Comparing U.S. and Japanese Inquiry-Based Science Practices in Middle Schools

The importance of scientific inquiry in K-12 education has been strongly advocated by the science education establishment in the United States through the publication of the National Science Education Standards (NESE) (National Research Council, NRC, 1996) and state standards. In spite of the emphasis on scientific inquiry, the literature suggests that little inquiry-based teaching is taking place in actual classrooms. For example, the Third International Mathematics and Science Study (TIMSS) 1999 Video Study of science (Roth et al., 2006) showed that 66% of the 88 U.S. science lessons for eighth graders were conducted mainly by student acquisition of facts, definitions, and algorithms. The study also showed that only 17% of the U.S. lessons developed science content through inquiry. In contrast, the same study showed that 57% of the Japanese eighth-grade science lessons developed content by making connections through inquiry. It seems that Japanese science lessons are more inquiry based than U.S. lesson.

However, there is some evidence that Japanese science teachers are just following prescribed ways of teaching science as inquiry (Ogura, 2004). In the independent research study based on the TIMSS 1999 videotaped lessons, Ogura found that U.S. science teachers were critical of Japanese lessons; U.S. teachers who watched 10 of the Japanese videotaped lessons commented that Japanese teachers were typically concerned with giving students “the right answers” to the results of their activities. Moreover, Japanese science teaching has been strongly criticized by Japanese media as not helping students develop their critical thinking abilities (Knipprath, 2010; Yomiuri Newspaper, 2007). In fact, the Course of Study, which is Japan’s national curriculum, was fully modified in 2008 and the new version will be in effect in 2012 (Ministry of Education, Culture, Sports, Science, and Technology, MEXT, 2008). The new Course of Study will place more emphases on student critical thinking, decision making, and communication abilities in elementary and secondary education.

Given a set of contradictory findings, it is important to conduct an in-depth study on science teaching in the United States and Japan. The TIMSS Video Study (Roth et al., 2006) showed that inquiry was more widely used in middle school science in Japan than in the U.S.

based on observational data collected by governmental agencies. The literature shows that U.S. and Japanese comparative studies were conducted before, focusing on student science achievement (Lawson, 1990; National Center for Educational Statistics, 2008; Organization for Economic Cooperation and Development, 2007; Takemura et al., 1985) or on general instructional approaches in elementary science (Linn, Lewis, Tsuchida, & Songer, 2000; Takemura & Shimizu, 1993). However, no study has shown a comparative analysis on the levels of science teachers’ inquiry orientation in the United States and Japan. This study identifies what might have been missed in the TIMSS Video Study by observing lessons in more private settings, and more importantly from a perspective that focuses on teachers’ support for students’ cognitive engagement. This study also examines what teachers understand inquiry-based teaching to be by conducting teacher interviews. This study further identifies how their understanding of inquirybased teaching is related to their practices. The examinations would enable us to gain insight on how to increase inquiry-based teaching in each of the countries.

The overall goal of the research project is to find out the interplay among middle-school science teachers’ thoughts, beliefs, and practices in regard to the use of inquiry, and to examine how U.S. teachers’ interplay would be different and similar to Japanese teachers’ interplay. However, for the purpose of organized presentation of the research findings, this paper is devoted to the examination of teachers’ thoughts and practices. A survey analysis was performed to measure U.S. and Japanese teachers’ attitudes and beliefs toward inquiry-based teaching as a part of the research project. A full discussion of the survey analysis will be presented in a separate paper because of the limitation in space. However, a brief description is provided later in this paper and the results are integrated in the discussions.

The following research questions guided the present study:

1. What are the similarities and differences between U.S. and Japanese middle-school science teachers’ teaching practices in terms of their use of inquiry?

2. What is U.S. and Japanese middle-school science teachers’ understanding of inquiry-based teaching?

3. How do U.S. and Japanese middle-school science teachers’ teaching practices relate to their understanding of inquiry-based teaching?

