Teacher-Directed Versus Inquiry-Based Science Instruction: Investigating Links to Adolescent Students’ Science Dispositions Across 66 Countries

ABSTRACT

Teacher-directed and inquiry-based science instructional practices have been shown to influence students’ performance on science assessments. However, only a small body of research has examined the associations of teacher-directed and inquiry-based science instructional practices with science-related dispositions among adolescent students using nationally representative samples drawn from countries across the globe. Hence, the present study, employing multilevel path analyses as an analytic strategy, investigated the relations of teacher-directed and inquiry-based science instruction to students’ science-related dispositions, such as enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science, among 428,197 adolescent students from 15,644 schools in 66 countries. Results of multilevel path analyses, after controlling for student-, school-, and country-level demographic and socio-economic factors, revealed that teacher-directed science instruction was significantly positively related to adolescent students’ enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science. Similarly, inquiry-based science instruction was also significantly positively linked to enjoyment of science, interest in broad science topics, instrumental motivation to learn science, and science self-efficacy. However, inquiry-based science instruction was not significantly associated with students’ epistemological beliefs about science. The findings of the current study suggest that a blend of teacher-directed and inquiry-based science instruction may be more appropriate for developing and nurturing students’ positive dispositions toward science. However, science teachers may require sufficient training and support to successfully implement the blended instruction model in their classrooms.

KEYWORDS:

• Teacher-directed instruction
• inquiry-based instruction
• enjoyment of science
• interest in science
• instrumental motivation
• science self-efficacy
• epistemological beliefs
• self-determination theory

There is growing evidence that the number of students choosing to pursue STEM programs and careers has been dwindling over the past few decades in many countries across the world (see Jones et al., 2018; Marginson, Tytler, Freeman, & Roberts, 2013; Murphy, MacDonald, Danaia, & Wang, 2018; Smith & White, 2018). However, building a highly competent, educated, and versatile STEM workforce is crucial for developing and sustaining a global knowledge-based economy (see Freeman, Marginson, & Tytler, 2015). It is therefore of critical importance to identify the factors that may enthuse students to pursue STEM disciplines, such as science that cuts across all STEM subjects. One of the proximal factors that might have a large bearing on students’ interest and motivation to pursue studies in science-related STEM fields is classroom practices in science instruction, such as teacher-directed and inquiry-based science instructional methods.

Although a growing body of research has examined the relations of both teacher-directed and inquiry-based science instructional methods to students’ performance on science assessments (e.g., Areepattamannil, 2012; Areepattamannil, Freeman, & Klinger, 2011; Cairns & Areepattamannil, 2019; Lavonen & Laaksonen, 2009; McConney, Oliver, Woods-McConney, Schibeci, & Maor, 2014; Organization for Economic Co-operation and Development [OECD], 2016), there is a dearth of research investigating the relations of these instructional approaches to students’ science-related dispositions, such as enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science among adolescent students employing nationally representative samples of students drawn from countries across the globe. Such research is imperative for better understanding the relative importance of teacher-directed and inquiry-based science instruction in predicting adolescent students’ positive dispositions toward science. Furthermore, how well teacher-directed and inquiry-based science instructional practices interact with students’ science-related dispositions to affect their science achievement is still largely unknown. Hence, the purpose of the present study was to investigate the relative strength of their roles—teacher-directed and inquiry-based science instruction—in predicting students’ science-related dispositions, such as enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science, across 66 Program for International Student Assessment (PISA) 2015 participating countries.

Literature review

Inquiry-based and teacher-directed instruction

It is widely accepted by the science education community that inquiry-based instruction is a highly effective methodology (Furtak, Seidel, Iverson, & Briggs, 2012), and with this comes the implication that inquiry-based instruction is a more effective instructional strategy than other methods such as teacher-directed instruction.

There is a strong theoretical case for inquiry-based instructional approaches (see Constantinou, Tsivitanidou, & Rybska, 2018; Spronken-Smith, 2008). Inquiry-based instruction is largely based on constructivist theory and is an inductive teaching (“bottom-up”) approach (Prince & Felder, 2006). That is, the individual learner constructs their own understandings (of reality) from their experiences and interactions during the learning process. This new information is filtered through knowledge structures, known as schemas, which allow for knowledge that aligns with, builds on, or replaces existing schemas to be incorporated into the learner’s long-term memory. Through this process they experience inductive learning. Inquiry-based instruction, however, is a highly complex model of teaching and learning and is not limited to only constructivist and inductive learning theories or indeed other specific approaches, such as project-based or problem-based learning (Constantinou et al., 2018). Although there is much debate regarding the definition of inquiry-based instruction (Anderson, 2002; Minner, Levy, & Century, 2010), one simple principle can be applied to all learning activities and teaching approaches used therein and that is, “do not give the answers in advance”. This implies an increased level of autonomy and self-direction combined with a shift in responsibility regarding moving to the next level of learning (Constantinou et al., 2018).

Inquiry-based instruction also aligns with theories relating to approaches to learning. For example, Säljö (1979) reported that “deep approaches” to learning result in the learner perceiving reality in a different way than prior to the learning experience. This is in contrast to more traditional “surface approaches” leading to the accumulation of information for reproduction at a specified time, such as an exam. Furthermore, inquiry-based instructional approaches offer opportunities for learners to develop higher-order cognitive skills that allow for the application of their deeper understanding of scientific principles to everyday phenomena, and personal and societal issues (Anderson, 2002; Constantinou et al., 2018). Further, including the features of authentic science inquiry (such as generating research questions and controlling for multiple dependent variables) in school-based tasks encourages the development of epistemological awareness and a better understanding of how scientific knowledge is created (Chinn & Malhotra, 2002). In terms of learning cycles, inquiry-based instructional methods are compatible with experiential learning models and “involve[s] a degree of autonomy” (Constantinou et al., 2018, p. 9). The active, more self-directed, and student-centered nature of inquiry-based instruction allows students to act as participants in knowledge generation rather than as passive audiences that receive knowledge (see Jungert & Koestner, 2015; Patall, Hooper, Vasquez, Pituch, & Steingut, 2018). Although research efforts in this area often focus on the research-teaching nexus in higher education (see Brew, 2003), the argument for inquiry-based instruction in schools (particularly in science lessons as part of a simulated research environment) is to reflect the authentic practices of scientists as they generate scientific knowledge (National Research Council, 1996).

Unlike inquiry-based instruction, the term “teacher-directed instruction” is not fully conceptualized in the literature. However, there is some agreement that high-quality teacher-directed instruction in science should involve instructional features such as guided practice with regular feedback, modeling, clearly defined goals, learning intentions and success criteria, independent practice, and hands-on practical work (Cobern et al., 2010; Hattie, 2009). As such, high-quality teacher-directed instruction should not be conflated with practices such as rote learning, exposition, and highly directed laboratory experiments (Ausubel, 1961). Many features of inquiry learning are likely to be present in (high quality) directed instruction. Note that “experientially-based” instruction and “active student engagement” are certainly advantageous for effective science learning but such “hands-on” and “minds-on” aspects can occur when using inquiry-based or teacher-directed instructional approaches (Cobern et al., 2010, p. 82).

