The role of interest in climate change instruction
Abstract
As climate change becomes an increasingly important topic for science educators, it is critical to learn how teachers may be able to increase students' knowledge about it. We conducted two consecutive quasi-experimental studies that investigated the role of interest in predicting middle school students' knowledge gains from a unit about how scientists use mathematical models to predict climate change's impacts on forests. The studies measured the intervention's effects on students' knowledge about climate change and examined how their interest in the topic and related factors were associated with their knowledge before and after the intervention. Participants in the two studies included 467 treatment and 177 comparison group students (Study 1) and 363 treatment and 219 comparison group students (Study 2). Multilevel modeling analyses revealed increases in students' knowledge about climate change after participating in the unit. Path modeling analyses showed that students' interest in climate change was indirectly related to their knowledge about climate change, mediated by students' existing desire to learn more, their interest in the unit, their belief that climate change was important (Studies 1 and 2), as well as their behavioral self-efficacy (Study 2). Students' interest in science was positively associated with their knowledge about climate change (Study 1) but their perception of threats posed by climate change was not (Study 2). Findings suggest that science educators can improve students' knowledge about climate change by connecting the topic to students' lives and ensuring they feel empowered to act on climate change.
1 INTRODUCTION
Climate change is one of the defining social and scientific challenges of the 21st century (IPCC, 2014, 2018), and is, therefore, a critically important educational topic (Anderson, 2012). Yet relatively little is known about how to engage students with climate change in ways that increase their foundational knowledge about this topic (Shapiro Ledley et al., 2017). Foundational knowledge refers to “what” students need to know (Kereluik et al., 2013, p. 130), including core content knowledge, digital literacy, and cross-disciplinary thinking skills (Kereluik et al., 2013). Foundational knowledge is an important component of 21st century knowledge, i.e., knowledge that students will need to address future social and scientific challenges (Kereluik et al., 2013), including climate change. Foundational knowledge about climate change is especially important because youth need to understand climate change to make effective decisions about response options (NOAA, 2009; Shapiro Ledley et al., 2017). Yet multiple studies have found that many youth have scientific misperceptions about climate change (Leiserowitz et al., 2011; Shepardson et al., 2014), such as being unable to name carbon dioxide as the primary greenhouse gas, to differentiate climate from weather (Leiserowitz et al., 2011), or to describe climate change effects on vegetation (Shepardson et al., 2014).
Moreover, climate change is becoming an important topic in science classrooms. In the United States, the Next Generation Science Standards (NGSS) explicitly include climate change concepts in its Disciplinary Core Ideas and in performance expectations for middle school and high school (NGSS Lead States, 2013). As these standards are adopted more widely, there is a need to better understand to how educational interventions can support increases in students' foundational knowledge about climate change.
The NGSS encourage promoting student interest as a way to increase their science engagement and learning in the classroom (NGSS Lead States, 2013). Interest broadly refers to a student's autonomous preference for a topic, object, or activity (Freeman et al., 2002; Schiefele et al., 1992) and includes both cognitive and affective components such as intrinsic inclination toward a topic (Schiefele et al., 1993), broader interest in related topics, perceiving that the topic is important (Renninger & Hidi, 2016; Schiefele, 1991), and positive affect regarding the topic (Ainley & Hidi, 2014; Renninger & Hidi, 2016; Schiefele, 1991).
We explore which components of interest were most relevant in students' gaining foundational knowledge about climate change within the context of a climate change education intervention. The research reported here seeks to advance scholarship on the role of interest in students' knowledge about climate change by presenting results from two studies. The two quasi-experimental studies were conducted during two consecutive school years, within the context of a 2-week educational intervention implemented in a public school district's seventh-grade classes (ages 12–13). The first study explored students' interest in climate change and its relationship with their knowledge about climate change before and after the intervention. The second study further investigated this link before and after a revised version of the unit, with a greater emphasis on the factors mediating the relationship.
1.1 Past empirical research
There is a large body of educational research that explores how students develop an interest in science and specific science topics as well as the effects of interest on learning and other long-term outcomes (e.g., Archer et al., 2013; OECD, 2007; Palmer et al., 2017; Sjøberg & Schreiner, 2010). However, only four studies to date have measured students' interest in climate change. Two of these studies measured changes in students' interest in climate change as a result of an educational intervention. These studies sought to explore if educational interventions can change students' interest in climate change, and to identify the practices and factors associated with changes in their interest in the topic. Nussbaum et al. (2016) found that ninth-grade students' (ages 14–15) interest in local climate change impacts increased after participating in a computer-based game designed to teach them about the effects of climate change on major local water resources. The authors attributed students' increased interest in the game's ability to make climate change impacts personally relevant. This is in contrast to the other study, conducted by the authors (Carman et al., 2017), which found that seventh-grade students' (ages 12–13) interest in climate change, as well as their desire to learn more about the topic, did not change after using computer-based activities to scientifically model the impacts of climate change on regional trees and forests. That study, however, showed that students' interest in science and hands-on learning activities played a greater role in their interest in climate change than their perception of the threats posed by climate change. Moreover, that particular study indicated that students' interest in climate change is distinct from their desire to learn more about the topic (Carman et al., 2017) and thus, that these two variables should not be treated as a single construct (cf. Dijkstra & Goedhart, 2012; Nussbaum et al., 2016). The two remaining cross-sectional studies assessed the relationships between students' interest in climate change, their beliefs, and other outcomes. One of these studies explored college students' beliefs and trust in information about climate change (Bråten et al., 2009). When controlling for age, gender, and knowledge about the topic students' interest only helped to explain their beliefs about information sources related to climate change (Bråten et al., 2009), but not their beliefs in the certainty of climate change evidence. The other study found that secondary school students' (age 18) interest in climate change initially predicted enhanced performance on a reading comprehension task on the topic, but it no longer had a significant effect after prior knowledge about climate change, ease of comprehension of the text itself, and personal memorization abilities were considered (Strømsø et al., 2010).
It is important to note that these studies measured interest in different ways. In the study by Nussbaum et al. (2016), students were asked to rate their level of interest (on a four-point scale from no interest to high interest) on several topics relating to the unit, including knowledge-related items such as “Understanding better how human activity impacts climate change” (Nussbaum et al., 2016, p. 799) and action-related items such as “Learning more about water conservation in the home,” (Nussbaum et al., 2016, p. 799) treating interest as the same as wanting to learn more about the topic. Similar to Nussbaum et al. (2016), Bråten et al.'s (2009) and Strømsø et al.'s (2010) constructs to measure interest in climate change included items describing both students' self-reported level of interest in educational topics (e.g., “I am interested in what conditions influence the Earth's climate,” Bråten et al., 2009, p. 539), and their desire to engage in related behaviors (e.g., “I am concerned with how I myself can contribute to the reduction of environmental pollution," Bråten et al., 2009, p. 539). In contrast, the authors' previous study (Carman et al., 2017) included question items that asked students separately about their interest in, and their desire to learn more about, topics included in the unit, such as “How climate change affects forests,” and did not include question items about conservation behaviors.
Researchers have found that many adolescents have misperceptions about the causes of climate change and know little about climate systems (Bodzin et al., 2014; Corner et al., 2015; Leiserowitz et al., 2011; Shepardson et al., 2014). In response, several authors have investigated instructional practices to improve students' knowledge about climate change. Lombardi et al. (2013) found that a seventh-grade (ages 12–13) unit with activities linking a model and evidence resulted in increased knowledge about climate change. When a similar unit was enacted with high school students (grades 9–12, ages 14–18), however, knowledge outcomes varied by the teacher (Lombardi et al., 2018). Sellmann and Bogner (2013) reported that a 1-day field trip increased high school students' (ages 15–19) knowledge about climate change. Bodzin and Fu (2014) found that using geospatial technologies (e.g. Google Earth) to teach about the geographic effects of climate change resulted in improved knowledge about climate change among eighth-grade (ages 13–14) students. Otieno et al. (2014) learned that sensational (i.e., emotion-focused) presentations of climate change information, in contrast to neutral, facts-focused presentations, led to higher performance on posttests of knowledge about climate change and invasive species among college psychology students (mean age = 22.6), when controlling for their prior knowledge about that topic. Varela et al. (2018) found that seventh-grade students (age 12–13) developed more sophisticated mental models of climate change after a 5-day unit that included both discussion of climate science and simulated negotiation activities. Most recently, Kuthe et al. (2019) found that teenagers (ages 13–16) who developed customized projects to address climate change throughout a yearlong program did not increase in their knowledge about climate change if they were disengaged with the topic. Unfortunately, these authors did not report what changes occurred among engaged students.
Few researchers have explored how adolescents' attitudes toward and interest in science relate to their knowledge about climate change, and these studies have found mixed results. Otieno et al.'s (2014) study found that students' attributions of the causes of climate change (human-caused or natural) had no effect on their performance on a multiple-choice test about climate change and invasive species. Dijkstra and Goedhart (2012), however, identified a positive correlation between young adults' (ages 12–21) interest in science and their knowledge about climate change. Our study adds to this body of research by accounting for various other attitudinal variables that may mediate the relationship between interest in and knowledge about climate change.
1.2 The present studies
1.2.1 Theoretical approach
Our overall theoretical approach incorporates several types of interest, as well as factors related to interest that may mediate the relationship between interest and knowledge. Educational researchers have classified interest in two main types: situational and individual (Hidi, 2000; Renninger, 2000). Situational Interest denotes a short-term preference for specific activities as they occur. Individual Interest is associated with a personal disposition and characterized by repeated engagement with a topic or object on one's own. Situational interest is easier to prompt through classroom activities, but is generally short-lived (Bergin, 1999; Hidi & Renninger, 2006), whereas individual interest develops more slowly but is linked to more autonomous information-seeking over time (Hidi & Renninger, 2006; Renninger, 2000). These two types of interest affect each other over time. Situational interests can deepen into individual interests over time as individuals choose to immerse themselves more in a topic (Hidi & Renninger, 2006; Renninger, 2000). Moreover, existing individual interests can prime students to have greater situational interest in learning activities on new but related topics (Azevedo, 2018; Hidi & Renninger, 2006). Students' interest in a classroom topic—sometimes called topic interest—integrates elements of both situational and individual interest, as students can have a pre-existing interest in a topic, be engaged through activities related to that topic, or both (Ainley et al., 2002; Renninger & Hidi, 2016).
The role that interest might play in building foundational knowledge about climate change is not yet well understood. Research on fostering interest suggests that interest includes positive affective factors such as enjoyment and fun (Ainley & Hidi, 2014; Renninger & Hidi, 2016), but climate change often promotes negative affective responses such as worry, fear, and stress (Doherty & Clayton, 2011), including among youth (Busch & Osborne, 2014; Ojala, 2012). Climate change education research also suggests that focusing only on one's own emotions regarding climate change may have negative consequences as it may cause youth to take on problem-avoidant coping strategies (Ojala, 2012). However, science and educational research on interest suggests that interest can be helpful to adolescents' science learning (Krapp, 1999; Potvin & Hasni, 2014). Moreover, because interest contains many subcomponents other than positive affect, some of these other components, such as perceived importance, may play a role in students' learning about climate change. For example, students' interest in the environment and in environmental education activities has been linked to proenvironmental behaviors (Fröhlich et al., 2013), including climate change mitigation actions (Hermans & Korhonen, 2017), suggesting that interest-related factors may support positive engagement with climate change. Fostering situational interest in the unit was thus expected to indirectly support subsequent interest in, and learning about, climate change (Hidi & Renninger, 2006).
