Comparison of problem-based learning studies in science education in Turkey with the world: Content analysıs of research papers

The studies in which PBL method was used in science education in the journals examined were subjected to a content analysis and the findings related to each research question are given below. Among the 104 papers whose content analysis was made 15.4% were in national journals while 84.6% were international journals (See Table 1). And 21.1% of the studies were written in Turkish while 78.9% of them were written in English. Of all the papers in this study, 42 of them are papers of Turkish authors in national and international journals and 62 are papers of foreign authors in international journals.

Table 1. Classification by journal type and language of paper and nationality of the authors

Type of Journal	f	%	The language of paper	f	%	Nationality of the authors	f	%
National	16	15.4	Turkish	22	21.1	Turkish	42	40.3
International	88	84.6	English	82	78.9	Mixed	62	59.7

Table 2 and Figure 1 are made to show the development of PBL method in science education.

Table 2. Number of paper related to PBL over years (1986-2012; N=104).

	Year	1986	1995	1997	1998	1999	2000	2001	2002	2003	2004	2005	2006		2007	2008	2009	2010	2011	2012		Total
TR		--	--	--	--	2	--	1	--	--	--	7		2	6	6	7	5	3		3	42
INT		1	2	2	2	3	1	1	4	5	5	3		8	3	3	7	4	3		5	62
Total		1	2	2	2	5	1	2	4	5	5	10		10	9	9	14	9	6		8	104

When Table 2 and Figure 1 are examined, it is seen that the first publication on PBL in science education was in 1986 (Birch, 1986) in “Studies in Higher Education” journal and the second publication was in 1995 (Gallagher et al., 1995; Savery & Duffy, 1995) in “School Science and Mathematics” and “Educational Technology” journals. In Turkey, on the other hand, the first publication on PBL method in science education was published in 1999 (Cakir & Tekkaya, 1999; Unal, 1999) in the journal of “Hacettepe University Journal of Education” and “Marmara University Journal of Ataturk Educational Faculty of Educational Sciences”. Among these two journals “Hacettepe University Journal of Education” has been indexed in ERIC since 2004 and in SCI/SSCI since 2007.

According to Table 2 and Figure 1, the highest number of papers on PBL in science education by foreign authors hit the peak in 2006 while the peak was hit in 2005 and 2009 for Turkish authors. While foreign authors had predominance over Turkish authors until 2005 in terms of papers on PBL in science education, as of 2005, except the year 2006, the number of papers of Turkish authors has been equal to/higher than the number of papers by foreign authors until 2012. And in 2012, again the papers of foreign authors on PBL in science education outnumbered those of Turkish authors.

Figure 1. Number of national and international papers published across years (1986-2012).

As a result of classifying the papers according to “Paper Classification Form”, Figure 2 was formed., It is seen that 35.6% of the studies are mixed, while 30.8% are on chemistry, 13.5% are on science and technology, 11.5% are on biology and 7.7% are on physics (see Figure 2). Also, according to Figure 2, of all the studies on PBL in science education by Turkish authors, 33.3% is mixed, while 23.8% is on chemistry, 19.1% is on science and technology and 11.9% on physics and biology. Of all the papers of foreign authors in this study, on the other hand, it is seen that 37.1% is mixed, 35.5% is on chemistry, 11.3% is on biology, 9.7% is on science and technology and 4.8% is on physics.

Figure 2. Discipline

Identifying which subject matters both Turkish and foreign authors prefer in their papers on PBL in science education is another research question in this study. In Figure 3, it is seen that 76% of the studies whose content analysis was made is about teaching and concept analysis (40.4% teaching and 35.6% concept analysis). Identifying attitude-interest and developing course/ project/ activity/ scenario for PBL make up the 7.7% of all studies.

It is seen that in studies on PBL in science education Turkish authors prefer teaching in 59.5% of their studies, concept analysis in 26.2%, attitude-interest determining in 11.9% and scale-test development - translation in 2.4% of their studies. Foreign authors, on the other hand, prefer concept analysis in 41.9% of the studies, teaching in 27.4% of the studies, developing course/project/activity/scenario for PBL in 12.9% of the studies, teacher training and computer-aided teaching in 3.2% of the studies and learning, assessment and evaluation based on PBL and workshop based on PBL in 1.6% of the studies.

Table 3 details which sub topics are examined in learning, teaching and teacher training subjects, which are frequently studied both by Turkish and foreign authors. The table was prepared by taking into consideration if studies cover more than one sub-topics of teaching, learning and teacher training.