Inquiry-based instruction and achievement

There is a considerable body of research that reports the positive effects of inquiry-based instruction in terms of both attainment and dispositions to learning science (see Furtak et al., 2012). Several meta-analyses carried out in the 1980s reported the benefits of an inquiry-based curriculum approach (Bredderman, 1983; Lott, 1983; Shymansky, Kyle, & Alport, 1983; Weinstein, Boulanger, & Walberg, 1982; Wise & Okey, 1983). Shymansky et al. (1983) found that inquiry-based instruction led to high effect sizes but was unrelated to the quantity of inquiry-based instruction content in the teaching and learning materials. Wise and Okey (1983) analyzed a range of methodologies in their meta-analysis and reported an effect size of 0.41 for inquiry-based methodologies when compared with traditional methods of teaching; however, the term “traditional teaching” was not defined. Hattie’s (2009) synthesis of meta-analyses included four studies involving inquiry-based instruction (Bangert-Drowns & Bankert, 1990; Shymansky, Hedges, & Woodworth, 1990; Sweitzer & Anderson, 1983) which produced an effect size of 0.31. A recent, comprehensive meta-analysis of 37 studies reported an effect size of 0.50 for inquiry-based teaching techniques (Furtak et al., 2012).

However, when analyzing data drawn from the PISA 2006 using multilevel modeling techniques, negative associations between inquiry-based instruction and achievement are observed. Using an inquiry-based instruction construct based on student investigations and hands-on activities, negative correlations with science attainment were reported in Qatar (Areepattamannil, 2012), Canada (Areepattamannil et al., 2011), Finland (Lavonen & Laaksonen, 2009), Italy (Gee & Wong, 2012), and Mexico (Gee & Wong, 2012). Similar findings were demonstrated by McConney et al. (2014) for over 40,000 students across Australia, Canada, and New Zealand and in a study of 54 countries including 170,474 students by Cairns and Areepattamannil (2019). Recently, Jerrim, Oliver, and Sims (2019), using the data for England from the most recent cycle PISA and the National Pupil Database (NPD), found that frequent use of inquiry-based instruction was not associated with higher achievement in science. Furthermore, secondary analyses of the Trends in International Mathematics and Science Study (TIMSS) 2015 Norwegian data, employing multilevel structural equation modeling as the analytic technique, also attest that frequent use of inquiry-based instruction is negatively related to science achievement (Teig, Scherer, & Nilsen, 2018). The negative correlations reported from these cross-sectional studies suggest that inquiry-based instruction may be less effective as a method of science instruction than previously reported by the experimental literature. In other words, in most studies, students that experienced high frequencies of inquiry-learning approaches in their science lessons, on average, also exhibited lower levels of science achievement.

Teacher-directed instruction and achievement

There are empirical studies that suggest employing inquiry-based instruction in science education over teacher-directed instruction (e.g., Chang, 2001), teacher-directed instruction over inquiry-based instruction (e.g., Klahr & Nigam, 2004), and studies that find no statistical difference between the methodologies (e.g., Cobern et al., 2010). Although Klahr and Nigam (2004) compared teacher-directed instruction and discovery learning, the students engaged in discovery learning “were allowed to design experiments on their own, without any guidance or feedback from the experimenter” (Klahr & Nigam, 2004, p. 3). The findings of this study showed that, when teaching the control-of-variables strategy (CVS) to 3rd and 4th graders, a greater number of students that experienced teacher-directed instruction mastered the concepts of CVS and, this group, were equally as adept at applying these concepts to broader, authentic contexts as students that mastered CVS through discovery learning (Klahr & Nigam, 2004). As explained by Kirschner, Sweller, and Clark (2006) this type of open inquiry can lead to cognitive overload and result in students not accessing the main learning intentions of an activity. These sub-optimal open inquiry conditions may have contributed to the relative success of the teacher-directed instruction treatment (Furtak et al., 2012; Hmelo-Silver, 2011; Salamon, Perkins, & Globerson, 1991). The effectiveness of explicit, directed instruction and inquiry-based research when teaching different types of outcomes was investigated by Flick (1995). Outcomes that required the acquisition of factual knowledge and specific science process skills benefited from directed instructional approaches, whereas inquiry-based teaching methods led to improved attainment in outcomes that measure higher-order thinking but only in a supportive classroom environment with highly trained teachers and higher ability students.

Cobern et al. (2010) attended to validity issues when comparing the two modes of instruction by building features into their research framework such as fidelity, specificity, transparency, subjects and settings. With these measures in place, no statistical difference was observed for science concept development between teacher-directed and inquiry-based instructional methodologies (Cobern et al., 2010). Hitherto, however, there is a lack of evidence for the benefits of inquiry-based learning when compared to teacher-directed instruction in terms of science attainment in the research literature.

A more recent study investigated a blended approach involving a discovery and explicit instruction phase (Chase & Klahr, 2017). The results of this study indicated that the order of receiving these modes of instruction to learn CVS did not significantly affect learning outcomes (or far transfer of these concepts) and that other contextual factors were likely dominant. Another study, produced by McKinsey & Company, suggested the need for a blended approach to science teaching using the PISA 2015 results for Europe. They reported that the optimum frequency of experiencing inquiry-based instruction approaches was shown to be “in some classes” with teacher-directed instruction recommended as a pre-requisite for successful learning (Denoël et al., 2017).

Student dispositions

In this study, we investigated the associations between the previously discussed instructional approaches (inquiry-based and teacher-directed instruction) and student dispositions. Students’ attitudes toward science learning are of interest because the research literature indicates that positive student dispositions are related to improved science attainment. For example, a study by Singh, Granville, and Dika (2002), using data from the National Education Longitudinal Study 1988, revealed a strong relationship between motivation and a positive attitude toward academic success in science. A project that examined the motivation of 14-16-year-old students to learn science and apply for the Advanced Placement Program found higher levels of intrinsic motivation, self-efficacy, self-determination, and achievement in those that applied (Bryan, Glynn, & Kittleson, 2011). The self-reported reasons for these differences included collaborative learning, inspiring teachers, and future career interests. Another investigation of 422 secondary students in different education systems found a positive correlation between attitudes toward science learning and science achievement (Narmadha & Chamundeswari, 2013).

The dispositions included in this study are enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs. The attitudinal scales measured by the PISA 2015 questionnaire are derived from the motivational constructs of intrinsic and instrumental motivation (OECD, 2016). These constructs are underpinned by self-determination theory (Ryan & Deci, 2009) and expectancy-value theory (Wigfield & Eccles, 2000), respectively.

Science self-efficacy

The interrelation between knowledge and action (and the often-observed misalignment) has been explained through self-percepts of efficacy that, in turn, affect motivation and behavior (Bandura, 1982). Self-efficacy is described as a mechanism that allows people to generate and regulate events in their lives. A strong initial sense of self-efficacy is also more likely to lead to persistence and mastery of an action—leading to higher attainment (Bandura, 1982).

Science self-efficacy is defined by the OECD as referring to an individual’s assessment of their competency in performing science-related tasks such as designing experiments, explaining scientific phenomena, and interpreting and making inferences from data (OCED, 2016). The importance of science self-efficacy stems from the additive effect on attainment. Namely, a student that performs well when learning science will likely receive positive feedback from teachers, peers, and parents leading to increased levels of self-efficacy. This can, in turn, lead to further improvements in attainment through higher levels of perseverance and effort. Domain-specific self-efficacy also tends to improve in accuracy in older children (e.g. 15-year-olds) as they develop a greater ability to decode feedback from teachers, parents, or their peers (Wigfield & Eccles, 2000).