Other interest-related factors that can predict adolescents' knowledge about climate change have yet to be explored. One such promising factor may be the students' desire to learn more about this topic. As students' temporary situational interests develop into longer-term individual interests, students are also more likely to increase their autonomous information-seeking related to that interest due to a desire to learn more about the subject of their interest (Hidi & Renninger, 2006). Our past study (Carman et al., 2017) found a strong link between students' interest in climate change and their desire to learn more about this topic both before and after an educational intervention but did not measure whether this desire to learn more was associated with increases in knowledge about climate change. Another relevant interest-related factor is the importance or meaningfulness that students assign to a topic, which has been theoretically linked to both their interest in and knowledge about that topic (Schiefele, 1991). Empirically, PISA 2006 found a positive link between how much students value science and their science literacy (OECD, 2007).
In summary, our study incorporated three premises from interest theory: first, that existing (i.e., preintervention) individual interests can positively affect situational interest in learning activities on a related topic second, that situational interests positively predict individual interests and related factors after an intervention; and third, that the relationship between students' interest in, and their knowledge about, climate change will be mediated by other factors such as their desire to learn more and perceived topic importance.
1.2.2 Description of intervention
The intervention consisted of a 2-week middle school unit, Climate Change and Michigan Forests (http://climatechangeandforests.org). The unit focuses on how scientists use mathematical modeling to predict the impacts of climate change on regional trees and forest ecosystems, based on authentic data from one coauthor's forest ecology research. Though the 2-week unit is relatively short, it is somewhat longer than many existing climate change educational interventions (e.g., Nussbaum et al., 2016; Sellmann & Bogner, 2013; Varela et al., 2018).
Climate Change and Michigan Forests was designed consistent with backward design (Graff, 2011; Jones et al., 2009; Prideaux, 2003), the 5E learning cycle (Bybee, 2014; Karpudewan et al., 2015; Songer, 2006), and the U.S. Next Generation Science Standards (NGSS Lead States, 2013). More details about the unit's development are provided in Appendix B. Since the completion of the studies reported on here, the unit has been adopted by the local school district's seventh-grade science curriculum, vetted by National Science Teaching Association (NSTA) curators based on the EQuIP rubric, and as result of that review, added to the organization's NGSS resource page (https://ngss.nsta.org/Resource.aspx?ResourceID=451).
The unit has two key features that were expected to support students' learning about climate change based on alignment with recognized best practices (Monroe et al., 2019). First, the unit includes a computer-based activity in which students construct simplified linear models of tree growth by entering and analyzing tree growth data. Second, as part of the unit students participate in a 2-hour field trip to a forest located within walking distance of their school, to conduct their own measurements from trees. Thus, the unit was designed to provide students with the opportunity to gain firsthand experience with how scientists collect data to predict the impacts of climate change on trees and forests. There is evidence that scientific modeling (Freeman et al., 2002; Inkinen et al., 2020) and hands-on activities (Potvin & Hasni, 2014; Swarat et al., 2012), including hands-on learning with trees (Jung et al., 2019), can foster youth's situational interests in science lessons. Moreover, modeling and mathematical analysis activities have been shown to improve students' knowledge about climate change (Pruneau et al., 2010).
A number of climate change education efforts have used scientific modeling activities in attempts to increase students' understanding of climate change (e.g., Lombardi et al., 2018; Pruneau et al., 2010). To the best of the authors' knowledge, however, none have drawn on real-world tree and forest data. We chose to focus on trees and forests partly because of the positive roles they play in children's lives (Lohr & Pearson-Mims, 2005; Sobel, 1995). Moreover, we specifically chose to focus on local and regional ones because these trees and forests should be more tangible, relatable, and meaningful to students (Anderson, 2012; Busch & Osborne, 2014; Corner et al., 2015) than distant ones, such as tropical rain forests. Another reason was that students' situational interest has been positively impacted by learning activities they perceive as important or meaningful (Jack & Lin, 2014; Swarat, 2008) and connect to their existing individual interests (Tapola et al., 2013).
1.2.3 Research questions
- 1.
To what extent does a short climate change education intervention change students' knowledge about, climate change?
- 2.
What are the relationships among students' interest in climate change, related factors, and their knowledge about climate change?
2 METHODS
2.1 Study design and recruitment
Data for both studies were collected from one Midwestern U.S. city's public school district. The studies employed a quasi-experimental design (Shadish et al., 2002), including a treatment group of students who experienced Climate Change and Michigan Forests as part of their science classes, and a comparison group of students who experienced the district's regular science curriculum, which did not include formal climate change education. Study 2 was conducted one year after Study 1, at about the same time of year and with a new cohort of students.
All of the district's science teachers had the opportunity to participate in the study and chose to be in the treatment or comparison group. Teachers in the treatment group completed a 1-day professional development before implementing the unit. Teachers received a stipend in return for their participation in the study. Teachers who participated in the treatment group received a larger stipend than the teachers who participated in the comparison group because they were asked to participate in professional development to teach the unit and to keep logs of their enactment of the unit. All materials needed for teaching the lessons (except computers) were provided by the research team.
Treatment group teachers completed logs to track implementation of the unit so that the study team could obtain information about enactment. Teachers reported which lessons they taught, modifications made, and suggestions for revisions.
Treatment group students completed online pre and postintervention questionnaires just before and after participating in the unit. The pre and postintervention questionnaires were identical for treatment and comparison group students, with the exception of questions asking treatment group students about their interest in the unit and field trip. Comparison group students completed the questionnaires 2 weeks apart to match the duration of the unit, at about the same times as treatment group students. Signed parent permission forms for students to participate in the study were obtained.
2.2 Statistical analysis
All quantitative analyses were conducted using Stata v.14. The total number of multiple-choice and true-false questions students answered correctly were tabulated to generate their Topic Knowledge score. All other factors were reduced using confirmatory factor analysis. Factor scores consisted of the mean score of the respective measures, and these scores were used for all subsequent analyses.
Multilevel analyses (Gelman & Hill, 2007; Raudenbush & Bryk, 2002) were conducted to test for pre–post-intervention differences in the repeated factors in both studies while accounting for the dependence in outcomes due to repeated measures per student and students being nested within teachers. These models allowed for the exploration of within- and between-student and teacher variability. Intraclass correlation coefficients were computed and indicated that a significant proportion of variation in the six factors was due to the repeated measures being nested within students and the clustering of students within teacher. The multilevel models were fit with a random intercept for student and teacher and estimated with restricted maximum likelihood estimation. Fixed effects in the model included gender, time, group (treatment or comparison), and time–group interaction. Gender was included as a covariate because it has been associated with students' interest in science generally (Lamb et al., 2012; Osborne et al., 2003) and in the environmental and life sciences specifically (Potvin & Hasni, 2014). Time (pre/postintervention) was included in the model to assess the differences between the two time points, group to measure differences between the treatment and comparison groups, and time–group interaction to identify the differences between the treatment and comparison groups in the two time periods.
Path analyses were conducted using treatment group student data to explore the extent to which the hypothesized factors, directly and indirectly, explained students' postintervention Topic Knowledge. Manual backward selection techniques and modification indices were used to arrive at the final model. The model was fit with full information maximum likelihood estimation and standardized results were requested. To control for teacher effects on student interest development and learning (Logan & Skamp, 2013; Lombardi et al., 2018; Osborne et al., 2003), the final path model included clustered robust standard errors by teacher.
Model fit was assessed through several frequently used indicators (Kline, 2011): the χ2 statistic, comparative fit index (CFI), Tucker–Lewis Index (TLI), and root mean square error of approximation (RMSEA). The χ2 should be low and nonsignificant to attest to a good fit between the sample and theoretical model (Kline, 2011), the CFI should be above 0.95, TLI not below 0.9 (Hu & Bentler, 1999), and RMSEA less than 0.08 for a reasonably close fit (Hu & Bentler, 1999).
3 STUDY 1
3.1 Procedure
3.1.1 Hypothetical model
Multilevel analyses were conducted to answer the first research question. To answer the second question, we developed and subsequently tested a hypothetical model of relationships between students' interest(s) and knowledge about climate change and forests (Figure 1).
The model included select relationships between interest variables and knowledge about climate change and forests. A direct relationship between topic interest and topic knowledge was not included because Strømsø et al. (2010) found that this relationship is not significant when controlling for other factors, including prior topic knowledge. Instead, we measured the direct relationship between students' desire to learn more and their knowledge, because the latter can be one of the mediators between topic interest and topic knowledge (Laukenmann et al., 2003; Schraw et al., 2001). Perceived topic importance and interest in science were included as moderating variables because of their relationships to topic interest and topic knowledge (Dijkstra & Goedhart, 2012; OECD, 2007). Finally, the situational interest variables (i.e., Situational Interest: Unit, Situational Interest: Field Trip, Situational Interest in Hands-On Activities) were hypothesized to predict topic interest but not to directly predict knowledge, because individual interest factors (e.g. interest in science) are theoretically linked to increased knowledge but situational interests are not (Renninger & Hidi, 2011).
Lastly, the hypothetical model posits that preintervention variables influence their respective postintervention counterparts. We included these factors to control for the effects of prior knowledge and interests, which are known to influence subsequent knowledge and interest development (Hidi & Renninger, 2006; Tobias, 1994).
3.1.2 Sample
Treatment teachers taught the unit in their life sciences class and had a mean of 13 years of full-time teaching experience (range: 4–34 years). Two of these teachers participated in the pilot study the previous year. Class sizes ranged from 25 to 33 students per class. In their teacher logs, all six teachers reported completing all of the unit's lessons, including the field trip. One teacher reported making modifications to add more challenging content for the class including eighth-graders.
The treatment group consisted of 467 seventh- and eighth-grade students taught by six teachers at five schools in the district (Table 1). Five of these six teachers taught seventh-grade students (ages 12–13), and one teacher taught a combined class of seventh- and eighth-grade students (ages 12–14). The eighth-grade students in the combined class had not been taught similar content before. The comparison group consisted of 177 seventh-grade students, taught by four teachers from three of the same schools. Unique identifiers allowed the matching of 76% of treatment group students and 70% of comparison group students. Only data from matched questionnaires were analyzed.
| Treatment | Comparison | |
|---|---|---|
| Number of teachers | 6 | 4 |
| Number of classes | 16 | 7 |
| Number of schools | 5 | 3 |
| Mean class size | 29.3 | not collected |
| Student Sample | ||
| Total students (N) | 467 | 177 |
| Students completing preintervention questionnaire (% of N) | 429 (92%) | 177 (100%) |
| Students completing postintervention questionnaires (% of N) | 399 (85%) | 158 (89%) |
| Matched questionnairesa (% of N) | 355 (76%) | 124 (70%) |
| Student gender (student self-reported, % of matched questionnaires n) | ||
| Boy | 168 (47%) | 63 (51%) |
| Girl | 187 (53%) | 61 (49%) |
| Student race/ethnicity (teacher-reported, aggregate % only) | ||
| White or Caucasian | 79% | not |
| Black or African American | 7% | Collected |
| Hispanic or Latino | 3% | |
| Asian or Pacific Islander | 3% | |
| Native American | 1% | |
| Other (including multiracial) | 6% | |
- a The number of students whose pretests could be matched to corresponding posttests based on unique identifiers.
