Faculty of Social and Behavioral Sciences
Graduate School of Child Development and Education
Running head: MENTOR SUPPORT: A SUPPORTIVE SCIENCE CONTEXT?
Searching for a Supportive Science Context in Primary Schools:
A Mentor Support Intervention
Research Master Child Development and Education Research Master Thesis
Niels J. de Ruig (10002531)
Supervisors: dr. E. H. de Bree and dr. M. Zee
Reviewers: dr. H. M. Y. Koomen and dr. F. C. Jellesma November 6, 2016
2 Preface
Working as a primary school teacher myself, I feel and experience every day how challenging it is to keep students motivated for science education. That is why I wanted to use what I’ve learned at the Research Master Child Development and Education (RESMA) to study what could contribute to students’ science motivation. Mentor support showed to be promising and my director, Hans Gigengack, and colleagues were very open to testing this with an
intervention study. I am really thankful to them that they provided me with this opportunity, with the aim of improving science education at our school. Dr. Francine Jellesma and Prof. dr. Peter de Jong guided me in the early stages of this study. We had discussions about the
research question and on how to compile the questionnaire. They also provided me with useful literature and tips for the data collection process. The input of dr. Debora Roorda during these early stages and for the improvement of the questionnaire were very valuable. I want to emphasize that I took the initiative in every stage of this process. For example, I designed the intervention myself, compiled the questionnaire, contacted the school and recruited the participants. This was a valuable learning process as a first years RESMA student. Bennie Mooren and Tom van Eijck helped me to collect the data, after which I conducted all the statistical analyses myself and wrote a first draft of the results section. Dr. Elise de Bree stepped into the process once the first analyses were conducted. She did not only provided me with really supportive feedback, but also helped me with minor details on APA and the analyses. Furthermore, she cheered me up in difficult times during the research process, which makes me glad that I got the chance to work with her. Elise also introduced me to dr. Marjolein Zee, who joined the process after the first draft of the thesis excluding the discussion was completed. With all her knowledge, Marjolein helped to get this study to the next level. She provided me with very useful feedback on content of the introduction which is mainly based on the Self-Determination theory. She also helped me to get the focus of the study straight and double checked the statistical assumptions and analyses. In addition, she explained to me how to conduct a confirmatory factor analysis in Mplus and provided me with the outcomes of this analysis. I would also like to thank the mentors who participated voluntarily in this study. They really made the project. Finally, I would like to thank Çisem Gürel, my Turkish friend, for providing me with helpful feedback. To conclude, I want to address that I am proud that so many people were involved in this study. It showed to me how powerful the combination is of working in primary education and conducting research on the improvement on education at the same time. I hope this study inspires educators and
3 Abstract
This study explored the effectiveness of a mentor support intervention for satisfying basic psychological needs (autonomy, competence, and relatedness, Self-Determination Theory) of third and fourth grade students. Using a pretest-posttest experimental design, 81 students in an intervention group and 71 students in a control group were asked to fill in a questionnaire about their needs satisfaction. Teachers of both groups were asked to fill in a questionnaire about their students’ scientific attitude (motivation, initiative and self-regulation, social attitude, and creativity and innovativeness) and research skills (wondering, collecting answers, processing data, drawing conclusions, and presenting findings). Results indicated that neither the students from the intervention group or those in the control group, reported an increase in their basic psychological needs. However, teachers with students in the mentor support group reported an increase in motivation towards science education and higher initiative, self-regulation and social attitude in comparison to the control group. Teachers did not note that students’ research skills increased in comparison to the control group, except for the skill ‘drawing conclusions’. Important conclusion of this study is that a discrepancy was found with regard to what students reported about their needs satisfaction and what teachers reported about motivation towards science. Possible explanations for these findings and limitations of the study are discussed.