Epistemological beliefs about science

There is a long-held assumption that an individual’s epistemological beliefs affect comprehension leading to a number of studies in the 1980s relating to personal epistemologies (e.g., Schoenfeld, 1983). Schommer (1990) showed that epistemic beliefs were multi-faceted and that epistemological persuasions such as the belief in the existence of “certain knowledge” led to poor performance in conclusion writing. Furthermore, a belief in “quick learning” led to poor performance overall. There is evidence that epistemological beliefs have a degree of domain generality (Buehl, Alexander, & Murphy, 2002; Schommer & Walker, 1995), but they are, to some extent, also domain-specific (Buehl et al., 2002; Hofer, 2000; Trautwein & Lüdtke, 2007).

According to Hofer (2000, 2004); Hofer and Pintrich (1997) epistemological beliefs within the scientific knowledge domain consist of four dimensions: simplicity of scientific knowledge, which includes beliefs about knowledge as a collection of discrete facts or as complex, interconnected and subject to interpretation; certainty of scientific knowledge concerns beliefs about whether knowledge is fixed or constantly developing; source of scientific knowledge includes beliefs about whether knowledge is outside the self and transferred from sources of authority to seeing oneself as a constructor of knowledge; justification of knowledge is identified as the perception of knowledge on the basis of what feels right to the use of the inquiry process to test the validity of knowledge. According to conventional definitions of epistemology, the dimensions of certainty and simplicity belong to the area known as the nature of knowledge (what is knowledge?) and the dimensions of source and justification belong to the nature or process of knowing (how do we know?).

The effects of epistemological beliefs on science learning have been studied in various contexts. Mason, Gava, and Boldrin (2008) investigated the effect of the epistemological dimensions of the nature and development of scientific knowledge (simple vs. complex and fixed vs. continuously evolving) in relation to improving conceptual understanding in science. They observed high levels of conceptual change for students that had more advanced beliefs about scientific knowledge.

Well-developed scientific epistemological beliefs were found to be related to higher academic performance in a natural science course and (particularly in terms of the dimension of “certainty/simplicity”) a higher grade point average (GPA) overall, for first-year college students (Hofer, 2000). These studies indicate that advanced epistemological beliefs are beneficial when acquiring and understanding scientific conceptual knowledge. In the PISA 2015 student questionnaire, participants were asked about their epistemic beliefs about science through questions pertaining to the validity of science knowledge and the processes by which is it generated (OECD, 2016).

Intrinsic and instrumental motivation

The overall construct of intrinsic motivation consists of the enjoyment of science and interest in broad science topics scales. Enjoyment of science in this study is considered to be one component of intrinsic motivation. Namely, the motivation to engage in an activity is inherent in the enjoyment derived directly from the activity itself (OECD, 2016). The PISA 2015 documentation implies that this construct is grounded in self-determination theory, which consists of inherent tendencies toward intrinsic motivation and integration (Ryan & Deci, 2009). Here, integration refers to the process by which students that feel cared for, safe and valued want to internalize knowledge and skills (i.e., learn) even if the topic or practice in question is not inherently interesting or enjoyable.

Another component of intrinsic motivation is interest, in this case, interest in broad science topics. The distinguishing motivating characteristic of interest, as opposed to enjoyment, is the necessity of an object of interest. The object can be a concrete thing, a certain topic or even a more abstract concept (Krapp & Prenzel, 2011). In the case of this study, the person in question is interested in specific content: science-related topics.

As stated above, the instrumental motivation construct employed in the PISA 2015 cycle is based on the expectancy-value theory model of achievement proposed by Eccles (1985). This model predicts achievement choices (e.g. a person deciding to enroll as a science major) as linked to two core beliefs: the individual’s expectations for success by making this choice, and the value individual places on the various options available (Eccles, 1985). As such, the student questionnaire used in PISA 2015 asks participants to what extent learning science is relevant to their future prospects, in terms of the next phase of their studies or career choices (OECD, 2016).

Instructional methods and motivation

The affective benefits of inquiry-based instructional approaches in schools are well documented in the literature (Andre, Whigham, Hendrickson, & Chambers, 1999; Crawford, 2000; Gott & Duggan, 1995; Holbrook & Kolodner, 2000; Mac Iver, Young, & Washburn, 2001; Madden, 2011; McConney et al., 2014; Palmer, 2009; Patrick, Mantzicopoulos, & Samarapungavan, 2009; Tuan, Chin, Tsai, & Cheng, 2005). Students involved in an investigation-based project, ranging from 11 to 15 years old, reported considerably higher levels of interest and enjoyment when learning science using this approach (Gott & Duggan, 1995; McConney et al., 2014). A recent study involving PISA 2006 data for Australia, Canada, and New Zealand demonstrated a positive correlation between the reported frequency of inquiry-based instruction and the reported enjoyment of learning science for 15-year-olds (McConney et al., 2014).

There is also evidence for motivational benefits of inquiry-based instruction in science education (Palmer, 2009; Tuan et al., 2005). Extrinsic motivational attitudes to learning science, such as future-career-related instrumental motivation, were found to develop as early as pre-school age in a study by Andre et al. (1999). Andre et al. (1999) recommended tackling this issue at the early stages of elementary school education through parental education and increased use of inquiry-based instructional approaches. An inquiry-based instructional approach that led to increased motivation for 8th grade students was reported by Tuan et al. (2005) where student motivation (as measured by the students’ motivation toward science learning questionnaire) was significantly higher for the treatment group than the control group experiencing “traditional” instruction. Overall confidence in the learning of science (self-efficacy) has been shown to improve through the inquiry-learning approach (Madden, 2011; Patrick et al., 2009).

There is a paucity of studies relating to the effect of direct instruction (as the treatment) on student motivation in science. Furthermore, research projects typically use terms such as traditional, didactic, and expository teaching (e.g., Lord & Orkwiszewski, 2006; Schroeder, Scott, Tolson, Huang, & Lee, 2007; Selim & Shrigley, 1983; Tuan et al., 2005) which can be falsely associated with direct instruction. This methodology, then, is usually the control condition for quantifying the effect of another treatment such as discovery learning and is poorly defined. However, some features of high-quality direct instruction (as described previously), such as goals coupled with learning intentions and success criteria, have been related to attitudinal variables. A study involving 1016 Australian students ranging from Year 7 to Year 12 revealed a positive relationship between challenging, specific goals and motivation and engagement (Martin, 2006), and informative, regular feedback that is linked to the learning intentions and success criteria was shown to have positive effects on motivation (e.g., Narciss & Huth, 2004). Clearly, as stated above, the research in this area is very limited.