The purpose of this study is to inform science educators in the United States and Japan on what critical elements might be missing in their implementation of inquiry-based teaching in each place. Teaching is a cultural activity (Stigler, Gallimore, & Hiebert, 2000). It is difficult to see some of the taken-for-granted teaching practices and beliefs in one’s own culture. By comparing U.S. and Japanese teachers’ practices and thoughts in terms of their use of inquiry, this study reveals what is tacitly implied as inquiry-based teaching in each place. This study further provides U.S. and Japanese science educators in higher education with information about what should be more strongly emphasized in preservice and inservice teacher education programs in order to improve middle school science teachers’ instruction through the use of inquiry.

The theoretical framework guiding this study is constructivism, in which the notion that knowledge is actively constructed by learners through a viable interpretation of their experiences is addressed (von Glasersfeld, 1995). This study also adopts a view of constructivism that addresses the notion that knowledge represents negotiated meanings through social interactions (Vygotsky, 1978). Constructivism further stresses the importance of knowledge to be compatible with socially accepted viewpoints (Driver, Asoko, Leach, Mortimer, & Scott, 1994). The National Science Education Standards (NRC, 1996) are aligned with the tenets of constructivism, and this study follows NSES’s definition of inquiry-based teaching.

In this study, student learning of science through inquiry means their active development of understanding of scientific concepts by combining “hands-on” and “minds-on” activities (NRC, 1996). The NRC (2000) defined typical student inquiry activities as essential features of classroom inquiry. They are (1) student engages in scientifically oriented questions, (2) student gives priority to evidence in responding to questions, (3) student formulates explanations from evidence, (4) student connects explanations to scientific knowledge, and (5) student communicates and justifies explanations. The NSES suggest that the teacher’s role in the classroom is to facilitate these activities for students by guiding, focusing, challenging, and encouraging learning. Inquiry-based teaching in this study refers to teachers’ such teaching through facilitations. It should be noted that the NSES do not recommend a single teaching method or strategy as inquiry-based teaching. The NSES state that teachers should combine various instructional strategies in order to ensure student engagement in learning science through inquiry.

This study uses both quantitative and qualitative approaches to examine U.S. and Japanese science teachers’ teaching practices and understanding of inquiry-based teaching. Data on teaching practices were collected though observations. The observational data were coded based on a rubric. A quantitative treatment of the numerical values of the coding followed. Observational data were also treated qualitatively through fieldnotes, coders’ descriptive notes, and video transcripts. The qualitative data provide information about the context of each of the lessons, including the content, the teacher, and the students. Data on teachers’ understanding of inquiry-based teaching were collected qualitatively through interviews which were then transcribed and coded. Types of teachers’ understanding of inquiry-based teaching were identified through typical data analysis techniques involved in qualitative research such as coding, pattern finding, and recursive processing with consistency checking (Glesne, 2006).

Participants were first recruited through mailings to randomly selected middle schools. Participants were also recruited through a request that was sent out through listserv networks for science teachers in the United States and Japan. A total of 24 teachers in 21 schools participated in this study, including nine U.S. and 15 Japanese teachers. One of the Japanese teachers participated only in an interview due to a schedule conflict. All the participants were full-time science teachers at public middle schools (Grades 6-8 in the United States and Grades 7-9 in Japan). Their teaching experience ranged from 3 months to more than 10 years. Three Japanese participants were teachers at a national public middle school in Tokyo (selective-admission school that serves as a model school attached to a national university). Twelve Japanese teachers were science teachers at typical public middle schools in the Tokyo area. All nine U.S. teachers were teaching science at typical local public middle schools in Massachusetts. Among the 21 schools, 13 schools (4 United States and 9 Japanese) were in inner-city school districts, whereas eight schools (4 United States and 4 Japanese) were in suburban school districts.