In Table 3, it is seen that in 25% of all papers, the impact of teaching on success was examined while in 24% the impact of teaching on attitude and in 20,2% the impact of teaching on scientific process skills were examined. In 42.9% of the studies of Turkish authors, the effect of teaching on success was examined; while in 38.1% the impact of teaching on attitudes and in 30.9% the impact of teaching on scientific process skills were examined. In 12.9% of the studies by foreign authors on PBL in science education, on the other hand, the impact of teaching on success was examined; while in 14.5% the impact of teaching on attitudes and in 12.9% the impact of teaching on scientific process skills were examined.

Figure 3. Frequently investigated subject matters by the paper on PBL science education researches

Again Table 3 demonstrates that success level and identifying learning styles under learning; preservice teacher education and in-service training under teacher training, and comparing effectiveness of different teaching method, methods under teaching constitute a small number of the studies whose content analysis was made, and these subjects were only preferred by foreign authors.

Table 3. Frequently investigated sub-topic of learning, teaching and teacher training

Subject matter		TR		INT		Total
Subject matter		f	%	f	%	f	%
Learning	To determine the level of success	---	---	1	1.6	1	0.9
Learning	Learning style identification	---	---	1	1.6	1	0.9
Teaching	Teaching methods on students’ academic achievement	18	42.9	8	12.9	26	25.0
	Teaching methods on students’ attitudes	16	38.1	9	14.5	25	24.0
	Teaching methods on students’ scientific process skills	13	30.9	8	12.9	21	20.2
	Compared the effectiveness of different teaching methods	---	---	1	1.6	1	0.9
Teacher training	Preservice Teacher Education	---	---	2	3.2	2	1.9
Teacher training	Inservice Training	---	---	1	1.6	1	0.9

The research methods used in the papers on PBL in science education are given in Figure 4. According to Figure 4, while qualitative research methods were preferred in 50.9% of the studies, quantitative research methods were preferred in 33.6% of the studies and mixed research methods were preferred in 10.6% of the studies. It is understood that Turkish authors preferred quantitative research methods (61.9%) more than qualitative (28.6%) and mixed research methods (2.4%), while foreign authors preferred qualitative research methods (66.1%) more than quantitative research methods (14.5%) and mixed research methods (16.1%).

Figure 4. Frequently used research design/methods on PBL in science education

Figure 5 includes which of qualitative, quantitative and mixed methods- are preferred by both Turkish and foreign authors in their studies on PBL in science education.

It is seen that in 35.6% of the studies, non-interactive research, a qualitative research method, was used. Experimental research as a, quantitative research method was preferred in 27.9%. Interactive, mixed and non-experimental research methods; were used in 19.2%, 10.6%, and 5.8%, respectively. It is also seen that in the papers of Turkish authors, 57.1% preferred experimental research method while 26.2% preferred non-interactive research methods; 7.1% preferred interactive research methods, 4.8% preferred non-experimental methods; 2.4% preferred mixed research methods. Foreign authors, on the other hand, preferred non-interactive research methods in 41.9% of their studies; interactive research methods in 27.4% of their studies; mixed research methods in 16.1% of their studies; experimental in 8.1% of their studies and non-experimental research methods in 6.5% of their studies.

Figure 5. Frequency of research design/methods on PBL

Figure 6 is formed to identify which research methods among qualitative, quantitative and mixed research methods are preferred in studies on PBL in science education.

Looking at Figure 6, it is understood that Turkish authors preferred quantitative research methods in their studies on PBL in science education and in 47.6% of their studies they used quasi-experimental research; while in 23.8% of their studies they preferred concept analysis; in 9.5% of the studies they preferred pre-experimental; in 7.1% of the studies they preferred case study and in 2.4% of the studies they preferred review, correlation, survey and triangulation. It is also figured out from the figure that foreign authors, on the other hand, preferred qualitative research methods in their studies and preferred concept analysis in 35.5% of their studies; case study in 27.4% of their studies; triangulation in 11.3% of their studies; quasi-experimental and review in 6.5% of their studies; survey, comparative and explanatory research design in 3.2% of their studies; pre-experimental and exploratory research design in 1.6% of their studies.

Figure 6. Frequently used research design/methods on PBL in science education

Figure 7 is made to show the frequency of the research methods that Turkish and foreign authors preferred in their studies by years.

Figure 7. Trends in research designs across years (1986-2012)

According to Figure 7, in 11.9% of the studies that Turkish authors published in 2009, they preferred quantitative research methods the most. Foreign authors, however, seem to have preferred qualitative research methods most in 11.3% of their studies in 2006.

Frequently preferred data collection tools in studies on PBL in science education are given in Table 4. While forming Table 4, whether more than one data collection tool is used in one study was considered. And while calculating the percentages for Turkish authors, the number of the data collection tools used was calculated by dividing them by total number of papers that Turkish authors made, whereas the same thing was done by dividing the number of data collection tools by total number of papers that foreign authors published.