There is evidence in the literature that inquiry-based instruction can be used as a vehicle for developing epistemological beliefs, leading to improved science attainment (Furtak et al., 2012; Minner et al., 2010). In a synthesis of research from 1984 to 2002, Minner et al. (2010) found that instruction that focused on developing the epistemic domains of inquiry led to greater learning gains. Some critics of the inquiry-based instructional approach argue that the epistemology of an expert working within their domain is considerably different than that of a novice (Kirschner, 1992) and subsequently, the epistemology of a practicing scientist is not a good basis for the pedagogy used to teach children science (Kirschner, 2009). Klahr and Nigam (2004) showed that not only was teacher-directed instruction more effective for teaching inquiry process skills (in this case, the controlling of variables in experiments) but also students that experienced teacher-directed instruction were equally as proficient at transferring this knowledge to authentic contexts as students that experienced a discovery learning approach (Klahr & Nigam, 2004). These findings call into question the claim that direct instruction is not an appropriate pathway for students to learn about and succeed in authentic inquiry contexts, particularly as efforts to align the epistemology of practicing scientists with that of inquiry-based instruction in schools are largely driven by this principle (Chinn & Malhotra, 2002)

The present study

The review of the literature clearly indicates that there is still no consensus amongst researchers as to the relative effectiveness of two predominant instructional practices, teacher-directed and inquiry-based instruction, for teaching science in schools. Moreover, most studies to date that examined the efficacy of these two instructional practices primarily focused on gains in learning and achievement (e.g., Furtak et al., 2012; Minner et al., 2010; Stockard, Wood, Coughlin, & Khoury, 2018). However, there is a paucity of research on the relative strengths of teacher-directed and inquiry-based science instruction as potential determinants of science-related dispositions, such as enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science, among adolescent students in countries across the globe. Although researchers have employed a wide array of motivational frameworks for exploring the relations of teachers’ instructional behaviors with students’ academic dispositions, one perspective that appears particularly promising and pertinent for the current study is Deci and Ryan’s (1985) motivational framework, self-determination theory. Hence, the present study, employing self-determination theory (Deci & Ryan, 1985; Ryan & Deci, 2000, 2002) as the theoretical framework, sought to investigate how well these two instructional practices are associated with students’ positive dispositions toward science. Self-determination theory postulates that teachers’ instructional behaviors can either support or hinder students’ learning dispositions (Jang, Reeve, & Deci, 2010; Reeve, Jang, Carrell, Jeon, & Barch, 2004). Teachers who adopt an autonomy-supportive instructional style tend to promote more positive dispositions toward learning among their students, whereas teachers who adopt a controlling instructional style tend to thwart students’ inner motivational resources (Reeve & Jang, 2006). Unlike teacher-directed instructional behaviors, inquiry-based instructional behaviors offer a broad range of autonomy supports rather than autonomy thwarts (Rogat, Witham, & Chinn, 2014). Reeve at al. (2014) elaborate on the ways in which teachers’ instructional behaviors act as vehicles of or obstacles to autonomy support as follows:

Autonomy-supportive teachers tend to adopt their students’ perspectives, welcome their students’ thoughts, feelings, and actions into the flow of the lesson, and support their students’ developing capacity for autonomous self-regulation, while controlling teachers tend to adopt only their own perspective, intrude into their students’ thoughts, feelings, and actions, and pressure their students to think, feel, and behave in a teacher-prescribed way. (p. 94)

Thus, teachers’ instructional behaviors, autonomy-supportive inquiry-based instructional style or teacher-directed controlling instructional style, may affect students’ learning dispositions and academic performance. However, for a deeper understanding of the nature of the relationship between teacher-directed as well as inquiry-based science instructional practices and science achievement, it is important to account for the effects of proximal factors that may mediate the relations between science instructional practices and science achievement. A voluminous body of research has demonstrated that science-related dispositions, such as enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science, are proximal indicators or antecedents of science achievement (e.g., Cairns & Areepattamannil, 2019; Greene, Cartiff, & Duke, 2018). Moreover, there is growing evidence that adolescent students who hold positive dispositions toward science are more likely to choose STEM majors or careers (e.g., Lamb et al., 2018; Sahin, Ekmekci, & Waxman, 2017). Thus, science-related positive dispositions play a pivotal role in shaping adolescent students’ academic and career trajectories. Nonetheless, little is known about which factors or combination of factors account for differences in adolescent students’ science-related positive dispositions. One of the proximal factors that might help develop adolescent students’ positive motivational and learning-related beliefs is the instructional practices of high school teachers (Anderman, Andrzejewski, & Allen, 2011). The intentional creation of supportive instructional contexts in high school classrooms may promote and sustain adolescent students’ positive motivational and learning-related beliefs (Anderman et al., 2011). Therefore, prior to investigating the mediational roles of science-related dispositions in the association between science instructional practices and science achievement, it is crucial to investigate the direct effects of science instructional practices, teacher-directed and inquiry-based instruction, on students’ science-related dispositions. Such an examination may help determine the relative importance of teacher-directed and inquiry-based science instruction in predicting adolescent students’ science-related dispositions. Further, findings of such an investigation may help science teachers to design and implement empirically proven instructional interventions that are capable of developing and nurturing adolescent students’ positive dispositions toward science. The following two research questions addressed the purpose of this investigation:

1. Does inquiry-based science instruction predict 15-year-old students’ dispositions toward science across 66 countries?

2. Does teacher-directed instruction predict 15-year-old students’ dispositions toward science across 66 countries?

Method

Data

This study used the Organization for Economic Cooperation and Development’s (OECD) PISA 2015 student and school questionnaire data and student science assessment data (http://www.oecd.org/pisa/data/2015database). PISA 2015 focused on science literacy. More than half a million 15-year-old students from over 18,000 schools in 72 countries and economies took part in PISA 2015 surveys and assessments (Organization for Economic Cooperation and Development [OECD], 2017). The country-level data were drawn from the World Bank database (https://data.worldbank.org/indicator) and the Central Intelligence Agency (CIA) world factbook (https://www.cia.gov/library/publications/the-world-factbook). The sample of the present study included 428,197 15-year-old students (male = 210,867 (49%), female = 217,330 (51%); M_age = 15.78 years, SD =.29) from 15,644 schools in 70 countries and economies (66 countries in total). The list of countries and economies included in the study is given in Table 1.

Table 1. List of countries/economies included in the study.

Algeria	Lithuania
Argentina	Luxembourg
Australia	Macao
Austria	Malaysia
Belgium	Malta
Brazil	Mexico
Bulgaria	Moldova
Canada	Montenegro
Chile	Netherlands
Chinese Taipei	New Zealand
Colombia	Norway
Costa Rica	Peru
Croatia	Poland
Czech Republic	Portugal
Denmark	Qatar
Dominican Republic	Romania
Estonia	Russian Federation
Finland	Singapore
France	Slovak Republic
Georgia	Slovenia
Germany	Vietnam
Greece	Spain
Hong Kong	Switzerland
Hungary	Thailand
Iceland	Trinidad and Tobago
Indonesia	United Arab Emirates
Ireland	Tunisia
Italy	Turkey
Japan	FYROM (Former Yugoslav Republic of Macedonia)
Kazakhstan	United Kingdom
Jordan	United States
Korea	Uruguay
Kosovo	Beijing-Shanghai-Jiangsu-Guangdong (China)
Lebanon Latvia	Spain (Regions) Ciudad Autónoma de Buenos (Argentina)

Measures

The PISA 2015 original measures used in the current study (see Table 2) were constructed employing the item response theory (IRT) scaling methodology (see OECD, 2017). “IRT models show the relationship between the ability or trait (symbolized θ) measured by the instrument and an item response” (DeMars, 2010, p. 3). PISA 2015 used one of the most widely recognized IRT scaling models based on Masters’ (1982) partial credit model (PCM), the generalized partial credit model (GPCM; Muraki, 1992), to develop IRT-based scales. The GPCM is a mathematical model for constructing measures using items with two or more ordered response categories (Muraki, 1997). The scaling procedures and the construct validation procedures of the PISA 2015 context questionnaire derived variables and measures are described in detail in the PISA 2015 technical report (http://www.oecd.org/pisa/data/2015-technical-report).

