-Methods/dp/080394246X . http://www.efdergi.hacettepe.edu.tr/makale_goster.php?id=1005 http://gefad.gazi.edu.tr/article/view/5000078593 http://ucmaz.home.uludag.edu.tr/PDF/egitim/htmpdf/2011-24(1)/M10.pdf http://ucmaz.home.uludag.edu.tr/PDF/egitim/htmpdf/2011-24(1)/M10.pdf http://ucmaz.home.uludag.edu.tr/PDF/egitim/htmpdf/2012-25(1)/M7.pdf http://www.amazon.com/Meta-Analytic-Procedures-Research-Applied-Methods/dp/080394246X http://www.amazon.com/Meta-Analytic-Procedures-Research-Applied-Methods/dp/080394246X International Online Journal of Education and Teaching (IOJET) 2018, 5(4), 849-865. 865 Serttürk, M. (2008). Fen öğretiminde proje tabanlı öğrenme yaklaşımının ilköğretim 7. Sınıf öğrencilerinin fen başarısı ve tutumuna etkisi. Yüksek Lisans Tezi, Sakarya Üniversitesi Sosyal Bilimler Enstitüsü, Sakarya. Şahin, M. C. (2005). İnternet tabanlı uzaktan eğitimin etkililiği: Bir meta analiz çalışması. Yüksek Lisans Tezi. Çukurova Üniversitesi Sosyal Bilimler Enstitüsü, Adana. Şahin, M. (2009). İlköğretim fen ve teknoloji dersinde proje tabanlı öğrenme yönteminin uygulanması ile ilgili öğretmen ve öğrenci görüşleri. Yüksek Lisans Tezi, Ondokuz Mayıs Üniversitesi Sosyal Bilimler Enstitüsü, Samsun. Toprak, E. (2007). Proje tabanlı öğrenme metodunun ilköğretim 5. Sınıf öğrencilerinin fen ve teknoloji dersindeki akademi başarısına etkisi. Yüksek Lisans Tezi, Marmara Üniversitesi Eğitim Bilimleri Enstitüsü, İstanbul. Tortop, H. S. (2010). Yapılandırmacı yaklaşıma göre hazırlanan alan gezisi ile desteklenmiş proje tabanlı öğrenme modelinin güneş enerjisi ve kullanım alanları konusunda uygulanması. Doktora Tezi, Süleyman Demirel Üniversitesi Fen Bilimleri Enstitüsü, Isparta. Tuncer, M., & Taşpınar, M. (2010). Sanal ortamda proje tabanlı öğrenme yönteminin öğrenci başarısına etkisi. Akdeniz Eğitim Araştırmaları Dergisi, (7), 104-140. Ünal, S., Coştu, B., & Karataş, F. Ö. (2004). Türkiye'de fen bilimleri eğitimi alanındaki program geliştirme çalışmalarına genel bir bakış. Gazi Üniversitesi Gazi Eğitim Fakültesi Dergisi, 24(2), 183-202. Retrieved from http://gefad.gazi.edu.tr/article/view/5000078796/5000073014 . Wolk, S. (1994). Project-based learning: pursuits with a purpose. Educational Leadership, 52(3), 42-45. Retrieved from http://web.a.ebscohost.com/ehost/pdfviewer/pdfviewer?vid=2&sid=95a3a9f9-7a78- 4ff7-a420-836d55d1c4e9%40sessionmgr4002&hid=4214 . Yıldız, N. (2009). Meta-analizinde heterojenliğin ve farklı varyans tahmin yöntemlerinin incelenmesi. Doktora Tezi, Marmara Üniversitesi Fen Bilimleri Enstitüsü, İstanbul. Yurttepe, S. (2007). İlköğretim fen bilgisi dersinde proje tabanlı öğrenmenin öğrenci başarısına etkisi. Yüksek Lisans Tezi, Eskişehir Osmangazi Üniversitesi Fen Bilimleri Enstitüsü, Eskişehir. http://gefad.gazi.edu.tr/article/view/5000078796/5000073014 http://web.a.ebscohost.com/ehost/pdfviewer/pdfviewer?vid=2&sid=95a3a9f9-7a78-4ff7-a420-836d55d1c4e9%40sessionmgr4002&hid=4214 http://web.a.ebscohost.com/ehost/pdfviewer/pdfviewer?vid=2&sid=95a3a9f9-7a78-4ff7-a420-836d55d1c4e9%40sessionmgr4002&hid=4214