Treatment group teachers reported their students' race/ethnicities as aggregate percentages. Comparison group teachers did not report students' race/ethnicities, but they were likely similar because treatment and comparison teachers taught in the same schools.
3.1.3 Questionnaire design
The study's seven interest-related factors were constructed using 37 items (Appendix A). These items were selected and adapted from existing scales. These factors were selected to measure different theoretical components of interest other than positive affect, including topic interest, situational interest, desire to learn more, perceived topic importance, and broader interest in related topics (science interest).
| Treatment | Comparison | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| PreIntervention (T1) | PostIntervention (T2) | PreIntervention (T1) | PostIntervention (T2) | |||||||
| Factor name | n | Mean | SD | Mean | SD | n | Mean | SD | Mean | SD |
| Topic Knowledge | 355 | 8.68 | 4.48 | 13.72 | 5.03 | 125 | 8.04 | 4.73 | 8.15 | 4.44 |
| Science Interest | 345 | 2.91 | 1.10 | 2.96 | 1.11 | 122 | 2.96 | 1.16 | 3.03 | 1.11 |
| Situational Interest: Hands-On Activities | 352 | 3.49 | 0.94 | 3.34 | 1.02 | 124 | 3.42 | 0.89 | 3.35 | 0.93 |
| Perceived Topic Importance | 351 | 4.08 | 0.72 | 4.17 | 0.78 | 124 | 3.79 | 0.87 | 3.85 | 0.78 |
| Topic Interest | 353 | 3.43 | 0.81 | 3.34 | 0.97 | 123 | 3.40 | 0.81 | 3.30 | 0.84 |
| Desire to Learn More | 325 | 3.41 | 0.85 | 3.10 | 0.98 | 111 | 3.25 | 0.85 | 3.30 | 0.88 |
| Situational Interest: Unita | 354 | N/A | N/A | 3.03 | 1.10 | |||||
| Situational Interest: Field Tripa | 312 | N/A | N/A | 3.50 | 1.00 | |||||
- a Situational Interest: Unit and Situational Interest: Field Trip were only measured for the treatment group, postintervention.
Topic Interest measured students' individual interest in climate change and its impacts on forests. This factor was measured using the stem “This sounds interesting” (Swarat et al., 2012), followed by a list of topics addressed by the unit. This approach to assessing topic interest, by measuring interest in relevant subtopics, was validated by Drechsel et al. (2011) using data from PISA 2006 (OECD, 2007) and was also employed by the Relevance of Science Education (ROSE) Project (Sjøberg & Schreiner, 2010).
Situational interest was operationalized using three factors: Situational Interest: Hands-On Activities, Situational Interest: Field Trip, and Situational Interest: Unit. The Situational Interest: Hands-On Activities scale was designed to measure students' interest in these types of activities, independent of the unit (Hidi & Renninger, 2006). The Field Trip scale measured students' situational interest (Hidi & Renninger, 2006) in the unit's field trip. The Unit scale measured situational interest in the unit overall.
Desire to Learn More about climate change was measured using the stem “I would like to learn more about this,” followed by the same list of topics as for Topic Interest, matching Swarat et al. (2012)'s approach. Perceived Topic Importance measured the extent to which students believed that climate change was an important issue to understand. Science Interest measured students' broader interest and engagement with science, including their desire to potentially pursue a career in science.
Students' Topic Knowledge about climate change and its impacts on forests was measured using 17 multiple-choice (five response options including “I don't know”) and four true-false questions (three options including “I don't know”), totaling 21 questions. Questions focused on climate change were drawn from Leiserowitz et al.'s (2011) and Shepardson et al. (2009) studies of adolescents' knowledge about climate change. The remaining questions were designed specifically for this study, aligned with the unit's learning objectives. One coauthor, who is a forest ecologist, validated the scientific accuracy of all multiple-choice questions. Sample questions are included in Appendix A. Classical test analysis (Crocker & Algina, 2006) results confirmed the appropriateness of these questions (Table S1).
3.2 Results
3.2.1 Descriptive statistics
Treatment students rated both the unit and field trip as moderately interesting (Table 2). Situational Interest: Hands-On Activities, Topic Interest, and Desire to Learn More were also moderate pre and postintervention for both the treatment and control groups. In some contrast, students in both groups rated their Interest in Science slightly lower, and their Perceived Topic Importance slightly higher after the intervention. Topic Knowledge was moderate at the pretest for treatment and comparison groups, but higher for the treatment group at the posttest.
3.2.2 Factor analysis and path model fit
Confirmatory factor analyses showed that factor loadings and reliabilities for the seven factors were satisfactory to high (range: 0.48–0.93 preintervention and 0.66–0.94 postintervention).
The model fit indices for the hypothesized model indicated that further refinement of the model was necessary to achieve an acceptable fit (χ2 = 322.43, p < 0.001, CFI = 0.91, TLI = 0.87, RMSEA = 0.11). The subsequent, revised model (Figure 2) demonstrated sufficient overall model fit (χ2 = 199.40, p < 0.001, CFI = 0.95, TLI = 0.94, RMSEA = 0.08), and predicted students' postintervention Topic Knowledge quite well (R2 = .43). The model's significant path coefficients ranged from 0.07 to 0.78 and can be found in Figure 2 and are detailed in the Supporting Information Materials (Tables S2 and S3).
3.2.3 Changes in students' interest in and knowledge about climate change
The time–group interaction coefficients in the multilevel analyses of the factor means indicated that the treatment group differed statistically significantly from the comparison group in two of the six repeated factors as a result of the intervention (Table 4). The time coefficient for treatment students' Topic Knowledge was positive, indicating an increase in treatment students' knowledge after the intervention, whereas the time coefficient for Desire to Learn More was negative, indicating a decline in their desire to learn more. Moreover, the statistically significant time–group interaction coefficient suggests that the treatment group's postintervention Desire to Learn More was lower than the comparison group's.
| Preintervention (T1) | Postintervention (T2) | |||||||
|---|---|---|---|---|---|---|---|---|
| Treatment | Comparison | Treatment | Comparison | |||||
| Factor name and items included | Score | α | Score | α | Score | α | Score | α |
| Factor: Science Interest | ||||||||
| I sometimes think about becoming a scientist when I grow up. | 0.85 | 0.89 | 0.86 | 0.89 | 0.83 | 0.91 | 0.85 | 0.89 |
| I would like science to be a part of my job one day. | 0.89 | 0.91 | 0.88 | 0.89 | ||||
| I plan to take more science classes in the future. | 0.74 | 0.83 | 0.81 | 0.83 | ||||
| Jobs in science are extremely interesting to me. | 0.87 | 0.86 | 0.92 | 0.88 | ||||
| My friends and I discuss science-related topics. | 0.62 | 0.57 | 0.64 | 0.45 | ||||
| Factor: Perceived Topic Importance | ||||||||
| Climate change is a very important issue to me. | 0.69 | 0.83 | 0.67 | 0.85 | 0.75 | 0.87 | 0.70 | 0.86 |
| Trees and forests are very important to me. | 0.48 | 0.70 | 0.65 | 0.77 | ||||
| Climate change will be an important issue in the future. | 0.79 | 0.74 | 0.81 | 0.72 | ||||
| I think it's important to know how climate change impacts forests. | 0.75 | 0.78 | 0.81 | 0.80 | ||||
| Trees and forests are valuable to society. | 0.59 | 0.61 | 0.64 | 0.63 | ||||
| Climate change will impact forests in ways that affect all people. | 0.73 | 0.77 | 0.76 | 0.74 | ||||
| Factor: Desire to Learn More | ||||||||
| I would like to learn more about … Trees (in general). | 0.73 | 0.90 | 0.68 | 0.89 | 0.82 | 0.93 | 0.73 | 0.92 |
| … Forests (in general). | 0.70 | 0.55 | 0.78 | 0.61 | ||||
| … Climate change (in general). | 0.63 | 0.73 | 0.76 | 0.78 | ||||
| … How scientists study climate change. | 0.65 | 0.61 | 0.78 | 0.77 | ||||
| … How trees are adapted to different climates. | 0.81 | 0.67 | 0.87 | 0.85 | ||||
| … How climate change affects forests. | 0.86 | 0.83 | 0.89 | 0.83 | ||||
| … How climate change affects trees. | 0.88 | 0.89 | 0.93 | 0.85 | ||||
| Factor: Topic Interest | ||||||||
| This sounds interesting … Trees (in general). | 0.68 | 0.88 | 0.69 | 0.87 | 0.80 | 0.94 | 0.65 | 0.89 |
| … Forests (in general). | 0.67 | 0.64 | 0.80 | 0.59 | ||||
| … Climate change (in general). | 0.56 | 0.72 | 0.76 | 0.72 | ||||
| … How scientists study climate change. | 0.71 | 0.58 | 0.79 | 0.73 | ||||
| … How trees are adapted to different climates. | 0.80 | 0.66 | 0.85 | 0.81 | ||||
| … How climate change affects forests. | 0.80 | 0.68 | 0.92 | 0.85 | ||||
| … How climate change affects trees. | 0.84 | 0.75 | 0.91 | 0.79 | ||||
| Factor: Situational Interest: Hands-On Activities | ||||||||
| This sounds interesting … Working with charts and graphs. | 0.68 | 0.72 | 0.56 | 0.69 | 0.77 | 0.79 | 0.69 | 0.66 |
| Working with real-life tree samples. | 0.58 | 0.49 | 0.69 | 0.54 | ||||
| Taking scientific measurements. | 0.87 | 0.75 | 0.85 | 0.75 | ||||
| Factor: Situational Interest: Unita | ||||||||
| Boring … Exciting | 0.88 | 0.87 | ||||||
| Worthless … Valuable | 0.71 | |||||||
| Dull … Interesting | 0.92 | |||||||
| Factor: Situational Interest: Field Tripa | ||||||||
| I liked the field trip. | 0.89 | 0.92 | ||||||
| It was fun to take scientific measurements. | 0.85 | |||||||
| I learned a lot during the field trip. | 0.73 | |||||||
| The field trip was fun. | 0.89 | |||||||
| What we did during field trip helped me understand what we learned in class. | 0.66 | |||||||
| Taking scientific measurements was interesting. | 0.75 | |||||||
- a Situational Interest: Unit and Situational Interest: Field Trip were only measured for the treatment group, postintervention.