Table 2. PISA 2015 scales and scale items.

Teacher-directed instruction The teacher explains scientific ideas. A whole class discussion takes place with the teacher. The teacher discusses our questions. The teacher demonstrates an idea. Inquiry-based instruction Students are given opportunities to explain their ideas. Students spend time in the laboratory doing practical experiments. Students are required to argue about science questions. Students are asked to draw conclusions from an experiment they have conducted. The teacher explains how a school science idea can be applied to a number of different phenomena (e.g., the movement of objects, substances with similar properties). Students are allowed to design their own experiments. There is a class debate about investigations. The teacher clearly explains the relevance of broad science concepts to our lives. Students are asked to do an investigation to test ideas. Enjoyment of science I generally have fun when I am learning broad science topics. I like reading about broad science I am happy working on broad science topics. I enjoy acquiring new knowledge in broad science I am interested in learning about broad science. Interest in broad science topics Biosphere (e.g., ecosystem services, sustainability) Motion and forces (e.g., velocity, friction, magnetic and gravitational forces) Energy and its transformation (e.g., conservation, chemical reactions) The Universe and its history How science can help us prevent disease Instrumental motivation to learn science Making an effort in my school science subject(s) is worth it because this will help me in the work I want to do later on. What I learn in my school science subject(s) is important for me because I need this for what I want to do later on. Studying my school science subject(s) is worthwhile for me because what I learn will improve my career prospects. Many things I learn in my school science subject(s) will help me to get a job. Science self-efficacy Recognize the science question that underlies a newspaper report on a health issue. Explain why earthquakes occur more frequently in some areas than in others. Describe the role of antibiotics in the treatment of disease. Identify the science question associated with the disposal of garbage. Predict how changes to an environment will affect the survival of certain species. Interpret the scientific information provided on the labeling of food items. Discuss how new evidence can lead you to change your understanding about the possibility of life on Mars. Identify the better of two explanations for the formation of acid rain. Epistemological beliefs about science A good way to know if something is true is to do an experiment. Ideas in broad science sometimes change. Good answers are based on evidence from many different experiments. It is good to try experiments more than once to make sure of your findings. Sometimes broad science scientists change their minds about what is true in science. The ideas in science books sometimes change.

Inquiry-based science instruction

The inquiry-based science instruction scale (IBTEACH; Cronbach’s α = 0.74 to 0.92 across national samples) comprised nine items (e.g., Students are asked to draw conclusions from an experiment they have conducted), rated on a 4-point Likert-type scale, ranging from 1 (never or hardly ever) to 4 (in most lessons).

Teacher-directed science instruction

The teacher-directed science instruction scale (TDTEACH; Cronbach’s α = 0.69 to 0.89 across national samples) included four items (e.g., A whole class discussion takes place with the teacher), rated on a 4-point Likert-type scale, ranging from 1 (never or almost ever) to 4 (every lesson or almost every lesson).

Interest in broad science topics

The interest in broad science topics scale (INTBRSCI; Cronbach’s α = 0.72 to 0.89 across national samples) was based on five items (e.g., How science can help us prevent disease), rated on a 4-point Likert-type scale, ranging from 1 (not interested) to 4 (highly interested).

Instrumental motivation to learn science

The instrumental motivation to learn science scale (INSTSCIE; Cronbach’s α = 0.76 to 0.96 across national samples) included four items (e.g., “Making an effort in my school science subject(s) is worth it because this will help me in the work I want to do later on”), rated on a 4-point Likert-type scale, ranging from 1 (strongly disagree) to 4 (strongly agree).

Science self-efficacy

The science self-efficacy scale (SCIEEFF; Cronbach’s α = 0.73 to 0.94 across national samples) comprised eight items (e.g., Discuss how new evidence can lead you to change your understanding about the possibility of life on Mars), rated on a 4-point Likert-type scale, ranging from 1 (I couldn’t do this) to 4 (I could do this easily).

Epistemological beliefs about science

The epistemological beliefs about science scale (EPIST; Cronbach’s α = 0.69 to 0.94 across national samples) included six items (e.g., It is good to try experiments more than once to make sure of your findings), rated on a 4-point Likert-type scale, ranging from 1 (strongly disagree) to 4 (strongly agree).

Student-, school-, and country-level control variables

The student-, school-, and country-level demographic and socioeconomic control variables included in the study were gender (1 = female, 0 = male), immigration status (1 = non-immigrant, 0 = immigrant), index of economic, social, and cultural status (ESCS; a composite score of highest level of education of parents, highest parental occupational status, and home possessions; see OECD, 2017), school ownership type (1 = public, 0 = private), school location (1 = rural, 0 = urban), index of schools’ science-specific resources (see OECD, 2017), income Gini coefficient (CIA, 2017; World Bank, 2017), and log GDP per capita (CIA, 2017; World Bank, 2017).

Analytic strategy

We employed multilevel path analysis, an extension to path analysis for analyzing hierarchically structured data (Cheung & Au, 2005), as the analytic strategy to test the hypothesized model and to answer the two research questions. Multilevel path analysis effectively combines the best of both methodologies, path analysis and multilevel modeling (Kaplan, 2009). For nested data, the use of path analysis may lead to serious specification errors (Kaplan, 2009). The use of path analysis alone may result in biased parameter estimates and may mask complex indirect and simultaneous effects within and across levels of nestedness (Kaplan, 2009; Preacher, Zhang, & Zyphur, 2016).

The hypothesized multilevel path model is shown in Figure 1. The statistical software, Mplus Version 8.2 (Muthén & Muthén, 1998-2018), was used to perform multilevel path analyses. All student-, school-, and country-level continuous measures were grand-mean centered (see Brincks et al., 2017), whereas all binary variables were retained in their original metric. Grand-mean centering not only facilitates the interpretation of a multilevel model (Heck & Thomas, 2015) but also helps to mitigate multicollinearity (Tabachnik & Fidell, 2013). We used the maximum likelihood with robust standard errors (MLR) estimation procedure to estimate the multilevel path model. The MLR estimator is capable of effectively handling nonnormally distributed data and complex or clustered data (see Lei & Wu, 2012). The PISA 2015 student- and school-level sampling weights were included in the multilevel path model to produce unbiased estimates of standard errors (see Asparouhov, 2006; OECD, 2017). The full information maximum likelihood (FIML) method was employed to handle the missing data (see Enders, 2001; Enders & Bandalos, 2001). We used the following goodness-of-fit indices to assess the fit of the hypothesized multilevel path model (see Schreiber, Nora, Stage, Barlow, & King, 2006): root mean square error of approximation (RMSEA < .08), standardized root mean square residual (SRMR ≤ .08), comparative fit index (CFI ≥ .95), and Tucker–Lewis index (TLI ≥ .95). Given the large sample size in the current study, the alpha error rate for statistical significance was set at p < .01 to minimize the probability of Type I error (see Cohen, Manion, & Morrison, 2018).

Figure 1. The hypothesized multilevel path model.