Keywords: mentor support intervention, basic psychological needs, scientific attitude, research skills
4 Searching for a Supportive Science Context in Primary Schools:
A Mentor Support Intervention
Recent policy developments around the globe have increasingly emphasized the importance of structural science education in primary schools (Fensham, 2008; Kali, Linn, & Roseman, 2008; Lavigne, Vallerand, & Miquelon, 2007; Osborne, Simon, & Collins, 2003). This increased emphasis is based on the prediction that beta science related jobs will become critical to 21st century economic growth of the knowledge society. It is also based on research demonstrating a decline of general interest in beta sciences among Western students
(Griethuijsen et al., 2014; Krapp & Prenzel, 2011; Osborne et al., 2003). This is for example reflected in the decreasing number of graduates in all European countries who choose a beta related PhD (Fensham, 2008; Osborne & Dillon, 2008). The decision for such a beta career is typically made by students between the age of 10 and 14. Research has identified that this decision is largely influenced by primary school teachers (Griethuijsen et al., 2014; Osborne & Dillon, 2008; Osborne et al., 2003). These teachers may help students feel motivated for science education at an early age, by creating a supportive science context (Osborne et al., 2003). Primary school teachers can create a supportive science context by including three dimensions in their teaching behavior (Skinner & Belmont, 1993). These three dimensions are: involvement (the interpersonal relationship of students with peers and the teacher), structure (clear information about how to realize desired outcomes) and autonomy support (freedom of students to express own behavior; Deci & Ryan, 2002; Skinner & Belmont, 1993). Teachers can express these dimensions by, amongst others, communicating about expectations, providing students with freedom of choice, showing affection, or responding with positive and constructive feedback (Skinner & Belmont, 1993). If students perceive that their teacher has created a supportive context, positive outcomes such as motivation towards science and the decision to choose a beta career may turn out to be positive (Ryan & Deci, 2000). The present study’s aim is therefore to explore if an intervention targeted at primary school teachers is able to help them create a supportive science context.
Underlying this aim is the well-developed Self-Determination Theory (SDT; Deci & Ryan, 2002). It is believed that this theory can help to assess a supportive science context by
measuring students’ satisfaction or dissatisfaction of three basic psychological needs (Deci & Ryan, 2002; Lavigne et al., 2007; Vallerand & Ratelle, 2002). These three basic needs are competence, autonomy, and relatedness (Deci & Ryan, 2002). Competence is students’ feelings of being successful in expressing own capacities by overcoming challenges created by their teachers (Schunk, Pintrich & Meece, 2010). Autonomy is students’ needs to act from
5 own interests and values (Deci & Ryan, 2002). Finally, relatedness is students’ needs for connection with and belongingness to their teachers (Ryan & Deci, 2000; 2002). Based on this SDT framework, motivation towards science is assumed to be present if the basic psychological needs of students are satisfied. Future science outcomes such as persistence in beta science education and future intensions to become a beta scientist are then expected to be positive (Ryan & Deci, 2000; Vallerand, Fortier & Guay, 1997; Vallerand & Ratelle, 2002). However, if one or more of these needs is not satisfied by a teacher, negative outcomes for students in science education are expected (Deci & Ryan, 2002). For example, empirical research shows that students reported lower academic achievement (Boggiano, Flink, Shields, Seelbach & Barret, 1993; Flink, Boggiano, & Barret, 1990), lower positive emotionality (Patrick, Skinner, & Connell, 1993), lower creativity (Koestner, Ryan, Bernieri, & Holt, 1984) and lower motivation (Deci & Ryan, 2002) in comparison to students with a supportive context. These outcomes show how important it is for teachers to create a supportive science context that is able to satisfy all students basic psychological needs.
Self-Determination Theory and Science Education
Empirical evidence supporting the Self-Determination Theory for science education has also been found (Lavigne et al., 2007). Results show that high school students with satisfied basic psychological needs had intrinsic motivation towards science education. Intrinsic motivation is the motivation for a task or assignment because of its content, which is challenging and relates to the interests of the student (Bender, 2010). These students were also found to have significantly higher intentions to become a beta scientist than students who were extrinsically motivated. Extrinsic motivation is student’s motivation to work on a task only because he believes that the outcomes are desirable to avoid problems or negative results (Schunk, Pintrich, & Meece, 2009). Importantly, students for whom motivation towards science was found to be present also reported that their teachers were significantly more autonomy supportive than teachers of students who were extrinsically motivated (Lavigne et al., 2007). In addition to these results, students reported that their needs (competence and autonomy) were satisfied by a teacher who showed to be autonomy supportive. This also resulted in motivation towards science education. Furthermore, direct effects of autonomy and
competence on intentions to become a beta scientist were found. These findings illustrate the importance of a teacher in satisfying basic psychological needs, so that students remain motivated for science education and develop intentions to become a beta scientist.