| Factor Name | Topic Knowledge | Science Interest | Perceived Topic Importance | Desire to Learn More | Topic Interest | Situational Interest: Hands-On Activities | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Potential values | 0–21 | 1–5 | 1–5 | 1–5 | 1–5 | 1–5 | ||||||||||||
| Model coefficients | β | SE | p | β | SE | p | β | SE | p | β | SE | p | β | SE | p | β | SE | p |
| Time (effect on treatment) | 5.03 | 0.20 | *** | 0.06 | 0.04 | ns | 0.08 | 0.04 | * | −0.32 | 0.04 | *** | −0.08 | 0.04 | * | −0.16 | 0.05 | *** |
| Group (treatment = 0) | −0.79 | 0.95 | ns | 0.08 | 0.14 | ns | −0.17 | 0.18 | ns | −0.20 | 0.12 | ns | 0.03 | 0.14 | ns | −0.09 | 0.11 | ns |
| Time–Group interactiona | −4.93 | 0.40 | *** | 0.01 | 0.08 | ns | −0.03 | 0.07 | ns | 0.38 | 0.09 | *** | −0.02 | 0.08 | ns | 0.09 | 0.09 | ns |
| Gender (male = 0) | −0.30 | 0.38 | ns | −0.37 | 0.09 | *** | −0.10 | 0.06 | ns | −0.16 | 0.07 | * | −0.23 | 0.07 | *** | −0.29 | 0.08 | *** |
| L1 residual variance | 7.34 | 0.26 | 0.25 | 0.32 | 0.32 | 0.40 | ||||||||||||
| L2 intercept variance (students nested within teacher) | 13.75 | 0.93 | 0.29 | 0.49 | 0.42 | 0.51 | ||||||||||||
| L3 intercept variance (teacher) | 1.55 | 0.01 | 0.06 | 0.01 | 0.02 | 0.00 | ||||||||||||
- Abbreviation: ns, not significant at p < 0.05.
- a Time–Group Interaction measures comparison group effects at the postintervention period. Negative β coefficients indicate a negative effect of being in the comparison group after the intervention (and thus a positive effect of being in the treatment group).
- *p < 0.05, **p < 0.01, and ***p < 0.01.
3.2.4 Relationships among students' interest in climate change, their knowledge about climate change, and related factors
Students' situational and individual interests, mediated by their desire to learn more, had no effects on topic knowledge before or after the intervention. There was, however, a small, but statistically significant, indirect relationship between students' preintervention Desire to Learn More (LearnCC1) and their postintervention Topic Knowledge (Knowledge2; indirect effects coefficient = 0.03, p < 0.01). This indirect effect emerged from the paths between students' initial desire to learn more, situational interest in the unit, postintervention perceived topic importance, and topic knowledge (LearnCC1⟶InterestUnit⟶ImportanceCC2⟶Knowledge2). Furthermore, students' situational interest in the unit had a small indirect effect on postintervention topic knowledge through its direct relationship with perceived topic importance within this same path (indirect effects coefficient = 0.06, p = 0.001).
Science Interest (InterestScience1) had a small direct relationship with Topic Interest (InterestCC1) before the intervention, and no direct relationship afterward. Thus, students who were more interested in science were also slightly more interested in climate change, but only before the intervention. Science Interest (InterestScience), however, had direct effects on Topic Knowledge (Knowledge) both before and after the intervention. This direct relationship was relatively large before the intervention, but much smaller afterward, indicating that students with a greater interest in science were more knowledgeable about climate change initially, but that their interest in science had little additional effect on their subsequent topic knowledge.
Perceived Topic Importance (ImportanceCC) had direct effects on Topic Interest (InterestCC) before and after the intervention. In other words, students who considered climate change an important topic were more likely to be interested in it. Perceived Topic Importance (ImportanceCC2) also had a moderate relationship with Topic Knowledge (Knowledge2) after the intervention, indicating that students who thought the topic was more important were more likely to know more about it after the unit.
4 STUDY 2
4.1 Procedure
4.1.1 Changes building on study 1
On the basis of Study 1 results, modifications were made to the intervention as well as the model of relationships between students' interest(s) and knowledge about climate change and its impacts on forests.
The modification to the intervention consisted of the addition of new content and activities about climate change and forest ecology careers. The decision to add these resources was based on findings from Study 1 showing that the existing unit did not increase students' interest in climate change and forests, and research indicating that information about science careers can foster students' situational interest in science activities (Tyler-Wood et al., 2012). One of the purposes of Study 2 was to test the revised unit in terms of its effectiveness on outcomes of interest. We also conducted analyses to determine if we could replicate key results from Study 1.
Modifications to the model exploring the relationship between students' interest(s) and knowledge about climate change and its impacts on forests consisted of a number of changes (Figure 3 and Table 5). One change involved removing Situational Interest: Field Trip and Situational Interest: Hands-On Activities. These two variables were removed because Study 1 did not suggest that these factors played an important role in predicting knowledge, and to allow for the inclusion of two other potentially relevant factors: Perception of Threat and Behavioral Self-Efficacy. Perception of Threat (i.e., the belief that climate change will impact oneself and others) was included because of the danger climate change poses to students (Busch & Osborne, 2014; National Research Council, 2012) and Behavioral Self-Efficacy (i.e., the belief that one can take action to address climate change) because of its strong influence on individuals' attitudes toward (Kellstedt et al., 2008; Milfont, 2012), and actions on climate change (Anderson, 2012; Stevenson & Peterson, 2015). Moreover, perception of threat has been linked to adolescents' topic interest (Carman et al., 2017; Sjöberg, 2007) to climate change information-seeking behavior (Mead et al., 2012), that is, a variable consistent with this study's Desire to Learn More.
| Study 1 Factor Complete Name (Abbreviation) | In Study 1 | In Study 2 |
|---|---|---|
| Interest-related variables | ||
| Situational Interest in the Unit Overall (InterestUnit)a | X | X |
| Situational Interest in the Field Trip (InterestFT)a | X | |
| Situational Interest in Hands-On Science Activities (InterestAct) | X | |
| Topic Interest in Climate Change and Forests (InterestCC) | X | X |
| Desire to Learn More about Climate Change and Forests (LearnCC) | X | X |
| General Interest in School Science (InterestScience) | X | |
| Self-Efficacy Regarding Ability to Address Climate Change Impacts (SelfEff) | X | |
| Perception of Threat Posed by Climate Change to Self and Others (Percep) | X | |
| Interest in Careers Related to Climate Change and Forest Science (InterestCareer) | X | |
| Perceived Importance of Climate Change and Forests to Self and Society (ImportanceCC) | X | X |
| Dependent variable | ||
| Topic Knowledge about Climate Change and Forests (Knowledge) | X | X |
- Note: Boldface indicates factors that are measured in both studies.
- a Factors that were measured in the posttest with treatment students only.
Another change to the model consisted of replacing Science Interest with Career Interest to measure students' interests in climate change and forestry careers. This change was made for three reasons. First, this new factor would measure the effects of the addition of career information to the unit. Second, science career interest has been found to be a more effective indicator of long-term science engagement than science interest (Archer et al., 2013). Third, two of the five items in our Science Interest scale addressed career interest (Table 3). Thus, testing both Science Interest and Career Interest would have been redundant.
Several of the variables in the model were retained for Study 2, to allow for replication of select Study 1 results. Situational Interest: Unit was retained because Study 1 results suggested it had indirect effects on postintervention Topic Knowledge. Desire to Learn More was retained to retest its relationship to Topic Knowledge. Perceived Topic Importance was retained because it had direct effects on Topic Interest and Topic Knowledge. Topic Interest was retained because Perceived Topic Importance had a direct effect on it, and Topic Interest had a direct effect on Desire to Learn More.
4.1.2 Revised hypothetical model
The revised hypothetical model for Study 2 differed from that for Study 1 with regard to changes in the factors that were included (Figure 3) and included some changes among the hypothesized relationships in the model. Because Study 1 results showed that Situational Interest: Unit had unexpected direct effects on both Topic Interest and Perceived Topic Importance, we retested both of these relationships in Study 2. Furthermore, we hypothesized that Situational Interest: Unit may play a greater role in predicting the unit's outcomes than captured in Study 1. Therefore, we also tested the relationship between Situational Interest: Unit and the three new factors (Career Interest, Perception of Threat, and Self-Efficacy). Next, although Study 1 results identified no statistically significant direct relationship between Desire to Learn More and Topic Knowledge at either time period, we retested this relationship in Study 2 due to past empirical evidence linking these two concepts (Laukenmann et al., 2003; Schraw et al., 2001). Finally, Study 2's hypothetical model again posits that preintervention variables influence their respective postintervention counterparts (Figure 3).
4.1.3 Sample
Treatment group teachers had a mean of 18 years of full-time teaching experience (range: 6–35 years). Class sizes ranged from 28 to 33 students per class. In their teacher logs, three treatment group teachers reported completing all of the unit's lessons, including all of the optional career components. Four teachers reported completing the field trip. One teacher reported that her students did not complete the field trip, and only completed seven of nine lesson days and three of five optional career activities. However, this teacher's results did not significantly differ from other teachers' results on any of the measured factors, so their results were retained for the final analysis.
The Study 2 treatment group consisted of 363 seventh-grade students (ages 12–13) taught by five teachers at three schools, and the comparison group consisted of 219 seventh-grade students taught by four teachers from the same schools (Table 6). Three treatment group teachers and one comparison group teacher from Study 1 participated in Study 2; none of these teachers changed groups. Unique identifiers allowed the matching of 63% of treatment group students and 91% of comparison group students. The relatively lower percentage of useable treatment group data occurred because a smaller percentage of these students returned permission forms to participate in the study (77%). As was the case for Study 1, only data from matched questionnaires were analyzed.
| Treatment | Comparison | |
|---|---|---|
| Number of teachers | 5 | 4 |
| Number of classes | 12 | 8 |
| Number of schools | 3 | 3 |
| Mean class size | 30.3 | 32.4 |
| Student sample | ||
| Total students (N) | 363 | 219 |
| Students completing preintervention questionnaire (% of N) | 273 (75%)a | 219 (100%) |
| Students completing postintervention questionnaires (% of N) | 250 (69%) | 211 (96%) |
| Matched questionnairesb (% of N) | 229 (63%) | 199 (91%) |
| Student gender (student self-reported, % of matched questionnaires n) | ||
| Boy | 91 (40%) | 76 (38%) |
| Girl | 118 (52%) | 101 (51%) |
| No responsec | 20 (9%) | 22 (11%) |
| Student race/ethnicity (student self-reported, % of matched questionnaires n) | ||
| White or Caucasian | 89 (39%) | 98 (49%) |
| Black or African American | 13 (6%) | 11 (6%) |
| Hispanic or Latino | 7 (3%) | 5 (3%) |
| Asian | 29 (13%) | 16 (8%) |
| Native American or Alaska Native | 2 (1%) | 0 (0%) |
| Native Hawaiian or Pacific Islander | 2 (1%) | 0 (0%) |
| Other (not including multiracial) | 14 (6%) | 10 (5%) |
| Multiracial | 39 (17%) | 29 (15%) |
| No responsec | 34 (15%) | 30 (15%) |
- a The low response rate for this group resulted from the relatively low percentage of students who obtained parent permission to participate in the study. Of 294 students (81%) who returned signed parental consent forms, only 281 students (77%) received permission to participate.