Table 4. Frequently used data collection tools in science education researches

Type of data collection tools	TR		INT		TOTAL
Type of data collection tools	f	%	f	%	f	%
Achievement tests	16	38.1	11	17.7	27	25.9
Questionnaires	17	40.5	16	25.8	33	31.7
Interviews	8	19.0	18	29.0	26	25.0
Observations	---	---	15	24.3	15	14.4
Alternative assessment tools	5	11.9	17	27.4	22	21.2
Perception, attitude, skill, personality tests etc.	13	30.9	5	8.1	18	17.3
Documents	1	2.4	1	1.6	2	1.9

According to Figure 4, questionnaires were used in 37.1% of the papers on PBL in science education; while achievement test was used in 25.9% of the studies; interviews were used in 25% of the studies; alternative evaluation tools were used in 21.2% of the studies. Other scales like perception, attitude, skill, personality tests etc. were used in 17.3% of the studies; observation was used in 14.4% and documents were used in 1.9% of the studies.

It is seen that Turkish authors preferred questionnaires the most (40.5%) in their studies on PBL in science education. It is followed by achievement tests (38.1%), perception-attitude-skill-personality etc. scales (30.9%), interviews (19.0%), alternative evaluation tools (11.9%) and documents (2.4%) respectively. Foreign authors, on the other hand, seem to have preferred interviews in most of their studies (29.0%); and interviews were followed by alternative evaluation tools (27.4), questionnaires (25.8%), observation (24.3%), achievement test (17.7%), perception- -attitude-skill-personality etc. scales (8.1%) - and documents (1.6%), respectively.

In the studies whose content analysis was made using paper classification form, the type of questionnaires preferred were identified as well as the type of questions used in achievement tests. The type of observations and interviews used in these studies were also included. Among the total 33 papers in which questionnaires were used as data collection tools, likert-type questionnaires were preferred in 26 of them - in 13 papers of Turkish and foreign authors each while in 9 papers open-ended questionnaires were preferred - in 4 Turkish and 5 foreign authors’ papers. And among 27 papers in which achievement test was used, multiple-choice questions were preferred in 24 of them- in 16 Turkish and 8 foreign authors’ studies- while open-ended questions were preferred in 10 - in 6 Turkish authors’ and 4 foreign authors’ studies- and questions grouped as “other” in 2 studies- in 1 Turkish author’s study and 1 foreign author’s study. In 21 of the studies, among 26 in which interview was used as data collection tool, semi-structured interview was used (in 7 Turkish and 14 foreign authors’ papers), while in 2 of them, focus group discussion was preferred (in 1 Turkish author’s and 1 foreign author’s paper). In 3 studies that belong to foreign authors, the type of interview was not specified. All 15 studies in which observation was used as data collection tool belong to foreign authors; and among these studies, nonparticipant observation was preferred in 11 studies. In 4 studies, the type of observation made was not given.

Figure 8 includes the frequency of data collection tools preferred by Turkish and foreign authors in their studies on PBL in science education. In 35.6% of the studies analyzed, there were no data collection tools used or specified. The rate of studies in which only one data collection tool was used was 25.0%; while the rate of studies in which 2 data collection tools were used was 19.2%, and the rate of studies in which 3 and more data collection tools were used was 20.2%.

Figure 8. Frequency of different data collection tools

Figure 8 reveals that in 30.2% of their studies, Turkish authors preferred 2 data collection tools while in 27.9% of their studies, they preferred one data collection tool and in 13.9% of the studies, they preferred three or more data collection tools. Foreign authors, on the other hand, preferred three or more data collection tools in 24.2% of their studies; one data collection tool in 22.6% of their studies and two data collection tools in 11.3% of their studies.

The frequently preferred samples in the studies on PBL in science education are given in Figure 9. In 36.5% of the studies, whose content analysis was made, samples were not given. In 35.6% of the studies, undergraduate students were chosen as the sample, while in 13.5% secondary school students (9-12) were chosen as the sample; in 9.6% primary school students (6-8); in 3.8%, teachers; in 1.9% primary school students (1-5) and in 0.9%, faculty members were chosen as study samples. In 42.9% of the studies carried out by Turkish authors on PBL in science education, undergraduate students were preferred; while in 16.7% secondary school students (9-12); in 11.9% primary school students (6-8) and in 2.4% primary school students (1-5), faculty members and teachers were preferred. In 26.2% of the studies by Turkish authors, the samples of the study was not specified. On the other hand, in 30.6% of the studies by foreign authors on PBL in science education, undergraduate students were preferred as the samples of the study; while in 11.3% secondary school students (9-12); in 8.1%, primary school students (6-8); in 4.8%, teachers and in 1.6%, primary school students (1-5) were preferred as the population of their studies. And in 43.6% of the studies of foreign authors, the samples of the study was not given.