6 Although empirical results of Lavigne and colleagues (2007) showed the importance of teachers for satisfying basic needs in general, their findings for gender differed. Girls perceived lower competence and had lower intentions to pursue a beta scientific career than boys (Lavigne et al., 2007). In addition, girls did not have lower motivation towards science education. This discrepancy between boys and girls could have been influenced by
stereotypes of beta science that demonstrate qualities that are more related and valued by boys than girls (Cheryan, Master, & Meltzoff, 2015; Lavigne et al., 2007; Osborne et al., 2003). Furthermore, research indicate that boys generally had higher interests in scientific beta topics (Anderson, 2006; Chang, Yeung, & Cheng, 2009). In addition, boys were more interested in becoming a beta scientist than girls (Griethuijsen et al., 2014). These results indicate that boys and girls may perceive supportive science contexts that are created by their teachers in
different ways. This might have been influenced by factors not directly attributable to teachers’ behaviors.
Despite the difference in findings for gender, it seems difficult for teachers to satisfy all students’ basic psychological needs when they are offering structural science education. This claim is in line with studies showing that science motivation uniformly declines throughout primary school (Martin, Mullis, Foy, & Stanco, 2012; Osborne et al., 2003; Vedder-Weis & Fortus 2011). According to SDT, the reason for this decline might have been that the basic psychological needs of the students were not (completely) satisfied by their teachers. A good example is the study of Murphy and Beggs (2003). They conducted a large-scale (N = 1000) survey study with 8 to 11 year old primary school children from 44 schools across Northern Ireland. Their results showed that 10 and 11-year-olds have significantly lower motivation towards science education than 8 and 9 year old students. Students described their science education as: absence of hands-on experiments or opportunities to discover new phenomena and/or facts, emphasis on test preparation and inappropriate curriculum content not related to their day-to-day experiences. In addition, teachers felt that they had a lack of scientific background knowledge which resulted in low confidence to teach science topics (Fensham, 2008; Murphy & Beggs, 2005). These results indicate that it may indeed be difficult for teachers to provide students with structural science education in primary schools and to satisfy their basic psychological needs at the same time (Lavigne et al., 2007). Therefore,
interventions that assist primary school teachers to create a supportive science context are assumed to be highly relevant.
7 Mentor support to create a supportive context
There are, to the best of the author’s knowledge, three similar studies that show promising results in assisting primary school teachers in science education by using mentor support (Forbes & McCloughan, 2010; Howitt, Rennie, Heard, & Yuncken, 2009; Wilson,
Krakowsky, & Herget, 2010). The MyScience project of Forbes and McCloughan (2010) has been repeatedly studied and is therefore taken as starting point for the intervention in the present study (Forbes & Skamp, 2013; 2014). MyScience started at two primary schools in Australia for children between 7 and 11 years old. Five elements underpinned the MyScience project (Forbes & McCloughan, 2010): (1) collaborative professional learning for primary teachers: interaction between teachers, administrators and mentors was facilitated within the school, to support teachers in transferring to a new science curriculum. In addition,
collaboration between schools was part of this first element as well. (2) clear achievement criteria: the achievement criteria were clear to students from the beginning. This helped them to know what was expected from them throughout the project. Most criteria were taken from the national standards. This also provided teachers with guidelines to teach and assess
students’ progress. (3) student support in their investigation through using an empirical cycle for students: teachers introduced research skills to students. After this introduction, students applied their learned skills together with support of a mentor. The cycle started with students’ own research questions based on a class theme. (4) student mentoring by scientist:
MyScience’s key characteristic is the use of mentors (volunteer scientists). These mentors interacted with teachers and helped them to develop scientific knowledge and understandings (Forbes & Skamp, 2014). In addition, mentors were available in the classroom or online to provide students with technical and scientific advice on their own research. For every six students, one mentor was available. The mentors were selected on science-based
qualifications or beta science related careers (Forbes & McCloughan, 2010). (5) celebrating achievement by sharing findings with the community: this last element increased the meaning of students’ research, because it gave teachers the opportunity to put emphasis on students’ achievements. Moreover, it involved the community (parents, nearby schools, scientists etc.) in the collaborative approach of MyScience.