- b Number of students whose pretests could be matched to corresponding posttests based on unique identifiers.
- c Includes students who selected the “I prefer not to respond” option or left this question blank.
Students in both the treatment and comparison groups self-reported their gender and race/ethnicity, and the gender and race/ethnicity distributions were about the same for both groups.
4.1.4 Questionnaire design
The study's eight interest-related factors were constructed using 38 items (Appendix A). We removed some items with lower loading scores from the Topic Interest, Desire to Learn More, and Perceived Topic Importance scales from Study 1 to reduce the length of the instrument. The Topic Knowledge questions were slightly revised in light of the unit's new content, re-reviewed by the same forest ecologist as in Study 1, and retested using classical test analysis (Supporting Information Materials, Table S4).
Perception of Threat items measured students' belief that climate change is occurring, caused by humans, and likely to affect themselves and others. Self-efficacy was measured by asking students to rate their perceived ability to address climate change, either alone or working with others. Career Interest measured students' interest in pursuing careers related to climate change science.
4.2 Results
4.2.1 Descriptive statistics
Treatment group students rated the unit as moderately interesting. Career Interest, Topic Interest, and Desire to Learn More were also moderate at both time periods for both treatment and control groups. Self-Efficacy, Perception of Threat, and Perceived Topic Importance were higher at both time periods for both groups (Table 7). Consistent with Study 1, Topic Knowledge was moderate at the pretests for both treatment and comparison groups and was higher for the treatment group in the posttest.
| Treatment | Comparison | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Preintervention (T1) | Postintervention (T2) | Preintervention (T1) | Postintervention (T2) | |||||||
| Factor Name | n | Mean | SD | Mean | SD | n | Mean | SD | Mean | SD |
| Topic Knowledge | 229 | 9.37 | 4.14 | 14.22 | 5.39 | 198 | 9.29 | 4.46 | 9.21 | 4.86 |
| Perceived Topic Importance | 225 | 3.90 | 0.90 | 4.09 | 0.88 | 199 | 3.77 | 0.88 | 3.65 | 0.88 |
| Self-Efficacy | 227 | 3.94 | 0.84 | 4.12 | 0.87 | 198 | 3.85 | 0.87 | 3.69 | 0.96 |
| Perception of Threat | 227 | 3.92 | 0.80 | 4.09 | 0.78 | 199 | 3.86 | 0.83 | 3.77 | 0.92 |
| Desire to Learn More | 213 | 3.48 | 0.93 | 3.35 | 0.94 | 188 | 3.36 | 0.82 | 3.06 | 0.90 |
| Career Interest | 229 | 2.83 | 0.95 | 2.88 | 1.06 | 199 | 2.77 | 0.97 | 2.73 | 1.01 |
| Topic Interest | 227 | 3.46 | 0.84 | 3.40 | 0.95 | 196 | 3.32 | 0.74 | 3.16 | 0.83 |
| Situational Interest: Unita | 227 | N/A | N/A | 3.34 | 1.08 | |||||
- a Situational Interest: Unit was only measured for the treatment group, postintervention.
4.2.2 Factor analyses and path model fit
Factor loadings for the eight factors (Table 8) were satisfactory to high (range: 0.56–0.93) and reliabilities were high (range: 0.77–0.94). Factor item means were used in subsequent multilevel and path analyses.
| Preintervention (T1) | Postintervention (T2) | |||||||
|---|---|---|---|---|---|---|---|---|
| Treatment | Comparison | Treatment | Comparison | |||||
| Factor name and items included | Score | α | Score | α | Score | α | Score | α |
| Factor: Perceived Topic Importance | ||||||||
| Climate change is a very important issue to me. | 0.83 | 0.80 | 0.74 | 0.77 | 0.89 | 0.78 | 0.84 | 0.79 |
| I think it's important to know how climate change impacts forests. | 0.69 | 0.77 | 0.64 | 0.71 | ||||
| Climate change will be an important issue in the future. | 0.78 | 0.73 | 0.71 | 0.75 | ||||
| Factor: Self-Efficacy | ||||||||
| By working with others, I can address climate change | 0.74 | 0.78 | 0.75 | 0.80 | 0.73 | 0.82 | 0.84 | 0.87 |
| There are things I can do that will address climate change | 0.72 | 0.74 | 0.76 | 0.81 | ||||
| If everyone does their part, we can address climate change | 0.79 | 0.80 | 0.85 | 0.84 | ||||
| Factor: Perception of Climate Change Threat | ||||||||
| Ann Arbor is already being affected by climate change. | 0.71 | 0.89 | 0.75 | 0.89 | 0.81 | 0.93 | 0.82 | 0.94 |
| Climate change already has an impact on people like me. | 0.77 | 0.72 | 0.83 | 0.79 | ||||
| Climate change already has an impact on people different from me. | 0.75 | 0.75 | 0.83 | 0.84 | ||||
| Michigan is already experiencing the effects of climate change. | 0.77 | 0.76 | 0.83 | 0.86 | ||||
| Climate change already poses a threat to forests. | 0.73 | 0.76 | 0.78 | 0.78 | ||||
| Ecosystems are already being affected by climate change. | 0.74 | 0.78 | 0.76 | 0.86 | ||||
| Climate change already impacts forests in ways that affect people all over the world. | 0.71 | 0.75 | 0.81 | 0.87 | ||||
| Factor: Desire to Learn More | ||||||||
| I would like to learn more about … Forests (in general). | 0.75 | 0.87 | 0.69 | 0.81 | 0.75 | 0.88 | 0.76 | 0.88 |
| … Climate change (in general). | 0.65 | 0.56 | 0.82 | 0.74 | ||||
| … How scientists study climate change. | 0.74 | 0.66 | 0.72 | 0.73 | ||||
| … How trees are adapted to different climates. | 0.82 | 0.75 | 0.78 | 0.79 | ||||
| … How climate change affects forests. | 0.86 | 0.73 | 0.87 | 0.86 | ||||
| Factor: Career Interest | ||||||||
| Jobs in … science (generally) are very interesting to me. | 0.61 | 0.87 | 0.60 | 0.88 | 0.67 | 0.91 | 0.63 | 0.91 |
| … Climate science. | 0.75 | 0.75 | 0.86 | 0.85 | ||||
| … Forest ecology. | 0.73 | 0.70 | 0.83 | 0.84 | ||||
| Someday I would like to become a … scientist. | 0.66 | 0.83 | 0.66 | 0.70 | ||||
| … Climate scientist. | 0.84 | 0.83 | 0.88 | 0.87 | ||||
| … Forest ecologist. | 0.83 | 0.63 | 0.86 | 0.87 | ||||
| Factor: Topic Interest | ||||||||
| This sounds interesting … Forests (in general). | 0.71 | 0.86 | 0.66 | 0.77 | 0.71 | 0.89 | 0.74 | 0.86 |
| … Climate change (in general). | 0.60 | 0.54 | 0.70 | 0.74 | ||||
| … How scientists study climate change. | 0.68 | 0.66 | 0.80 | 0.69 | ||||
| … How trees are adapted to different climates. | 0.79 | 0.70 | 0.85 | 0.72 | ||||
| … How climate change affects forests. | 0.82 | 0.74 | 0.90 | 0.87 | ||||
| Factor: Situational Interest: Unita | ||||||||
| Boring … Exciting | 0.88 | 0.88 | ||||||
| Worthless … Valuable | 0.75 | |||||||
| Dull … Interesting | 0.93 | |||||||
- a Situational Interest: Unit was only measured for the treatment group, post-intervention.
The hypothesized path model had an acceptable model fit (CFI = 0.95, TLI = 0.93, RMSEA = 0.08, χ2 = 151.37, p < 0.001) but included some nonsignificant paths. We subsequently removed these paths to present the most parsimonious model (Figure 4). This revised model also demonstrated acceptable fit (CFI = 0.95, TLI = 0.93, RMSEA = 0.08, χ2 = 162.64, p < 0.001). The model predicted students' postintervention Topic Knowledge quite well (R2 = 0.41). Complete results for the final path model are summarized in the Supporting Information Materials (Table S5 and S6).
4.2.3 Replication of key findings from study 1
The multilevel analyses' time–group interaction coefficients indicated that the treatment group differed significantly from the comparison group in five of the seven repeated factors as a result of the intervention (Table 9). The time coefficients for treatment students' Topic Knowledge, Perceived Topic Importance, Self-Efficacy, and Perception of Threat, were positive, indicating a statistically significant increase in these factors after the intervention. Although the time coefficient for Desire to Learn More was negative, indicating a decline in treatment students' desire to learn more, the statistically significant time–group interaction coefficient suggests that the treatment group students' postintervention Desire to Learn More was higher than the comparison group's.
| Variable Name | Topic Knowledge | Perceived Topic Importance | Self-Efficacy | Perception of Threat | Desire to Learn More | Career Interest | Topic Interest | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Potential Values | 0–21 | 1–5 | 1–5 | 1–5 | 1–5 | 1–5 | 1–5 | ||||||||||||||
| Model Coefficients | β | SE | p | β | SE | p | β | SE | p | β | SE | p | β | SE | p | β | SE | P | β | SE | p |
| Time (effect on treatment) | 5.07 | 0.27 | *** | 0.20 | 0.05 | *** | 0.22 | 0.05 | *** | 0.20 | 0.04 | *** | −0.12 | 0.06 | * | 0.07 | 0.05 | ns | −0.08 | −0.05 | ns |
| Group (treatment = 0) | −0.32 | 0.56 | ns | −0.16 | 0.12 | ns | −0.14 | 0.09 | ns | −0.07 | 0.11 | ns | −0.15 | 0.16 | ns | −0.11 | 0.18 | ns | −0.18 | 0.15 | ns |
| Time–Group Interactiona† | −5.05 | 0.40 | *** | −0.32 | 0.08 | *** | −0.35 | 0.09 | *** | −0.30 | 0.07 | *** | −0.18 | 0.09 | * | −0.11 | 0.08 | ns | −0.08 | 0.08 | ns |
| Gender (male = 0) | −0.34 | 0.43 | ns | 0.02 | 0.08 | ns | −0.02 | 0.07 | ns | 0.14 | 0.08 | ns | −0.04 | 0.08 | ns | −0.05 | 0.09 | ns | −0.06 | 0.08 | ns |
| L1 Residual Variance | 7.65 | 0.28 | 0.31 | 0.21 | 0.33 | 0.28 | 0.28 | ||||||||||||||
| L2 Intercept Variance (students within teacher) | 13.31 | 0.42 | 0.37 | 0.43 | 0.41 | 0.67 | 0.39 | ||||||||||||||
| L3 Intercept Variance (teachers) | 0.17 | 0.01 | 0.00 | 0.01 | 0.03 | 0.05 | 0.03 | ||||||||||||||
- Abbreviation: ns, not significant at p < 0.05.