Results

The descriptive statistics for all variables and measures included in the multilevel path model are given in Table 3. The correlations among the predictor and outcome measures are reported in Table 4. The hypothesized multilevel path model fitted the observed data well. All of the goodness-of-fit indices were within the recommended ranges: CFI = 0.997, TLI = 0.959, RMSEA = 0.006, SRMR _within = 0.016, SRMR _{between level 2} = 0.000, SRMR _{between level 3} = 0.001.

Table 3. Means, standard deviations, skewness, and kurtosis for variables and measures.

Variables/measures	M	SD	Skewness	Kurtosis
Student-level
Gender	0.50	0.50	−0.016	−2.000
Immigration status	0.88	0.32	−2.362	3.579
Economic, social, and cultural status (ESCS)	−0.31	1.13	−0.406	−0.198
Inquiry-based science instruction	0.11	1.01	0.010	2.310
Teacher-directed science instruction	0.07	1.01	−0.073	0.262
Enjoyment of science	0.23	1.07	−0.156	−0.120
Interest in broad science topics	0.15	0.95	−0.571	1.525
Instrumental motivation	0.35	0.95	−0.310	−0.289
Science self-efficacy	0.09	1.23	0.107	1.783
Epistemological beliefs	−0.02	0.97	0.097	0.740
School-level
School ownership type	0.78	0.42	−1.326	−0.242
School location	0.70	0.46	−0.879	−1.227
Science-specific resources	4.73	2.25	−0.359	−0.845
Country-level
Income Gini coefficient	0.36	0.08	0.616	−0.302
Log GDP per capita	10.05	0.76	−0.029	−0.867

All PISA 2015 student-level measures have a mean of 0 and a standard deviation of 1 across OECD countries. Positive values on the PISA 2015 scales indicate that mean scores are higher than the OECD average, while negative values on the PISA 2015 scales indicate that mean scores are lower than the OECD average (see OECD, 2017).

Table 4. Intercorrelations among the predictor and outcome measures.

	1	2	3	4	5	6
1. Inquiry-based science instruction	—
2. Teacher-directed science instruction	0.28
3. Enjoyment of science	0.17	0.24
4. Interest in broad science topics	0.15	0.19	0.53
5. Instrumental motivation	0.18	0.13	0.35	0.25
6. Science self-efficacy	0.20	0.13	0.31	0.28	0.27
7. Epistemological beliefs	0.05	0.18	0.34	0.26	0.15	0.17	—

All correlations are significant at p < 0.01.

The intraclass correlation coefficients (ICCs; see Table 5), i.e., the proportions of the total variance in the outcome measures that can be attributed to between-school differences and between-country differences, revealed that the outcome measures such as enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science varied between schools and between countries. The ICCs indicated that 3%, 3%, 2%, 2%, and 3% of the variances in enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science, respectively, were at the school-level. Furthermore, the ICCs suggested that 7%, 7%, 6%, 3%, and 2% of the variances in enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science, respectively, were at the country-level. The estimated student-, school-, and country-level variance components and the final R² estimates are given in Tables 6 and 7, respectively.

Table 5. Intraclass correlation coefficients (ICC) for the outcome measures.

	School-level	Country-level
Enjoyment of science	0.031	0.072
Interest in broad science topics	0.027	0.073
Instrumental motivation	0.023	0.060
Science self-efficacy	0.023	0.025
Epistemological beliefs	0.034	0.022

Table 6. Decomposition of variance into student-, school-, and country-level components.

Parameter	Estimate	SE	p
Student-level
Enjoyment of science	0.979	0.03	0.000
Interest in broad science topics	0.802	0.02	0.000
Instrumental motivation	0.821	0.03	0.000
Science self-efficacy	1.383	0.03	0.000
Epistemological beliefs	0.848	0.02	0.000
School-level
Enjoyment of science	0.037	0.00	0.000
Interest in broad science topics	0.025	0.00	0.000
Instrumental motivation	0.021	0.00	0.000
Science self-efficacy	0.035	0.00	0.000
Epistemological beliefs	0.030	0.00	0.000
Country-level
Enjoyment of science	0.060	0.01	0.000
Interest in broad science topics	0.048	0.01	0.001
Instrumental motivation	0.036	0.01	0.000
Science self-efficacy	0.030	0.01	0.000
Epistemological beliefs	0.020	0.00	0.000

Table 7. Proportion of variance in the outcome measures explained by the predictor measures (final R² estimates).

Parameter	Estimate	SE	p
Student-level
Enjoyment of science	0.08	0.01	0.000
Interest in broad science topics	0.07	0.00	0.000
Instrumental motivation	0.04	0.00	0.000
Science self-efficacy	0.08	0.00	0.000
Epistemological beliefs	0.05	0.00	0.000
School-level
Enjoyment of science	0.01	0.01	0.122
Interest in broad science topics	0.01	0.02	0.143
Instrumental motivation	0.03	0.01	0.007
Science self-efficacy	0.00	0.00	0.176
Epistemological beliefs	0.06	0.01	0.000
Country-level
Enjoyment of science	0.30	0.08	0.000
Interest in broad science topics	0.31	0.08	0.000
Instrumental motivation	0.36	0.08	0.000
Science self-efficacy	0.23	0.08	0.006
Epistemological beliefs	0.03	0.05	0.524

Inquiry-based science instruction and students’ science dispositions

Results of multilevel path analyses indicated that inquiry-based science instruction was statistically significantly and positively associated with students’ enjoyment of science, interest in broad science topics, instrumental motivation to learn science, and science self-efficacy (see Figure 2). On average, a one-point increase in inquiry-based science instruction was associated with a 0.11-point increase in enjoyment of science (B = 0.11, SE = 0.01, p < .001); a 0.09-point increase in interest in broad science topics (B = 0.09, SE = 0.00, p < .001); a 0.14-point increase in instrumental motivation to learn science (B = 0.14, SE = 0.01, p < .001); and a 0.22-point increase in science self-efficacy (B = 0.22, SE = 0.01, p < .001), after accounting for student-, school-, and country-level demographic and socio-economic factors.

Figure 2. Results of multilevel path analyses. The unstandardized regression coefficients and the standardized covariance estimates (student-level) are reported. Standard errors are in parentheses. All student-, school-, and country-level demographic and socio-economic factors were included in the analyses; but for clarity of presentation purposes, only the main measures of interest are shown.

Put another way, students whose science teachers frequently used inquiry-based methods of instruction tended to report higher levels of enjoyment of science, interest in broad science topics, instrumental motivation, and science self-efficacy than did their peers whose science teachers infrequently employed inquiry-based methods of instruction. However, inquiry-based science instruction was not statistically significantly related to students’ epistemological beliefs about science (B = 0.00, SE = 0.01, p = .855), suggesting that there is virtually no relationship between inquiry-based science instruction and students’ epistemological beliefs about science. Nonetheless, the standardized regression coefficients indicated that inquiry-based science instruction was most strongly associated with science self-efficacy (β = 0.18, SE = 0.01, p < .001), followed by instrumental motivation to learn science (β = 0.15, SE = 0.01, p < .001), enjoyment of science (β = 0.11, SE = 0.01, p < .001), and interest in broad science topics (β = 0.10, SE = 0.00, p < .001).