As part of the evaluation of MyScience, semi-structured interviews were conducted with eight primary teachers from three different MyScience schools (Forbes & Skamp, 2014). The aim of this study was to obtain insight in the view of teachers on MyScience. The questions of the interview were directed towards attributes of a Community of Practice and the nature of science (see overview Forbes & Skamp, 2013). Results of interviews revealed, among other
8 things, that mentors were able to create bridges between school science and the outside world. Mentors also provided students with knowledge, information and enthusiasm about science. Teachers were, because of the help of mentors, able to spend more time on every group than if the teachers were alone. In addition, mentors directly or indirectly taught teachers different concepts related to the nature of science. For example, what constitutes a data set or
hypothesis. This seemed to give teachers more confidence in teaching science. In addition, teachers reported that their students showed an increase in the understanding and application of research skills (Forbes & Skamp, 2014). Teachers also described that their role had shifted from being a director to a facilitator. They provided support to students while they were investigating, worked as mediator between mentors and students, raised interest about science among students and maintained a physically and emotionally safe and open classroom
environment. Students were found to be highly motivated and willing to remain involved in the program. These findings of the MyScience project showed that teachers appeared to be satisfied with mentor support. In addition, teachers indicated that their students showed increased motivation for structural science education with the help of mentors.
Present study
The MyScience project will be taken as starting point for designing an intervention in the present study. In addition to Forbes and McCloughan (2010), the present study’s aim is to explore if a mentor support intervention that is based on MyScience is effective in helping teachers to satisfy their students’ competence, autonomy and relatedness. Satisfaction or dissatisfaction of these needs will tell if teachers were able to create supportive science contexts with the help of mentors. Findings would complement the results of Forbes and Skamp (2013; 2014), because, as seen, the paradox in science education appears to be that structural science education often does not go hand in hand with a supportive science context (Martin et al., 2012; Osborne et al., 2003; Vedder-Weis & Fortus 2011). The research
question central to this study is therefore: what is the effectiveness of a mentor support intervention in satisfying the basic psychological needs of third and fourth grade students? Third and fourth grade students are included, because they are not likely to have made the decision yet whether to pursue a beta career (Griethuijsen et al, 2014; Logan & Skamp, 2013; Osborne & Dillon, 2008; Osborne et al., 2003). Based on the results of Forbes and
McCloughan (2010) and Forbes and Skamp (2013; 2014), it is expected that the students who receive mentor support show significantly higher perceptions of competence, autonomy and relatedness than students without mentor support. Such findings would indicate that students
9 with mentor support experience a supportive science context from their teachers and have motivation towards science education.