- a Time–Group Interaction measures comparison group effects at the postintervention period. Negative β coefficients indicate a negative effect of being in the comparison group after the intervention (and thus a positive effect of being in the treatment group).
- *p < 0.05, **p < 0.01, and ***p < 0.001.
These multilevel model results replicate findings from Study 1 of a significant positive time effect and time–group interaction effect on Topic Knowledge, a significant negative time effect on Desire to Learn More, and that neither time nor group had an effect on Topic Interest. In contrast, Study 2 results do not replicate the Study 1 finding that Desire to Learn More was lower for the treatment than the comparison group after the intervention. Instead, Study 2 results showed that Desire to Learn More was higher for the treatment group than for the comparison group after the intervention. Furthermore, Study 2 showed a positive time effect and time–group interaction effect for Perceived Topic Importance, indicating that the intervention had a positive impact on students' perceptions of climate change as an important topic. This is in contrast to the Study 1 model, which did not identify a significant time–group interaction coefficient for Perceived Topic Importance.
Many path model results were also similar to those from Study 1 (Figure 4). Like in Study 1, the path model in Study 2 identified no significant connection between Desire to Learn More and Topic Knowledge in either time period but showed a statistically significant positive connection between preintervention Desire to Learn More and postintervention Situational Interest: Unit. It also showed that Desire to Learn More had indirect effects on Topic Knowledge through the same path of preintervention Desire to Learn More (LearnCC1), Situational Interest: Unit (InterestUnit), postintervention Perceived Topic Importance (ImportanceCC2), and postintervention Topic Knowledge (Knowledge2; indirect effects coefficient = 0.09, p < 0.001).
Some path model results from Study 1 were not replicated in Study 2. First, the Study 2 path model showed a statistically significant positive connection between preintervention Perceived Topic Importance (ImportanceCC1) and preintervention Topic Knowledge (Knowledge1), whereas the connection between them was not significant in Study 1. Furthermore, the stability coefficient for Topic Interest (InterestCC1⟶InterestCC2) was statistically significant in Study 2, whereas it was not significant in Study 1. This significant stability coefficient in Study 2 appeared to replace the LearnCC1⟶InterestCC2 path identified in Study 1.
4.2.4 Additional path model relationships
The path model showed that Perceived Topic Importance and Self-Efficacy had statistically significant effects on Topic Knowledge during both time periods, whereas Perception of Threat and Career Interest did not. These results suggest that students' perceptions of the importance of climate change and the ability to take action to address it were positively linked to what they learned about climate change and its impacts on forests as a result of the intervention.
Furthermore, the path model showed that students' level of interest in the unit (Situational Interest: Unit) had indirect effects on their postintervention Topic Knowledge, mediated by Self-Efficacy and Perceived Topic Importance (indirect effects coefficient = 0.17, p < 0.001). This result is consistent with Study 1 findings suggesting that Situational Interest: Unit had significant indirect effects on postintervention Topic Knowledge, mediated by postintervention Perceived Topic Importance.
5 DISCUSSION
Given the impact that climate change already has and will increasingly have on students' lives, strengthening their knowledge about climate change is critical. Overall, relatively little is known about how best to accomplish this goal. We contribute to addressing this gap by exploring the roles different types of interests and related factors played in predicting two cohorts of predominantly seventh-grade (age 12–13) students' knowledge about climate change before and after a 2-week educational intervention. The two studies' overarching goals were to (1) assess to what extent a short climate change education intervention changed students' knowledge about the topic and (2) investigate to what extent students' interests and related factors predict their knowledge and learning about climate change, to help advance climate change education research and practice.
5.1 Effects of an educational intervention on student interest in and knowledge about climate change
Both studies' multilevel model analyses results showed that adolescents' knowledge about climate change increased as a result of the 2-week climate change unit. These results are important because adolescents have many misconceptions about climate change (Leiserowitz et al., 2011; Shepardson et al., 2014), and formal climate change education interventions, particularly in K–12 education, are often short in duration (e.g., Nussbaum et al., 2016; Varela et al., 2018). Our studies' findings are consistent with other researchers' conclusions that students are more likely to learn about climate change (Anderson, 2012; Busch & Osborne, 2014) and science in general (NGSS Lead States, 2013), when the topics that are addressed are relevant to students' lives, as has been shown to be the case with regional trees and forests (Sobel, 1995).
Furthermore, our results provide additional evidence that scientific modeling activities can help middle school students learn about climate change. Similar to Lombardi et al. (2018), who found that scientific modeling played an important part in adolescents' plausibility appraisals of climate change, the two studies' results suggest that students' participation in mathematical modeling activities can increase their understanding about the causes of climate change and its impacts. Like Lombardi et al. (2018), we suspect that this is the case because learning how to critically evaluate and draw conclusions from data helps students understand how researchers conduct climate science. Our results also support Monroe et al.'s (2019) recommendation, based on their literature review of climate change education research, that drawing on real-world climate science should be a climate change education best practice.
Another promising result was that while the unit increased students' perceptions of the threats posed by climate change, it also strengthened their belief that they can take actions to help address climate change (i.e., behavioral self-efficacy). Students' increased perception of the threats posed by climate change as a result of the revised unit is contrary to findings from an earlier study (Carman et al., 2017) when threat perception did not change in response to a pilot version of the unit. It is unclear why these findings are inconsistent but may be due to the improvements that were made to the pilot unit in collaboration with participating teachers. The increase in self-efficacy, which was not measured as part of that earlier pilot study, is relevant because we are not aware of climate change education interventions for this age group that has been able to achieve this change. Behavioral self-efficacy is a particularly important outcome because similar constructs such as hope have been linked to increased concern and likelihood to act on climate change (Ojala, 2012; Stevenson & Peterson, 2015). Moreover, self-efficacy is particularly challenging to strengthen because the threats posed by climate change can overwhelm and frighten students (Busch & Osborne, 2014).
In contrast to increasing students' knowledge about climate change as well as their perceived topic importance and self-efficacy for addressing it, revised versions of the unit did not change students' interest in climate change. This is consistent with interest theory, which suggests that students' topic interest may not be easily changed over a short time (Hidi & Renninger, 2006; Renninger, 1992). Nonetheless, some studies of relatively short educational interventions have found positive impacts on students' topic interest. We argue that it may be due to the lack of distinction these educational researchers have made between topic interest and desire to learn more. For example, upon closer examination of Nussbaum et al.'s (2016) study of using games to teach about climate change impacts on local water resources, we discovered that they asked students to describe their level of interest in “knowing more” and “learning more” about various topics (p. 799). In other words, rather than measuring only topic interest, they also measured students' desire to learn more. As our studies have shown, studies of students' interest should distinguish these two constructs.
At the same time, unlike Nussbaum et al. (2016), who found a small but significant increase in students' desire to learn more about climate change topics after an educational game, our unit did not have such an effect. In fact, consistent with a prior pilot study (Carman et al., 2017), students' desire to learn more declined as a result of the unit used in both of our studies. In the first study, the decline was greater for the treatment than the comparison group, whereas in the second study, the decline was greater for the comparison than the treatment group. Rotgans and Schmidt (2014), who studied students' situational interest in lessons about historical topics and ecological problems, suggested that students' “thirst for knowledge” (p. 37) may have been satisfied by what they learned. As a result, they may, therefore, not have felt motivated to learn even more about it.
The unit did not have consistent effects on the remaining measured factors. Treatment students' overall interest in hands-on activities did not differ from those of the comparison group in the first study. This may be because these interests build over time and are less likely to change in response to a single experience (Hidi & Renninger, 2006; Renninger & Hidi, 2011). Students' interest in science also remained the same, but this is not surprising given that students' general interest in science is unlikely to change as a result of a short intervention (Renninger, 1992). In contrast, the lack of change in the second study's students' interest in careers related to climate change and forestry was disappointing. Similar to Tyler-Wood et al. (2012), who engaged students in science and environmental issues using career materials, we anticipated that providing such information would increase students' interest in related careers. Yet, even after removing the students whose instructor did not teach the career activities, we found no significant change after the intervention. One reason for this result may be that in contrast to Tyler-Wood et al. (2012) our intervention was much shorter in duration (i.e., 2 weeks vs. 1 academic year). Another explanation may be that the students were already aware of climate-change-related career opportunities because of prior exposure to scientists, including climate scientists, from the large research university in the city in which this study was conducted. This possible explanation is consistent with research suggesting that interest in science careers is influenced by a variety of factors outside the classroom, including family members and out-of-school experiences (Hecht et al., 2019).
5.2 Relationships between middle school students' interest and knowledge regarding climate change
Altogether, our research suggests that the relationship between students' interest in climate change and the extent to which they subsequently learn about the topic after an intervention is indirect and very small. These findings are consistent with those from other researchers who have explored the link between interest and knowledge and found the relationship to be only an indirect one (e.g., Ainley et al., 2002). Notably, however, our findings are contrary to other theoretical studies' findings, which have suggested that the link between students' interest in a topic and their learning about the topic depends on their desire to learn more about the topic (Ainley, 2006; Voss & Schauble, 1992). Neither of our two studies identified such a link between students' desire to learn and knowledge at either time period. Instead, our results are consistent with Rotgans and Schmidt (2014), who found that when students' desire to learn is satisfied by a lesson, they gain in knowledge but decline in their desire to learn more. Nonetheless, we acknowledge that desire to learn more about a topic has been found to characterize interest development, particularly as situational interests develop into longer-term individual interests (Renninger & Hidi, 2011). Research is needed to determine whether desire to learn more may have different effects on student learning about climate change over longer time periods.
In contrast to students' interest in climate change, students' overall interest in science had a much stronger relationship with their knowledge about climate change. For one, students' overall interest in science had a direct effect on their preintervention knowledge about climate change, which is consistent with a cross-sectional study by Dijkstra and Goedhart (2012) that found a similar link. One potential reason for this finding is that students with a greater interest in science may have been more likely to explore information about scientific topics, including climate change, before the intervention. Researchers have found that individuals with broad subject interests, like in science, are more likely to seek out information on multiple sub-areas of that subject, leading to increased knowledge about those subareas (Renninger & Hidi, 2016; Tobias, 1994).
Our study builds on and extends past cross-sectional research by examining the relationship between interest and knowledge both before and after an educational intervention. As a result, we learned that students' overall interest in science has a strong direct relationship with their preintervention knowledge about climate change, which subsequently is positively associated with their knowledge after the intervention. In combination with the multilevel model results, these findings suggest climate change education can support gains in students' knowledge about climate change, regardless of their initial interest in science. Those with a greater interest in science, however, will have a stronger foundation to build on because they may know more about climate change than their peers who are less interested in science.
In both studies, situational interest in the unit had significant indirect effects on post-intervention knowledge through its direct effects on perceived topic importance and self-efficacy. These results are inconsistent with those by Tapola et al. (2013), who found no significant link between situational interest in a learning task and learning from that task but are consistent with those by Laukenmann et al. (2003), who found a positive correlation between situational interest and learning. One reason why our results are consistent with those by Laukenmann et al. (2003) may be that the latter measured students' perceived importance as part of situational interests. Our two studies' results suggest that perceived topic importance and situational interest should be treated as distinct concepts, and moreover, that perceived topic importance mediates the relationship between situational interest and learning. In other words, students who are interested in a climate change lesson may learn more from it, but the extent to which they will do so, will depend on how important of a topic they perceive climate change to be.