Observational Rubric. The observed lessons were coded based on a rubric that describes how each of the five essential features of inquiry-based teaching was practiced. The rubric included the X scale with three levels based on student self-directedness (structured inquiry = 1, guided inquiry = 2, and open inquiry = 3). The National Research Council (2000) describes the variation in four levels. This study adopted and modified it into a three-level rubric. Essential features of inquiry classroom at each level of student self-directedness are described in Table 1. A scientific fact referred in Table 1 means a verbal or mathematical statement of a scientific concept, whereas scientific knowledge means students’ cognitive construction of a scientific concept.

The variation in student self-directedness has been used in previous research studies (Smolleck, Zembal-Saul, & Yoder, 2006; Wee, Shepardson, Fast, & Harbor, 2007). However, the variation in student self-directedness does not necessarily represent the lesson’s orientation toward inquiry (NRC, 2000). It only represents how much structure the teacher provides students. The NRC describes that the level of student self-directedness in a lesson should be determined by the intended learning outcome. If students are not ready for carrying out inquiry activities, teachers should provide students with more structured lessons so that students would be able to develop their abilities to inquire. If the learning outcome is placed on the development of particular science concepts, guided inquiry would be most appropriate. If the development of understanding of the nature of scientific inquiry is the learning goal, more open inquiry might provide students with more opportunities for it (NRC).

In order to measure inquiry orientation of science lessons, another scale Y was incorporated in the rubric. The new scale describes the variations in how teachers provide cognitive scaffolds to support students in making connections among their experiences, ideas, data, and explanations (Borko, Kuffner, & Arnold 2007; Roth et al., 2006). The variation in conceptual links has not been used in the research literature before as an explicit, independent, continuum scale for measuring classroom orientation toward inquiry. The new scale measures, for example, as the essential feature of formulating explanations, how teachers facilitate activities for students by asking questions that would help them generate more sophisticated explanations that includes “a rich scientific knowledge base, evidence of logic, higher levels of analysis, greater tolerance of criticism and uncertainty” (NRC, 1996, p. 117). The Y scale measures the depth of conceptual links in five levels from 1 to 5. The level 1 represents a practice in which no facilitation for students to generate conceptual links occurs. The level 5 in the Y dimension represents a classroom practice in which teacher facilitates activities for students in such a way that they would be able to generate multiple links among their experiences, ideas, data, and explanations. Essential features of inquiry classroom at each level of depth of conceptual links are described in Table 2.

The rubric employed in this study has a feature of measuring the extent to which teachers practice inquiry-based teaching in two dimensions: student self-directedness and depth of conceptual links (see Figure 1). Each lesson received five X values for indicating the levels of student self-directedness for five essential features of inquiry classroom. Each lesson also received five Y values for indicating the depth of conceptual likes made for five essential features. A total of 10 numbers were given for each lesson. Descriptive notes and comments were also given for each of the 10 categories by coders so that the reasons for giving the numerical values were clearly recorded. The coders’ notes and comments were saved as data.

Widely used observational protocols for measuring teachers’ inquiry-orientation, such as the Reformed Teaching Observational Protocol (RTOP) (Sawada et al. 2000), do not differentiate the depth of conceptual links from the level of student self-directedness. Furthermore, RTOP does not explicitly include the five essential features of inquiry activities. It should be noted that the two-dimensional consideration for measuring classroom orientation toward inquiry is analogous to the diagrammatic representation of student learning proposed by Novak and Gowin (1984). Novak and Gowin proposed that the levels of student learning can be mapped on a two-dimensional plane with the reception-discovery learning dimension as the X axis and the rote-meaningful learning dimension as the Y axis. Novak and Gowin’s work was done for student learning. Obviously student learning and teacher teaching happen in the same classroom. The similarity to Novak and Gowin’s representation for student learning gives a support for the validity for the idea of incorporating two dimensions for analyzing teachers’ practices in the present study.