Besides measuring students’ needs satisfaction, teachers’ perceptions of the intervention are also thought to be of significance. Especially because the intervention is important to the school’s science curriculum. Therefore, two extra concepts are included in the present study. The first concept is students’ scientific attitude. This attitude consist of aspects that
characterize scientific thinking of students (Osborne et al., 2003). These aspects can be divided into general attitudes towards ideas and information (i.e. curiosity and creativity), attitudes towards the evaluation of those ideas and information and willingness to use a scientific approach (Chang et al., 2009; Gauld & Hukins, 1980). Gauld and Hukins (1980) describe that these aspects helped students to understand the nature of scientific processes more easily. In addition, scientific attitudes were found to be vital for students to encounter daily life problems. Students with a scientific attitude solved these problems by evaluating information and ideas in a scientific manner. This resulted in more satisfactory solutions in comparison to students who did not master a scientific attitude (Guald and Hukins, 1980). Although these results are positive, scientific attitudes are complex and therefore difficult to measure (Osborne et al., 2003). To help primary schools measure students’ scientific
attitudes, Van Keulen and Slot (2015) described four dimensions underlying scientific attitudes of primary school students. These four dimensions were based on the indicators that are described in the Dutch national standards (van Graft, Klein Tank, & Beker, 2014). The dimensions are: (1) motivation: students work independently, show high involvement, ask questions and enjoy the activities (2) initiative and self-regulation: students are constantly looking for situations or opportunities to apply their knowledge and skills. In addition, these students do not need much help from teachers. (3) social attitude: students work together with other students. This requires skills such as: listening to other students, sharing thoughts and ideas, make use of strengths of group members and being respectful to others. (4) creativity and innovativeness: students have the ability to come up with new ideas, explanations or solutions for the problems they encounter. They can use this knowledge to find solutions to problems they will come across in the future. To conclude, these four dimensions together will be used in the present study to measure teachers’ perceptions of students’ scientific attitude.
The second concept is students’ research skills. Research skills are skills that scientists use to conduct research, for example: formulating research questions and hypotheses, setting up experiments and presenting findings (van Graft et al., 2014). Empirical research showed that
10 high-school chemistry students who learned research skills by an inquiry approach, were more motivated to pose scientific questions and also had more questions in comparison to the control group (Hofstein, Navon, Kipnis, & Mamlok-Naaman, 2004). In addition, the questions they asked were of a higher cognitive level and more relevant. Another empirical study
showed that undergraduates who mastered research skills, had higher research self-efficacy and therefore higher intentions to become a beta scientist (Adedokun, Bessenbacher, Parker, Kirkham, & Burgess, 2013). Although both studies were not conducted under primary school students, they show the importance of mastering research skills. That is why research skills are incorporated in the present study, based on five different phases of an empirical cycle for students. This cycle will also be central to the present study’s intervention (Van Keulen & Slot, 2015). The five phases of this cycle are: (1) wondering (i.e. asking questions, using foreknowledge, exploring a problem), (2) collecting answers (i.e. making a research
proposal/planning, conducting data, observing), (3) processing data (i.e. organizing, editing and analyzing collected data), (4) drawing conclusions (i.e. critical discussion, generating follow-up questions) and (5) presenting findings to the class (i.e. sharing ideas). Based on the results of Forbes and McCloughan (2010) and Forbes and Skamp (2013; 2014), it is expected that teachers will observe that students with mentor support show increased research skills in comparison to students without mentor support. The same increase in comparison to the control group is expected for scientific attitudes of students with mentor support.
Method Participants
One hundred and fifty-five students divided over six classes from a Montessori primary school in Amsterdam, the Netherlands, participated in this study. Three classes were selected for the intervention group (N = 83, 38 boys) and three for the control group (N = 72, 36 boys), based on an equal distribution of boys and girls for each group. Students’ ages ranged from 8 to 11 years old (M = 8.99 years, SD =.71 years). All students had were Dutch and spoke Dutch fluently, although six students were born in foreign Western countries. One of these students was part of the intervention group and five of the control group. Table 1 displays information on the teachers participating is this study. One teacher started working at the school in the year of the intervention, but she already had five years’ experience at another school in Amsterdam.
11 Table 1
Information on Participating Teachers
Group Gender Age Grade Number of
students in class Experience at current school (years) Intervention group Female 64 3/4 28 43 Female 59 3/4 28 4 Female 36 3/4 27 14
Control group Male 57 3/4 27 35
Female 37 3/4 28 1
Female 34 3 17 5
Note. One control group consisted of 17 third grade students because the school was expanding in the period of
the research. The next schoolyear this class consisted of a third and fourth grade of 27 students.