5.3 Interest-related factors and knowledge
Our two studies show that there are important links between students' perception of the importance of climate change and their knowledge as well as learning about the topic. Specifically, our study suggests that students who believe that climate change is an important issue are more likely to know more about it, independent of their perception of the threats posed by climate change or beliefs that they can act to help address this challenge. Other scholars have similarly linked perceived topic importance to interest in, and knowledge about, science (Jack & Lin, 2014; OECD, 2007), but our study is the first to do so in the context of climate change.
Our second study, which measured behavioral self-efficacy, suggests that there are also important relationships between students' beliefs that they can act to help address climate change, knowledge, and subsequent learning about the topic. Our path model shows that students' who were more likely to believe that they can act to help address climate change, knew more about the topic both before and after the intervention. This finding is somewhat consistent with Mead et al.'s (2012) finding that adolescents with greater perceived ability to act in response to climate change threats, along with greater threat perception, were more likely to seek out more information about climate change. However, whereas their cross-sectional study suggested that threat perception plays a more important role in information-seeking, our experimental study shows that behavioral self-efficacy actually does so. This conclusion is aligned with other climate change education research showing that promoting students' behavioral self-efficacy benefits their mental health and well-being (Ojala, 2012) and increases their likelihood of acting on climate change (Anderson, 2012; Stevenson & Peterson, 2015).
Moreover, in contrast to prior research, our studies identified no linkage between perception of climate change threat and knowledge about the topic, either before or after an intervention. One of these prior cross-sectional studies by Stevenson et al. (2014) found that knowledge about climate change is an indirect predictor of climate change risk perception for seventh- and eighth-grade students with individualist worldviews. Another prior study by Aksit et al. (2018) found that undergraduate students' knowledge about climate change predicted their risk perception before an educational intervention, but did not measure the relationship after the intervention. Our results may differ from those found in these studies because they did not focus on knowledge about climate change as a dependent variable.
Lastly, we found no evidence that students' interest in careers related to climate change and forests contributed to their knowledge or learning about climate change. We expected that students' career interest would serve a similar role to their science interest (cf. Archer et al., 2013), but our results suggest that it did not. In the first study, the Science Interest factor, which included two career interest items, was relatively strongly related to both students' prior and subsequent knowledge about climate change, whereas the Career Interest factor in study two was not. However, there was a relatively strong relationship between the Career Interest and Topic Interest factors in the second study. Our findings thus suggest that interest in science and in careers may play different important roles in climate change education. Students' career interests support their interest in climate change, whereas students' interest in science supports their knowledge about the topic. These findings are similar to those by the ASPIRES longitudinal study showing that career interests are distinct from the overall interest in science (Archer et al., 2013). However, our results do not support their conclusion that interest in science is less important to learning than the interest in careers is, at least within the context of climate change education.
6 STUDY LIMITATIONS
Our study has five main limitations. First, our study was solely quantitative, and thus, cannot offer the same level of in-depth insights into interest development as qualitative investigations (Ardoin et al., 2014; Freeman et al., 2002). Second, our study's students were from a well-educated and relatively affluent school district, limiting the generalizability of study results. Environmental justice research suggests that climate change education interventions may be received differently by less affluent populations (Stevenson & Peterson, 2015; Taylor, 2014). Third, the district's community is politically liberal, and thus, students were more likely to have parents who express concern about climate change and support students' learning about it. Students from more politically conservative communities may have responded differently to the unit. Fourth, our two studies used path analysis because the sample sizes were not large enough for structural equation modeling analysis. In contrast to path analysis, structural equation modeling would have allowed for incorporating measurement error. A final limitation is that the sample for Study 1 consisted of one class that included both seventh- and eighth-grade students. As we only recruited seventh-grade teachers, this was not expected. We decided to use these students' data because our analysis allowed us to control for this class' higher pretest knowledge scores, and we determined that they experienced similar increases in knowledge as the remaining students. We acknowledge that this class may have experienced the unit somewhat differently from other classes.
7 CONCLUSIONS AND IMPLICATIONS
Students' interest, or more accurately, different types of interests, appear to play a complex but for the most part, a minor role in supporting the development of their foundational knowledge about climate change through formal education. Although most of the interests we investigated did not appear to be key to students' existing knowledge or learning about climate change, two interests are worthy of attention by those seeking to enhance students' foundational knowledge about climate change: students' situational interest in the unit and their general interest in science. Our studies show that students' interest in the unit had indirect but significant associations with students' postintervention knowledge about climate change. This finding suggests that more research is needed to determine how lessons about climate change can be made novel and meaningful for students (Bergin, 1999; Jack & Lin, 2014), and to incorporate these features into climate change education. Consistent with research on interest development (Hidi & Renninger, 2006), which finds that long-term interests are more likely to predict knowledge, we found that students with a greater interest in science were initially more interested in, and knowledgeable about climate change, as well as more likely to be situationally interested in climate change education and thus, to learn about the topic. To support students' learning about climate change, it is important that researchers continue to explore what activities, contexts, and other features may make climate change education units interesting to students, and prompt them to want to learn more about this topic, along with the broader challenge of increasing their overall interest in science. Because individual interest develops continually, it is also important to investigate how interest in climate change affects students in other ways over a longer time period. For example, future research could include a developmentally oriented measurement of interest (e.g., Rotgans, 2015) to assess whether students with more well-developed individual interest in climate change continue to perceive differences in the importance and threat of climate change.
The two studies reported here provide important insights into three interest-related factors and their respective roles, or lack thereof, in enhancing students' climate change knowledge. Consistent with the broader climate change research on fear appeals (O'Neill & Nicholson-Cole, 2009; Spence & Pidgeon, 2009), the authors' prior exploratory research (Carman et al., 2017), and our second study reported on here, raising students' perception of the threats posed by climate change does not appear to increase their learning about this topic. Instead, our study suggests that supporting students' ability to act on climate change has not just emotional and behavioral benefits (Anderson, 2012; Ojala, 2012; Stevenson & Peterson, 2015), but can also support their learning of foundational knowledge about climate change. Educators and others are, therefore, encouraged not to attempt to motivate students to learn about climate change by frightening them about predicted impacts. Instead, we urge educators to focus on how climate change may affect students' lives and empower them to help to respond to this challenge. Our research shows that there are relatively strong linkages between students' perceptions of the importance of climate change, self-efficacy, and their knowledge about climate change. Moreover, it shows that even short educational interventions can affect these two attitudes.
Lastly, to the best of our knowledge, the two studies reported here are the first to tease out perceived topic importance from topic interest in the context of climate change education. As a result, we were able to show that the former plays a far greater role in students' knowledge and learning than the latter. Perceived topic importance has been theorized to be a component of interest (Schiefele et al., 1993; Schiefele, 1991), but its role in science learning has not been tested separately from topic interest, including within the context of climate change education research. Moreover, we found that perceived topic importance can increase as a result of a short intervention, and is positively associated with situational interest, meaning that teachers can support the development of perceived topic importance relatively easily in the classroom. Given the novelty of this finding, more quantitative research is needed to replicate results that students' perceived importance of climate change can predict increases in their knowledge about the topic. Qualitative research is also needed to investigate educational methods to help students develop an appreciation for the importance of climate change in their lives in ways that both empower them and support their learning about climate change.
On the basis of our studies' findings, we offer several takeaways for science teachers. First, short educational interventions, like those currently implemented to teach students about climate change, can be effective in increasing students' foundational knowledge about climate change. Given the features of the unit enacted as part of our two studies, using educational techniques such as scientific modeling (Lombardi et al., 2018; Pruneau et al., 2010) and focusing on local impacts (Anderson, 2012; Busch & Osborne, 2014; Corner et al., 2015) to enhance students' knowledge about climate change may be beneficial practices. To improve students' situational interest in the unit, additional practices such as incorporating gaming and/or a focus on local communities (Nussbaum et al., 2016) appear promising. Moreover, our findings that behavioral self-efficacy and perceived topic importance play important roles in both interest and knowledge development suggest that techniques to enhance both of these factors may be beneficial. Relevant activities may include engaging students in deliberative discussions about how climate change might affect them and what they can do about it (Busch & Osborne, 2014; Monroe et al., 2019) and working on local projects to address climate change (Monroe et al., 2019).
Engaging students in the topic of climate change is a challenge that is different from engaging students in other science topics. Unlike many science topics that can be made enjoyable or fun to learn about (Ainley & Hidi, 2014; Jack & Lin, 2014), climate change tends to be stressful and upsetting for students (Busch & Osborne, 2014; Monroe et al., 2019). It is, therefore, important that climate change be taught in a way that that acknowledges its seriousness while also empowering students to act on it (Anderson, 2012; Ojala, 2012; Stevenson & Peterson, 2015). Our study shows that such an approach may not only have mental health benefits but contributes to students' learning about climate change. As climate change is increasingly incorporated into science curricula (NGSS Lead States, 2013) and students are more likely to recognize its influence on their lives (IPCC, 2014, 2018), having foundational knowledge about climate change will only grow in importance. To achieve this goal, it is, therefore, critically important that climate change educators both engage and empower students on this societal challenge, one that will profoundly impact the rest of their lives.
ACKNOWLEDGMENTS
We would like to thank Travis Hlavaty, Genevieve Leet, Janet Kahan, Annemarie McDonald, Benjamin Morse, Melissa Plegue, and Molly Watters for their various valuable contributions to this study. This study was supported by the USDA National Institute of Food and Agriculture McIntire-Stennis project (2013-32100-06099) and (2015-32100-06099); and the National Science Foundation (DEB 125664).
CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.