Interview Protocol. An interview protocol was created as part of this study. The interview questions were constructed in such a way as to collect data in three categories for identifying patterns in participants’ understanding of inquiry-based teaching: (a) definitions of inquiry-based teaching, (b) obstacles to inquiry-based teaching, and (c) advantages of inquiry-based teaching. The interview protocol included three questions regarding the observed lessons and five questions regarding teacher’s thoughts about inquiry-based teaching. Table 3 shows the preplanned interview questions in the protocol. The beginning questions enabled the participants to describe their thoughts about the use of inquiry on a concrete basis. This part of the interview provided the teachers with an opportunity to give an additional explanation about the lesson. In the latter part of the interview, the teachers were asked to describe their general point of view toward inquiry-based teaching. The teachers were asked to provide their thoughts about inquiry-based teaching in terms of its definition, obstacles, effectiveness, and the future practice. Interview data were all transcribed and coded to identify teachers’ understanding of inquiry-based teaching.

The newly created instruments were reviewed by three experts in science education independently for content validity. The content and format of the instruments were checked for their appropriateness. The wordings of interview questions were revised several times based on the experts’ comments. The interview protocol was then submitted to the Institutional Review Board for permission for conducting research on human subjects. Regarding the observational rubric, reliability was checked through the intercoder rating. Construct-related evidence of validity was established through checking the consistency between the numerical coding of the data and the qualitative description of the lessons denoted in the fieldnotes, coders’ notes, and video data.

The researcher visited each of the participants’ schools once to conduct a classroom observation and an interview. Artifacts including lesson plan and blank worksheets for students were also collected at the time of the site visit if the teacher had not submitted them electronically in advance. The participants responded to a survey that served to provide teachers’ demographic information. The researcher conducted school visits with 15 Japanese science teachers at 13 public middle schools in the Tokyo area in June and July, 2008. June and July are part of the first term in the Japanese school calendar. Nine U.S. teachers’ classrooms at eight public middle schools in the state of Massachusetts were also visited during the first term of their school year in late September and October in 2008. All of the 23 observed lessons were their regular lessons, and the lessons and teacher interviews were all videotaped with permission from the teachers and the principals of each school. Fieldnotes were taken by the researcher at the time of the site visits, and the participating teachers were asked to submit their lesson plan. However, half of the Japanese teachers refused to submit their lesson plans due to their busy schedules, so the analysis on the lesson plans was not possible. The fieldnotes, the lesson plans, and blank student worksheets were used to supplement the videotaped data.

Observational data were collected from nine U.S. and 14 Japanese science lessons in grades 6 to 9. Among the observed lessons, topics in physical science were taught in 18 classes (7 U.S. and 11 Japanese). Topics in life science were covered in four classes (1 U.S. and 3 Japanese). A topic in earth science was taught in one U.S. lesson. The average number of students in a lesson was 23.2 for U.S. lessons and 33.4 for Japanese lessons. The average lesson time was 56 minutes for U.S. lessons and 48 minutes for Japanese lessons. Textbooks were used in 33% of U.S. lessons and in 100% of Japanese lessons as an explicit source of information for activities and descriptions that students used in class.

All the observed lessons were coded by two individuals using the rubric described in the previous section. The researcher of this study was one of the coders and the cocoder was a graduate student at a major U.S. university. Both of the coders were fluent in English and Japanese languages. Intercoder correlations were calculated using the scores given by the two coders. Pearson’s two-tailed product moment was 0.735, showing acceptable intercoder reliability. Having shown reliable intercoder rating, the researcher’s data were chosen to be used for further analyses. This choice was made based on the assumption that better understanding of the phenomenon in question would take place with richer data in contextual dimensions (Glesne, 2006). Obviously, the researcher who was able to be in the actual classrooms could collect more data (i.e., fieldnotes) from the teachers than the cocoder who had only a limited view on the videotaped lessons. Table 4 shows the results of the mean scores and the standard deviations in the student self-directedness scale (X scale) and the depth of conceptual link scale (Y scale) for each of the essential features observed in U.S. and Japanese lessons.