Figure 9. Frequently studied samples.

Figure 10 is made regarding the sample sizes. When Figure 10 is analyzed, it is seen that in 39.4% of the studies on PBL in science education, sample sizes is not specified. It is seen that in 30.8% of the studies, sample size was between 31-100; while in 15.4%, it was between 101-300; in 7.7%, it was 11-30; in 2.9%, it was 1-10 and again in 2.9%, it was 301-1000 and in 0.9%, it was over 1000. It is seen from the Figure 10 that in 40.5% of the studies by Turkish authors and in 24.2% of the studies by foreign authors, sample sizes was between 31-100; while in 23.8% of the studies by Turkish authors and 9.7% of the studies by foreign authors, sample sizes was between 101-300. And in 7.1% of the studies by Turkish authors and in 8.1% of the studies by foreign authors, sample sizes was 11-30; while in 2.4% of the studies by Turkish authors and 3.2% of the studies by foreign authors, sample sizes was between 301-1000. Again, according to Figure 10, Turkish authors did not work with sample sizes between 1-10 and over 1000 and that in 26.2% of the studies of Turkish authors, they did not specify sample sizes of their studies. Foreign authors, on the other hand, preferred 1-10 sample sizes in 4.8% of their studies on PBL in science education and over 1000 sample sizes in 1.6% of their studies; and in 48.4% of their studies, they did not mention study sample sizes.

Figure 10. Frequently studied sample sizes.

Frequently preferred data analysis methods of both Turkish and foreign authors on PBL in science education are given in Table 5 and Figure 11.

Table 5. Frequently used data analysis methods and techniques

			TR		INT		TOTAL
QUANTITATIVE ANALYSIS	Descriptive statistics (%54.0)		f	%	f	%	f	%
		f / % tables	28	25.2	28	28.0	56	26.5
		Central tendency measures	26	23.4	19	19.0	45	21.3
		Charts	4	3.6	9	9.0	13	6.2
		Others	---	---	---	---	---	---
	Inferential statistics (%29.9)	t-testi	18	16.2	8	8.0	26	12.3
		ANOVA/ANCOVA	12	10.8	4	4.0	16	7.6
		MANOVA/MANCOVA	5	4.5	---	---	5	2.4
		Factor analysis	2	1.8	---	---	2	0.9
		Correlation	4	3.6	1	1.0	5	2.4
		Regression	1	0.9	3	3.0	4	1.9
		Non-Parametric Tests	2	1.8	3	3.0	5	2.4
QUALITATIVE ANALYSIS	Qualitative analysis (%16.0)	Content analysis	2	1.8	4	4.0	6	2.8
		Descriptive analysis	7	6.3	21	21.0	28	13.2
		Others	---	---	---	---	---	---
		Total	111	100	100	100	211	100

According to Table 5 and Figure 11, in 54.0% of the studies, whose content analysis was made, descriptive analysis, one of the quantitative data analysis methods, was used; while in 29.9% inferential analysis, another quantitative data analysis methods, was used; and in 16.0% qualitative data analysis method was preferred. In 26.5% of the studies, whose content analysis was made, frequency/percent tables were the most preferred methods; while in 21.3% central tendency measures; and in 13.2% qualitative descriptive analysis methods were preferred methods. In 25.2% of the studies of Turkish authors on PBL in science education, frequency/percent tables were preferred; while in 23.4%, it was the central tendency measures that was preferred; in 16.2% it was the t-test and in 10.8%, ANOVA/ANCOVA was preferred the most. In 28.0% of the studies by foreign authors on PBL in science education, on the other hand, frequency/percent tables were frequently used ; while in 21% of their studies it was the qualitative descriptive analysis and central tendency measures in 19.0% .

Figure 11. Frequently used data analysis methods and techniques

Figure 12 shows the number of data collection tools preferred both by Turkish and foreign authors in only one study.

Figure 12. Number of different data analysis methods combined in a study

When Figure 12 is examined, it is seen that two different data analysis methods were the most preferred methods in the studies. That rate is 36.5%. And in 13.5% of all the studies, one and three different data analysis methods were preferred. In 36.5% of the studies, whose content analysis was made, the data analysis method was not specified. Turkish authors preferred two different data analysis method in 52.4% of the studies they made while the rate is 38.1% in the studies of foreign authors. In 14.3% of the studies, Turkish authors preferred three different data analysis methods while foreign authors preferred three different data analysis method in 19.0% of their studies. And while Turkish authors preferred one data analysis method in 7.1% of their studies, foreign authors preferred one data analysis method in 26.2% of their studies.