APPENDIX A: QUESTIONNAIRE ITEMS
See Table A1
| Factor name and items included | Question source(s) | Response options | Included in Study 1 | Included in Study 2 |
|---|---|---|---|---|
| Science Interest | X | |||
| I sometimes think about becoming a scientist when I grow up. | Science Aspirations and Career Choice: Age 10–14 (ASPIRES) longitudinal study (Archer et al., 2013; DeWitt et al., 2013); Previous pilot study (Carman et al., 2017) | 1 = strongly disagree, 2 = somewhat disagree, 3 = neutral, 4 = somewhat agree, 5 = strongly agree | X | |
| I would like science to be a part of my job one day. | X | |||
| I plan to take more science classes in the future. | X | |||
| Jobs in science are extremely interesting to me. | X | |||
| My friends and I discuss science-related topics. | Science Interest Survey (Lamb et al., 2012) | X | ||
| Perceived Topic Importance | X | X | ||
| Climate change is a very important issue to me. | PISA 2006 “Personal Value of Science” and “General Value of Science” scales (modified; OECD, 2007) | 1 = strongly disagree, 2 = somewhat disagree, 3 = neutral, 4 = somewhat agree, 5 = strongly agree | X | X |
| Trees and forests are very important to me. | X | |||
| Climate change will be an important issue in the future. | X | X | ||
| I think it's important to know how climate change impacts forests. | X | X | ||
| Trees and forests are valuable to society. | X | |||
| Climate change will impact forests in ways that affect all people. | X | |||
| Desire to Learn More | X | X | ||
| Base Text: I would like to learn more about … | Instructional Episode Questionnaire (modified; Swarat et al., 2012); Previous pilot study | 1 = strongly disagree, 2 = somewhat disagree, 3 = neutral, 4 = somewhat agree, 5 = strongly agree | ||
| … Trees (in general) | X | |||
| … Forests (in general) | X | X | ||
| … Climate change (in general) | X | X | ||
| … How scientists study climate change | X | X | ||
| … How trees are adapted to different climates | X | X | ||
| … How climate change affects forests | X | X | ||
| … How climate change affects trees | X | |||
| Topic Interest | X | X | ||
| Base text: This sounds interesting … | Instructional Episode Questionnaire (modified; Swarat et al., 2012); Previous pilot study | 1 = strongly disagree, 2 = somewhat disagree, 3 = neutral, 4 = somewhat agree, 5 = strongly agree | ||
| … Trees (in general) | X | |||
| … Forests (in general) | X | X | ||
| … Climate change (in general) | X | X | ||
| … How scientists study climate change | X | X | ||
| … How trees are adapted to different climates | X | X | ||
| … How climate change affects forests | X | X | ||
| … How climate change affects trees | X | |||
| Situational Interest: Hands-On Activities | X | |||
| Base text: This sounds interesting … | Instructional Episode Questionnaire (modified; Swarat et al., 2012); Previous pilot study | 1 = strongly disagree, 2 = somewhat disagree, 3 = neutral, 4 = somewhat agree, 5 = strongly agree | ||
| … Working with charts and graphs | X | |||
| … Working with real-life tree samples | X | |||
| … Taking scientific measurements | X | |||
| Factor: Situational Interest: Unita | X | |||
| Base text: Learning about Michigan forests and how they are impacted by climate change was … | Study Interest Questionnaire (Schiefele et al., 1993) | |||
| Boring … Exciting | 1 = Boring, 5 = Exciting | X | X | |
| Worthless … Valuable | 1 = Worthless, 5 = Valuable | X | X | |
| Dull … Interesting | 1 = Dull, 5 = Interesting | X | X | |
| Situational Interest: Field Tripa | X | |||
| I liked the field trip. | Best practices for outdoor learning (Rickinson et al., 2004) | 1 = strongly disagree, 2 = somewhat disagree, 3 = neutral, 4 = somewhat agree, 5 = strongly agree | X | |
| It was fun to take scientific measurements. | X | |||
| The field trip was fun. | X | |||
| Taking scientific measurements was interesting. | X | |||
| I learned a lot during the field trip. | Science Outdoor Learning Environment Inventory (SOLEI) (modified; Orion et al., 1997) | X | ||
| What we did during field trip helped me understand what we learned in class. | X | |||
| Self-Efficacy | X | |||
| By working with others, I can address climate change. | Zint et al. (2002, 2014) locus of control scale | 1 = strongly disagree, 2 = somewhat disagree, 3 = neutral, 4 = somewhat agree, 5 = strongly agree | X | |
| There are things I can do that will address climate change. | X | |||
| If everyone does their part, we can address climate change. | X | |||
| Perception of Climate Change Threat | X | |||
| Ann Arbor is already being affected by climate change. | Climate Change in the American Mind (modified; Leiserowitz et al., 2019); Previous pilot study | 1 = strongly disagree, 2 = somewhat disagree, 3 = neutral, 4 = somewhat agree, 5 = strongly agree | X | |
| Climate change already has an impact on people like me. | X | |||
| Climate change already has an impact on people different from me. | X | |||
| Michigan is already experiencing the effects of climate change. | X | |||
| Climate change already poses a threat to forests. | X | |||
| Ecosystems are already being affected by climate change. | X | |||
| Climate change already impacts forests in ways that affect people all over the world. | X | |||
| Career Interest | X | |||
| Base Text 1: Jobs in … are very interesting to me. | Relevance of Science (ROSE) Project (modified; Sjøberg & Schreiner, 2010) | 1 = strongly disagree, 2 = somewhat disagree, 3 = neutral, 4 = somewhat agree, 5 = strongly agree | X | |
| … science (generally) … | ||||
| … climate science … | X | |||
| … forest ecology … | X | |||
| Base Text 2: Someday I would like to become a … | ||||
| … scientist | X | |||
| … climate scientist | X | |||
| … forest ecologist | X |
- a Items that were asked only to treatment group students in the posttest.
Sample multiple-choice questions to measure students' knowledge about climate change including its impacts on forests
Note: The following questions were used in both studies. Questions were shown to students one at a time.
- A.
An oak tree produces acorns, which are carried away and planted in the ground by squirrels
- B.
The leaves of a black cherry tree are an important food source for a variety of moth and butterfly caterpillars that return as adults to pollinate flowers
- C.
Dead trees provide nest sites for cavity-nesting birds such as eastern screech owls and black-capped chickadees
- D.
The needles of an evergreen tree in northern Michigan have a thick, waxy coating that stops water loss
- E.
I don't know
- A.
The average climate
- B.
Changing temperature based on Earth's distance from the sun
- C.
The state of the atmosphere at a specific time and place
- D.
The state of the atmosphere averaged over a long period of time
- E.
I don't know
- A.
An increase in the number of sunspots is causing the sun to get warmer
- B.
The hole in the ozone layer is letting in more radiation
- C.
Air pollution, like smog
- D.
Too much carbon dioxide in the atmosphere from burning fossil fuels
- E.
I don't know
- A.
Tree growth increases
- B.
Tree growth decreases
- C.
Tree growth stays the same
- D.
There is no relationship between temperature and tree growth
- E.
I don't know
- A.
Black locust
- B.
Northern red oak
- C.
White pine
- D.
White spruce
- E.
I don't know
APPENDIX B: ADDITIONAL INFORMATION ABOUT THE UNIT DEVELOPMENT
The unit used in this study, Climate Change and Michigan Forests, was developed by a team of middle school science curriculum developers, University of Michigan faculty members and students, and middle school science teachers. The pedagogical approaches that influenced the unit's design and development included backward design (Graff, 2011; Jones et al., 2009; Prideaux, 2003) and the 5E learning cycle (Bybee, 2014; Karpudewan et al., 2015; Songer, 2006). Student learning objectives were determined by the design team based on middle school science teachers' needs and NGSS recommendations for middle school life science curricula. An overview of the final learning objectives and activities can be found in Table B1.
| Lesson Name | Description | Learning Objectives (Students will be able to …) |
|---|---|---|
| Lesson 1: Get in Touch with Trees! | Students brainstorm and create a list of why forest ecosystems are important to humans and the environment. They take notes on a PowerPoint presentation discussing factors that affect plant growth, and characteristics of deciduous and coniferous trees. |
|
| Lesson 2: Connections to Climate Change | Students watch and take notes on a suite of videos produced by the National Science Foundation on climate change and its impacts on humans and the environment. |
|
| Lesson 3: Down to the Core! | Students learn how scientists study tree growth. They count and measure tree growth rings using enlarged photos of tree cores, and predict how temperature and precipitation influence tree growth. |
|
| Lesson 4: Scientific Modeling | Students describe characteristics of scientific models and create and interpret a scatterplot graph to describe the relationship between the tree growth data they collected and precipitation and temperature data. |
|
| Lesson 5: Making Sense of Data! | Students share and interpret their graphs to discuss and make predictions about the relationship between temperature and tree growth. They interpret scientific models showing the predicted growth of six tree species in a future greenhouse gas emissions scenario. |
|
| Lesson 6: Climate and Plant Growth | Students review the climatic factors that influence plant growth. They read and interpret climographs to characterize the climates of multiple biomes, and discuss why plant growing seasons differ by region. |
|
| Lesson 7: Regional Impacts and Predictions | Students discuss how climate change impacts weather, ecosystems, and human economic activity within five biomes. Students collect, interpret, and organize information to make a Regional Prediction and Explanation based on supporting evidence. |
|
| Lesson 8: Student Actions | Students define climate change mitigation and adaptation and discuss examples of each. They work in groups to collect information on how climate change impacts biomes, and actions they can take to mitigate and adapt to climate change. |
|
| Lesson 9: Student Conference | Students work in groups to collect information on how climate change impacts biomes, and organize their findings into an informational poster. They present their posters to the rest of the class and grade each others' work using a peer evaluation rubric. |
|
| Field Trip! | Students visit a local forest and learn about different data collection methods forest ecologists use. They take a tree core sample and collect other tree growth and forest ecology data. |
|
The development and implementation of this unit took place over three school years (2012–2015), and lesson development included contributions from all team members with the goal of strengthening the local school district's seventh-grade math and life science curricula. From the beginning of the development process, three of the authors met with the district's science curriculum development lead, and teachers to identify needs the proposed unit could help fill. Because the State was in the process of adopting the NGSS at the time of the unit's development, one of the team's main goals was to align the new unit with these standards from the very beginning. Teachers' input proved essential throughout and informed all aspects of the unit, including the technology developed for the modeling exercises and determining the timing for teaching the unit (i.e., after they had had time to address foundational ecological and mathematical concepts).
An initial draft version of the unit was piloted during an earlier school year (2012–2013) in five seventh-grade classes (Carman et al., 2017), which included teachers who had informed its development. On the basis of these teachers' feedback, the design team made significant changes, including expanding the unit from 7 to 10 days of lessons. The revised pilot version was used in Study 1 (2013–2014) by some of the same, as well as additional, teachers (due to changes in teacher assignments and turnover). The teachers who participated in Study 1 attended a 1-day professional development session before implementing the pilot unit. A teacher log was used to track information about how the unit was implemented, and one of the co-authors also conducted observations of at least one class with all participating teachers. Immediately after completion of the unit, two of the co-authors completed semistructured interviews with 30 students (2–4 students per class), and each treatment group teacher participated in an interview with the research staff. The purpose of these observations and interviews was to guide potential improvements to the clarity and usability of the unit content and instructions. Information gleaned from these formative evaluation methods as well as multiple focus groups involving additional teachers were used to inform changes to the final version of the unit. That unit was used in Study 2 (2014–2015). Teachers again participated in a professional development session, which was shortened to a half-day because many teachers had already taught the unit for 2 years. Both years of professional development workshops were developed in collaboration with teachers, and the second one was not only codeveloped but also coled by one of the teachers. New teachers were mentored by teachers in the same school who had taught it in the past, and all teachers were supported by one of the co-authors throughout their implementation of the unit.
Since the completion of the studies reported on here, the unit has been adopted by the local school district's seventh-grade science curriculum, in great part due to the support of the teachers who helped to develop and improve it and the studies' findings that it improved students' understanding of climate change science. The unit has also been vetted by the National Science Teaching Association (NSTA) curators based on the EQuIP rubric, and as result of a favorable review, has been added to the organization's NGSS resource page (https://ngss.nsta.org/Resource.aspx?ResourceID=451).