A multivariate analysis of variance with a country as a factor variable was conducted for 10 dependent variables. Using Wilks’ Lamda test, the combined dependent variables resulted in significant main effects for country with an F value of 14.39 and p = .001. Effect size measured by partial eta-squared was .954 which is considered to be large (Cohen, 1992). The univariate analysis of variance for each dependent variable followed to examine the effect of country on each of the 10 variables. Table 4 shows the results. There were statistically significant main effects for country in four of the categories. For example, when students engage in scientifically oriented questions, there was a significant main effect for country for the depth of conceptual links formed during the lesson with an F value of 6.88, p = .018, and η² = .30. However, the statistical results should be interpreted cautiously because of the small sample size (N = 23). In the following presentation of the results, examples found in qualitative data are supplemented to give context and support for the quantitative results.

Student self-directedness level was significantly higher in U.S. lessons (M = 2.33) than in Japanese lessons (M = 1.23) when students formulated explanations from the results of their activities. For example, in one of the U.S. lessons, “Students wrote down their explanations about the densities of the liquids on the worksheet based on the result they got in the experiment. Teacher asked them to write their explanations in just one sentence” (researcher’s coding notes). In contrast, in one of the Japanese lessons, the cocoder noted, “Teacher simply stated the correct answer. Complex cases were not discussed. Teacher proceeded to follow textbook and lesson material rather than attempting to explain the different answers that were presented by students.” In another Japanese lesson, the cocoder noted, “Teacher connected observations from the demonstration to scientific concepts under discussion. Alternate explanations were not given. Teacher derived her answers from observations in logical sequence”. These examples indicate that the student self-directedness in formulating explanations would be less emphasized in Japanese lessons than in U.S. lessons.

When students communicated explanations verbally or in written forms, the level of student self-directedness was shown to be significantly higher in U.S. lessons (M = 1.22) than in Japanese lessons (M = .79). In three of the Japanese lessons, students did not communicate their explanations and this resulted in lower mean scores. In both U.S. and Japanese lessons, however, 78% of teachers gave students direct instructions to communicate their explanations and these lessons received a score of 1 for this category.

In terms of the depth of conceptual links displayed in lessons, teachers’ facilitation for students in formulating conceptual links was at a significantly deeper level in the U.S. lessons (M=3.89) than in Japanese lessons (M = 2.93) when students were introduced to scientific questions at the beginning of lessons. The conceptual links were also formed at a significantly deeper level in U.S. lessons (M = 2.67) than Japanese lessons (M = 1.71) when student discuss the methods of investigation. The following transcript from video data illustrates how questions were introduced through a discussion based on the previously learned fact in one of the U.S. lessons:

In this U.S. lesson, the teacher further asked students to suggest procedure for conducting an experiment for determining the densities of salt water and fresh water. He also asked students to write down their predictions about the densities in numbers. The following transcript from another U.S. class shows how the teacher incorporated students’ daily experience to start the discussion of waves during earthquakes:

Then, the teacher referred to the earthquake video they watched before in order to introduce the idea of different kinds of waves. These examples show that the U.S. teachers were introducing the questions by soliciting student responses based on what they experienced or learned before. In contrast, the following transcript describes how one of the Japanese teachers started the lesson:

As the cocoders described this lesson, “Teacher simply presented the problem by drawing figures on the board and asking students to follow procedure. Teacher did not ask for student input and/or discussion.” The cocoder noted the following in another Japanese lesson:

These qualitative data suggest that students received more prompts for understanding the meaning of the questions in U.S. lessons than in Japanese lessons. The data also suggest that Japanese teachers often lectured on the concepts and asked students questions about procedures rather than questions about connections among the scientific concepts under the investigation.