Teacher-directed science instruction and students’ science dispositions

Results of multilevel path analyses also suggested that teacher-directed science instruction was statistically significantly and positively associated with students’ enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science (see Figure 2). On average, a one-point increase in teacher-directed science instruction was associated with a 0.20-point increase in enjoyment of science (B = 0.20, SE = 0.01, p < .001); a 0.14-point increase in interest in broad science topics (B = 0.14, SE = 0.00, p < .001); a 0.08-point increase in instrumental motivation to learn science (B = 0.08, SE = 0.01, p < .001); a 0.07-point increase in science self-efficacy (B = 0.07, SE = 0.01, p < .001); and a 0.16-point increase in epistemological beliefs about science (B = 0.16, SE = 0.01, p < .001). In other words, students whose science teachers frequently used direct instruction tended to report higher levels of enjoyment of science, interest in broad science topics, instrumental motivation, science self-efficacy, and epistemological beliefs about science than did their counterparts whose science teachers infrequently employed direct instruction.

The standardized regression coefficients indicated that teacher-directed science instruction was most strongly related to enjoyment of science (β = 0.20, SE = 0.01, p < .001), followed by epistemological beliefs about science (β = 0.17, SE = 0.01, p < .001), interest in broad science topics (β = 0.15, SE = 0.00, p < .001), instrumental motivation to learn science (β = 0.08, SE = 0.01, p < .001), and science self-efficacy (β = 0.06, SE = 0.01, p < .001). Furthermore, an examination of the standardized regression coefficients suggested that the strength of associations of teacher-directed science instruction with enjoyment of science and interest in broad science topics was stronger than the strength of associations of inquiry-based science instruction with these science dispositions (β = 0.20, 0.11; β = 0.15, 0.10, respectively). However, the strength of associations of inquiry-based science instruction with instrumental motivation to learn science and science self-efficacy was stronger than the strength of associations of teacher-directed science instruction with these science dispositions (β = 0.15, 0.08; β = 0.18, 0.06, respectively).

Discussion

The current study, employing self-determination theory as the theoretical framework, aimed at examining the relations of both teacher-directed and inquiry-based science instruction to students’ science-related dispositions, such as enjoyment of science, interest in broad science topics, instrumental motivation to learn science, science self-efficacy, and epistemological beliefs about science, across 66 PISA 2015 participating countries. Consistent with the findings of prior research (e.g., Cairns & Areepattamannil, 2019; Mac Iver et al., 2001; Madden, 2011; McConney et al., 2014; Palmer, 2009; Patrick et al., 2009), results of the study revealed that inquiry-based science instruction was significantly and positively associated with students’ enjoyment of science, interest in broad science topics, instrumental motivation to learn science, and science self-efficacy. One of the key principles underpinning the concept of inquiry-based science instruction is student autonomy in learning (Constantinou et al., 2018). As such, teachers employing inquiry-based science teaching and learning practices are more likely to adopt an autonomy-supportive motivating style toward their students during instruction (see Rogat et al., 2014). Although the myriad benefits of teachers’ autonomy support during instruction are well documented in the extant basic psychological need satisfaction literature (see Chirkov & Ryan, 2001; Gillet, Vallerand, & Lafrenière, 2012; Niemiec & Ryan, 2009; Reeve et al., 2004; Soenens & Vansteenkiste, 2005; van der Kaap-deeder, Vansteenkiste, Soenens, & Mabbe, 2017), a small but growing body of research has also demonstrated the beneficial outcomes of teachers’ autonomy support during science instruction. For example, Jungert and Koestner (2015) found that teachers’ use of autonomy-supportive practices during science instruction was significantly positively related to high school students’ autonomous motivation, self-efficacy, and science achievement.

Similarly, results suggested that teacher-directed science instruction was also significantly and positively related to students’ enjoyment of science, interest in broad science topics, instrumental motivation to learn science, and science self-efficacy. These findings highlight the importance of teacher-provided structure, i.e., communicating clear expectations to students and framing students’ learning activities with explicit directions and guidance (Jang et al., 2010, p. 588), in promoting students’ positive dispositions toward science. There is mounting evidence that teacher-provided autonomy support may not alone engage students in learning activities, and may need to be supplemented with teacher-provided structure (e.g., Hospel & Galand, 2016; Jang et al., 2010; Mouratidis, Michou, Aelterman, Haerens, & Vansteenkiste, 2018; Sierens, Vansteenkiste, Goossens, Soenens, & Dochy, 2009; Vansteenkiste et al., 2012).

For instance, Jang et al. (2010) examined the relations of these two important aspects of teachers’ instructional styles—provision of autonomy support and provision of structure—to high school students’ collective behavioral engagement and individual self-report engagement. The findings of correlational analyses showed that teacher-provided autonomy support and structure were strongly and positively associated with students’ classroom engagement. Further, results of hierarchical linear modeling (HLM) analyses revealed that teacher-provided autonomy support uniquely predicted students’ collective behavioral engagement as well as individual self-report engagement, while teacher-provided structure uniquely predicted students’ collective behavioral engagement.

Results of the current study, contrary to popular belief, indicated that teacher-directed science instruction was more strongly associated with students’ enjoyment of science and interest in broad science topics than inquiry-based science instruction, whereas inquiry-based science instruction was more strongly related to students’ instrumental motivation to learn science and science self-efficacy than teacher-directed science instruction. The finding that teacher-directed instruction has a stronger association than inquiry-based instruction with science enjoyment, is interesting in itself. Possibly, the appeal of the teacher-directed instruction within the tighter framework of students’ understanding of what they should understand, a less ambiguous structure wherein scope, topics, and learning objectives are clearly defined for students may lead to an enhanced sense of comfort and subsequently, enjoyment. This finding is perhaps unexpected, given the almost unchallenged views espoused in educational rhetoric nowadays, with Piagetian notions of constructing one’s knowledge conflated with views that inquiry-based approaches to learning are desirable. The heightened association for interest in broad science topics could also be, at least in part, due to the structured nature of learning where the teacher is guiding the student directly toward exemplar broad science topics, and where students have less autonomy to choose details of topics which interest them. Inquiry-based instruction, by contrast, involving degrees of self-direction (as it should if implemented authentically) may be providing a base upon which students’ own self-belief, confidence in their own ability and comfort in reliance upon oneself, flourish.

In the earlier literature review, we have shown the diversity of previous findings in the academic literature and the absence of a clear understanding of the relative effectiveness of the two instructional practices (teacher-directed versus inquiry-based) in terms of student dispositions. This underlines why studies such as the present one, which compares the relative effectiveness of the two, are important. Whereas well-designed, controlled experimental and quasi-experimental studies have provided empirical support for the effectiveness of direct, guided science instruction in improving students’ science learning and developing their positive science dispositions (e.g., Klahr & Nigam, 2004; Raes & Schellens, 2016; Zepeda, Richey, Ronevich, & Nokes-Malach, 2015), meta-analyses of experimental and quasi-experimental studies of inquiry-based science instruction and longitudinal studies of inquiry-based science programs have demonstrated the effectiveness of inquiry-based science teaching in improving students’ science learning and dispositions (e.g., Furtak et al., 2012; Gibson & Chase, 2002). These findings, that both teacher-directed and inquiry-based science instructional forms are associated with heightened students’ dispositions, are reflected in the present study too.