In spite of the stronger conceptual links displayed in the beginning of U.S. lessons, when it came to student making connections of their explanations to scientific knowledge, the average Y scores were lower for U.S. lessons (M = 1.19) than for Japanese lessons (M = 1.80). In four out of nine U.S. lessons, the coders could not identify scientific concepts to which the outcomes of the student activities should have connected. These lessons received low Y scores. The following examples drawn from cocoder’s comments made for three of the U.S. lessons illustrate this point. In a sixth grade class for the topic of mass and volume, “Teacher explained the procedure rather than the results. They talked about how to find mass using a triple-beam balance and how to find volume using the formula. The procedure provided answers for defining vocabulary terms rather than conceptual explanations.” In an eighth-grade lesson about the topic of mass, students measured the mass of chewing gum before they chewed and after they chewed using a triple-beam balance, and identified the mass difference as the mass of sugar included in the gum. The cocoder described the lesson in the following way: “Not sure what scientific concept is being discussed. Just because mass and changes in mass are being discussed does not mean an actual scientific concept is under discussion. The teacher did not provide an explicit discussion about mass”. In the density lesson quoted earlier as an example of high level of conceptual links at the beginning of the lesson, the teacher were concerned with the numerical values of the densities in the closure part of the lesson, and they did not have an explicit discussion about the concept of density. In addition, the lesson ended with student confusion because they obtained smaller numerical values for salt water density than the egg density. Perhaps the amount of salt given in the experimental procedure was too small to make the salt water saturated. Students did not test whether their salt water would float the egg or not.

In contrast, Japanese lessons on the average were able to provide some connections between student explanations and scientific concepts at the end of the lessons. The following example illustrates how a Japanese teacher connected student observations to a scientific concept:

Then, the teacher asked students to predict how the needle of a magnetic compass would be deflected differently when it is placed near each of the poles of the bar magnet or side of the magnet. The cocoder described the lesson in the following way: “Teacher attempted to connect facts about the patterns of iron filing to scientific concepts of magnetic field lines starting with basics configurations of magnets and expanding into more complex cases.” Although not all the Japanese lessons were successful in making connections between student explanations and scientific concepts during the observed lessons, the coders were able to identify the scientific concepts that the teachers intended to develop through the activities in all but one Japanese lessons.

Interview data were collected from nine U.S. and 15 Japanese teachers. All interviews were videotaped and transcribed by the researcher. The data were then coded according to the three categories described earlier. In addition, interview data were coded if the teachers’ descriptions included any of the following five categories: (a) what constitutes student learning, (b) the teachers’ depth of understanding of subject-specific knowledge, (c) the teachers’ understanding of effectiveness of different instructional strategies in inquiry-based teaching, (d) the teachers’ depth of understanding of the ways to transform scientific concepts into teachable forms in inquiry-based teaching, and (e) the teachers’ understanding of the nature of scientific inquiry.

Each of the interviews was summarized and tabulated according to the eight categories. In order to find patterns in teachers’ understanding of inquiry-based teaching, similar phrases and words used by teachers were identified in their definitions of inquiry-based teaching. Three types of teachers’ understanding of inquiry-based teaching were identified: (1) students’ hands-on activities, (2) students’ own explorations of concepts, and (3) teacher’s help for students to make connections. Two U.S. and one Japanese teacher thought that lessons that merely provide students with hands-on activities represent inquiry-based teaching. Seven U.S. and 13 Japanese teachers mentioned that inquiry-based teaching includes students’ own explorations of scientific concepts. Two teachers, one U.S. and one Japanese, mentioned that neither students’ own exploration, nor hands-on activities are the necessary components of inquiry-based teaching. Rather they mentioned that in inquirybased teaching students come up with mental questions continuously and the teacher helps them build logical connections in various ways so that students understand scientific concepts.

Additionally, six subcategories were identified among the teachers who thought that inquiry-based teaching includes students’ own exploration of scientific concepts. For this refinement of the classification, interview data in all eight categories were examined. Teachers’ comments on their instructional strategies and their thoughts on student learning were frequently used to examine how they embed inquiry in their practice. When all of the teachers were classified into types, the consistency was checked among the teachers who were in the same type. Reclassification of teachers took place and a subcategory was added or deleted when inconsistencies were found. The analysis followed a recursive pattern until consistency was achieved for all of the types of teachers. Table 5 shows the classifications and the number of teachers who belong to each of the types by country.

Among the teachers who defined inquirybased teaching as students’ own exploration of scientific concepts, the subcategories indicate the variations in their usage of inquiry. Brief explanations of each type follow. Two Japanese teachers whose definition of inquirybased teaching is Type 2a thought that students’ own exploration would not be practical in classroom practice because of constraints such as time, materials, and teacher’s help. Type 2b definition of inquiry-based teaching by a U.S. teacher is called algorithmic because he thought that inquiry-based teaching is an instructional practice that gives students train ing to develop logical, sequential thinking skills through their routine practice. The teachers whose definition is Type 2c thought that inquiry-based teaching is separate from students’ acquisition of scientific knowledge; they thought that students can acquire scientific knowledge more effectively through lecture-type lessons. The teachers who had Type 2d definition were strong advocates of studentcentered, open-ended activities and they would not call their lessons inquiry-based unless students control the direction of their learning. The four Japanese teachers’ definition of inquiry-based teaching in Type 2e is called scenario-driven teaching styles because they thought that inquiry-based teaching happens only in well-prepared lessons where scientific questions at the starting point and answers at the end point are fixed and students are expected to explore based on their own ideas in between. Lastly, the teachers who had the definition in Type 2f are the ones who thought that inquiry-based science teaching provides students with opportunities to conduct scientific investigations in a similar way as scientists do.

In order to find a relationship between teachers’ understanding of inquiry-based teaching and their practices, mean X and mean Y scores were calculated for each of the lessons. The mean X score indicates how much student self-directedness was involved in the lesson over the range of the five essential features of inquiry practice. Similarly, the mean Y score indicates how deeply conceptual links were made during the lesson. Figure 2 shows a scatter plot of the observational data in a twodimensional plane along the mean X and mean Y scales. The types of the teachers’ understanding of inquiry-based teaching are shown as the data labels. Figure 2 shows that the lessons delivered by Type 1 teachers received low scores both in the X and Y scales. This result makes sense because teachers who identify inquiry-based teaching just as hands-on activities would probably be less focusing on student self-directedness and conceptual links than merely carrying out activities during lessons. It should be noted that student engagement in hands-on activities does not guarantee their understanding of scientific concepts (NRC, 1996, 2000), and the definition of inquiry-based teaching by Type 1 teachers does not agree with the description given by the NSES. Besides the findings about Type 1 teachers, no apparent relationships between the data plot and the types of teachers’ understanding of inquiry-based teaching are identified. For example, teachers who were classified as Type 2f are scattered over the figure.

This study examined differences and similarities between U.S. and Japanese science teachers’ practices and their understanding of inquiry-based teaching at the middle-school level. The findings from the interview data showed that there are variations in teachers’ definitions of inquiry-based teaching among U.S. teachers as well as Japanese teachers. In spite of the variations, the observational data indicated that essential features of classroom inquiry (NRC, 2000) were observed to a certain extent in all of the U.S. and Japanese lessons in this study. The extent to which student self-directedness was emphasized was significantly higher in U.S. lessons than in Japanese lessons when students formulated and communicated explanations. The U.S. teachers also helped students form conceptual links significantly more explicitly than Japanese teachers did when they discussed questions and methods for investigation at the beginning of the lessons. When it came to connecting explanations to scientific knowledge, however, scientific concepts were difficult to be identified in some of the U.S. lessons. Scientific concepts under discussion were identified in all but one Japanese lesson. The findings indicate that students in U.S. lessons were more likely to have been exposed to the discussion of why and what they would do in the investigation, but they would have few chances to connect their results and explanations with scientific concepts that they need to understand. On the other hand, students in Japanese lessons were more likely to start activities without having many chances to understand what the experiment means, but they would have more chances to be mechanically presented with the science concepts involved in the lesson.