Finally, the results of the present study also suggested that teacher-directed science instruction alone was significantly positively linked to students’ epistemological beliefs about science. Without doubt, classroom teachers play a significant role in developing and promoting an understanding of the nature of science among their students (see Brickhouse, 1990; Leach, Hind, & Ryder, 2003; Maggioni & Parkinson, 2008). Given the positive association between teacher-directed science instruction and students’ epistemological beliefs about science, science teachers may be required not only to understand the epistemological underpinnings of science but also to build their students’ capacities to develop or increase epistemic insight (Billingsley & Fraser, 2018; Billingsley, Taber, Riga, & Newdick, 2013). Billingsley and Fraser (2018) rightly posit,

Adopting epistemic insight as a curriculum goal can potentially engage students’ intellectual curiosity, develop their interdisciplinary scholarly expertise and ability to find solutions to wicked problems which are rational and compassionate. Potentially, a curriculum which engages with epistemic insight may also widen the pipeline from school to science and science-related careers. (p. 3)

Implications for policy and practice

Unlike the findings of the previously reported studies on the relative effectiveness of teacher-directed and inquiry-based instruction in enhancing student learning and achievement (e.g., Alfieri, Brooks, Aldrich, & Tenenbaum, 2011; Klahr & Nigam, 2004), the findings of the current study underline the important roles of both teacher-directed and inquiry-based science instruction in predicting students’ science-related dispositions. These findings suggest that a blended instruction model—a combination of teacher-directed and inquiry-based science instruction—may be more beneficial for nurturing and developing students’ positive dispositions toward science. As Holliday (2006) rightly opines,

Like other teaching and learning strategies, what works best is a mixture of back-and-forth, non-linear combinations of implicit and explicit teaching depending on teachers’ professional judgments (p. 204)

Given the dwindling numbers of students opting to pursue STEM-related programs and careers in several countries across the globe (see Jones et al., 2018; Marginson et al., 2013; Murphy et al., 2018; Smith & White, 2018), science teachers may need to help students develop a vast array of positive dispositions toward science employing a blended instruction model. Prior research has also demonstrated that such a blended instruction model may lead to better student outcomes. For example, Denoël et al. (2017) examined the relations of teacher-directed and inquiry-based science instruction with science achievement among 15-year-old students in PISA 2015 participating countries. The authors documented that students who received a blend of teacher-directed and inquiry-based science instruction—teacher-directed instruction in most or almost all classes with inquiry-based instruction in some of them—performed better on the PISA 2015 science assessment than did their peers who received inquiry-based instruction in most or almost all classes without a strong foundation of teacher-directed instruction.

Because even experienced teachers may find it difficult to implement inquiry-based instruction in the classroom (see Dobber, Zwart, Tanis, & van Oers, 2017), students may not be able to employ inquiry-based instructional strategies without a thorough understanding of basic science concepts and their applications gained through teacher-directed instruction (Denoël et al., 2017; Holliday, 2006). Hence, in-service science teachers may need to engage in professional learning and development to successfully implement a blended instruction model in their classrooms. A well-designed professional development program aimed at fostering inquiry-based science instruction (see Tsivitanidou, Gray, Rybska, Louca, & Constantinou, 2018) is a mandatory prerequisite for in-service teachers to progress to a blended instruction model.

However, professional development programs targeted at addressing the needs of in-service science teachers may not alone suffice. Pre-service science teachers should also be given ample opportunities during their pre-service training (i.e., during both coursework and practicum experiences) to acquire the knowledge and skills to plan, implement, and evaluate effective blended instruction. Furthermore, there is a need to develop classroom observation and interview protocols that are congruent with the blended instruction model to accurately gauge actual classroom practice. Although there are a variety of classroom observation and interview protocols available to assess inquiry-based science instruction (see Minner et al., 2010), there is a dearth of classroom observation and interview protocols that focus solely on a blend of teacher-directed and inquiry-based instruction. Nevertheless, the success of a blended instruction model may largely depend on several instructional contextual factors, including, but not limited to the type of learner, the classroom culture, and the instructional content (Chase & Klahr, 2017).

Limitations of the study and directions for future research

There are three major limitations to the study. First, the PISA 2015 measures used in the study were self-report measures. Although self-report instruments have been widely used by psychologists and educators for decades, the use of such instruments in social and behavioral science research has been criticized on grounds of response bias factors, such as acquiescence and social desirability. Future research assessing instructional practices, such as teacher-directed and inquiry-based science instruction, may need to use classroom observation and interview protocols to provide an accurate picture of actual classroom practice. Second, the study was cross-sectional in nature. Hence, causality cannot be assumed. Further research, employing quasi-experimental and experimental research designs, is warranted to investigate the casual relations among the variables of interest in the current study. Additional research, using these research methods, is also needed to determine the mediating effects of students’ science-related dispositions on the relationship between teacher-directed and inquiry-based science instruction and science achievement. Finally, the countries included in the study vary hugely in terms of their social, economic, cultural, political, and religious characteristics. All potential country-level confounding factors were not accounted for in the current study. Given the diverse characteristics of PISA 2015 participating countries, future research may need to test the hypothesized model in each PISA 2015 participating country to unmask differences in structural relations among the variables, if any, between countries.

Conclusion

Despite these limitations, the findings of the present study provide empirical support for the positive relationships of both teacher-directed and inquiry-based science instructional practices with students’ science-related dispositions, such as enjoyment of science, interest in broad science topics, instrumental motivation to learn science, and science self-efficacy, among nationally representative samples of adolescent students drawn from 66 PISA 2015 participating countries across the world. Results of the study also indicated that teacher-directed science instruction was solely significantly and positively associated with students’ epistemological beliefs about science. A blend of teacher-directed and inquiry-based science instructional practices may be more appropriate for developing students’ positive dispositions toward science. This may allow for an effect benefiting from both models, i.e., providing sufficient structure, guidance, and conceptual foundation on one hand, which teacher-directed instructional practices appear to optimize, with opportunities for autonomy and self-direction which inquiry-based instructional approaches appear to aid, possibly leading to higher levels of self-efficacy. However, both pre-service and in-service science teachers may need to attend well-designed professional development programs for developing and honing their repertoire of teaching skills and strategies based on a combination of teacher-directed and inquiry-based pedagogical approaches. One of the findings with a reasonably straightforward conclusion is that of the positive relationship of both teacher-directed and inquiry-based instructional practices with science enjoyment (though with a stronger association for teacher-directed instruction). This requires the concept of enjoyment of science lessons to become less of a “luxury” and more of a “necessity” when planning science lessons, and requires appropriate emphasis of time and resources by both science teachers and school administrators. This would enable the rounded, balanced approach toward science teaching which our results indicate would impact positively on science-related dispositions. The strength of association of inquiry-based science instruction with science self-efficacy was stronger than those of teacher-directed instructional practices, suggesting that opportunities for autonomy which aspects of inquiry-based instruction (taught faithfully), such as self-direction of learning, and activity versus passivity, are linked positively with students’ growth of confidence and belief in their own abilities. Given the increasing demands upon students in STEM domains and later on as young adults within the STEM workforce, such as the application of 21st century skills, a need to demonstrate abilities to think critically and “outside of the box”, to be able to apply familiar theory to unfamiliar situations, this study illustrates the desirability of a teaching and learning model which blends elements of both teacher-directed and inquiry-based science instructional practices.

Disclosure statement

No potential conflict of interest was reported by the authors.