The survey analysis was performed separately from the observation and interview part of the research project. An online survey instrument that contains 36 Likert-type questions was created and measured teachers’ (a) general preference of inquiry-based teaching, (b) personal evaluation toward performing typical behavior in inquiry-based teaching, (c) beliefs about possible obstacles to inquirybased teaching, and (d) self-efficacy about the use of inquiry in science teaching (Tosa, 2009). A total of 191 (57 U.S. and 134 Japanese) middle-school science teachers responded to the survey. The results showed that the extent to which Japanese teachers felt comfortable in helping students with their activities and questions was significantly lower than U.S. teachers. These survey results are consistent with the observation data in which no incidence of teachers’ active help was observed during the Japanese lessons; the Japanese teachers did not take an active role for helping individual students understand what the activity meant or what the results meant. Perhaps for many Japanese teachers, helping students means giving them correct answers or prescribed methods of investigation. One of the Japanese teachers mentioned in the interview that “I don’t think student abilities to inquire would be developed if teacher provides them with everything. For example, students need to know the usefulness of graphs for finding a pattern in data and ask me for graphing paper instead of me providing graphing paper from the beginning”. Waiting for students until they come up with correct answers might be a well-known strategy in Japan (Fujita, 2009). However, this view seems to contradict their role described in the NESE (NRC, 1996) where teachers’ active involvement for helping students construct their own knowledge is emphasized. In contrast, in some of the U.S. lessons, effective instructional strategies such as Socratic questioning (Knezic, Wubbels, Elbers, & Hajer, 2010) were used and several U.S. teachers mentioned the importance of these interactive techniques in the interviews.

Through the interview data, it became clear that the majority of middle-school science teachers in the United States and Japan thought that having students explore their own ideas has to be an essential part of inquiry-based teaching. Although the ultimate goal for science teaching would be fully student selfdirected open inquiry that is scientifically meaningful, student active engagement does not necessarily mean open inquiry (NRC, 1996, 2000). Teacher-directed structured inquiry can generate student engagement as well as their solid understanding of concepts through teachers’ guidance and encouragement. Many teachers’ misidentification of open inquiry as an essential part of inquirybased teaching might be related to the lack of inquiry-based teaching found in this study. Perhaps many of the U.S. teachers were expecting students to “discover” scientific concepts and did not provide much cognitive help when they generated explanations. Perhaps many of the Japanese teachers were not aware of the importance of the teacher role for supporting student learning by asking thought provoking questions because inquiry means student own exploration. This study calls for a clarification of the meaning of inquiry-based teaching among science teachers. The emphasis placed on student self-directedness in scientific research community (Smolleck et al., 2006; Wee et al., 2007) should also be considered for an examination. Student self-directedness should be treated separately from student conceptual understanding in inquiry-based teaching. By using an instrument that has two dimensions, this study showed that a richer picture of inquiry-based teaching could be obtained across different countries than the previous studies.

The selection of the sample puts a serious limitation on this study. The U.S. and Japanese samples consisted of middle-school science teachers who voluntarily responded to the request for participating in this research. It is highly likely that the participating teachers are more favorable toward inquiry-based teaching than other teachers. Neither the U.S. nor the Japanese sample would represent all middleschool science teachers in the country. However, names of middle-school science teachers are not publicly accessible in the United States or in Japan, so for the feasibility of conducting research, the sample was selected on a limited basis. The sample size of 24 in the present study was small for a statistical treatment of data so the numerical results should be treated cautiously.

Reliability of the qualitative data analyses may need to be discussed for accuracy of data collection and coding. Data were collected in the United States and in Japan by the researcher alone who speaks both Japanese and English. Appropriate measures were taken in order to minimize the researcher’s bias. Observational data were cocoded by an individual who understands both Japanese and English. The intercoder reliability was checked and reasonable consistency was obtained. All of the interview data were transcribed and analyzed in the original languages. No threats for losing or changing the meaning of data in the process of translation happened in this cross-cultural study.

Further investigation is underway that includes teachers in elementary schools in Japan and the United States. It is expected that with more widely collected data, this study will reveal more clearly what teachers assume as inquiry-based teaching within their culture. It will be interesting to study further how teachers’ assumptions about inquiry-based teaching are connected with their practice. Through these analyses, it may be possible to suggest more effective ways to improve teacher education programs in both countries.

This work was in part supported by CAREER grant provided by the National Science Foundation under the grant number 0546513.