Intervention overview
An eight-week mentor support intervention was designed based on the MyScience study (Forbes & McCloughan, 2010) (see Appendix for an extensive overview of the intervention and how the five elements of MyScience were integrated). Central to the intervention was a collaboration between mentors, primary school teachers and students. Teachers were the didactic/pedagogic experts and mentors were the beta scientific experts. Both used their expertise to help students to investigate their own research question scientifically. In addition, mentors fed teachers with knowledge about the class theme and research skills and teachers fed mentors with knowledge about teaching. This mutual learning process of teachers and mentors took place by observing each other, but also through discussion after class or correspondence by e-mail. The primary aim of the intervention was to help teachers in creating a supportive science context that satisfied all their students’ basic psychological needs. The participation of mentors, teachers and teachers in the control group will be further elaborated upon in the following section.
Mentors Eight mentors participated in the three classes of the intervention group. Seven
mentors had work-related experience with the beta sciences, including an employee of the Port of Amsterdam, a PhD graduate in physics, a science-method developer for primary schools, and a neuro-scientist who taught science at a Teacher Training Program. One mentor was a musician of the Dutch Metropole Orchestra. He was still accepted as mentor because of his talent to think creatively, his degree in higher education and his interest in the project. All
12 mentors had an introductory meeting with the first author, in which their background and role as mentor was discussed. The first author paired mentors with diverse backgrounds in beta related work. Each class worked with at least two mentors. These mentors were at least six weeks present in the classroom for around one-and-half hour.
Teachers A prepatory meeting was held by the first author with the teachers from the
intervention group, in which the topic ‘magnetism’ was chosen out of the topics provided by the Dutch national standards of science education (Beker et al., 2009). Furthermore, teachers received instruction on the empirical cycle for students (Van Keulen & Slot, 2015, see
introduction for overview). This cycle was used throughout the intervention and instructed by the teachers to their students. In addition to the last step of the cycle (presenting findings to the class), parents from the school were invited to a scientific symposium organized by the students in their classrooms. Moreover, for every phase of the empirical cycle, a preparatory document was written by the first author to provide teachers with guidance. The preparation consisted of information on every phase of the empirical cycle and practical examples for teaching. The preparation was discussed during weekly meetings with the first author, in which the progression of the project was also evaluated.
Control group The control group followed the regular method for science education at
school. This method was primarily developed by teachers with a beta science background, who were not part of the control group. These teachers made several science containers with experiments relevant to a certain topic, together with a manual for their colleagues. Especially for the purpose of this study, a box of experiments and a manual of eight lessons about
magnetism was compiled by one science teacher (not related to this study), who followed a beta science track at the Teacher Training Program. The eight lessons consisted of doing experiments together with the teacher, such as making an electromagnet with a battery. The teachers in the control group had one preparatory meeting with the first author to discuss content of the science container and the research instruments.
Procedure
Students in the control and intervention group completed a questionnaire before and after the intervention in class under supervision of a trained assistant. The assistant received a 20-minute briefing, during which a standardized instruction for the students was explained. This standardized instruction informed students about mainly three topics: (1) their answers would not be shared and they could stop the questionnaire at any time, (2) the purpose of the
13 science education could be improved, (3) explanation about how to fill in the questionnaire. Written consent was obtained from each primary caregiver of all participating students, together with the assent from each student in advance for completing the questionnaire. Students were thanked for their participation afterwards. This research procedure was evaluated and approved by the Ethics Board of the University of Amsterdam (case number: 2015-CDE-4149 A).
All teachers within the control and intervention group were asked to digitally fill in the Skills Monitor Inquiry & Design (SMID, Van Keulen & Slot, 2015) for every individual student before and after the intervention. Written consent was obtained from every teacher. It was explained by the researcher that the purpose of the SMID was to provide feedback to the teachers themselves and to collect data on the research skills and scientific attitude of every student. Teachers were asked to save the SMID in the secure online school environment ‘ParnasSys’. To prevent from any bias and secure anonymity, results from the SMID and the students’ questionnaire were paired and all names were replaced by a code.
Instruments
Two different instruments were used to explore the effectiveness of the mentor support intervention (see Figure 1 for a visualization of the research design). The first instrument is a compiled questionnaire to measure students’ perceptions of their basic psychological needs. This questionnaire was also slightly adapted for science education. Examples of how the questionnaire was adapted will be given below. The second instrument is the Skills Monitor Inquiry & Design (SMID) to measure teachers’ perceptions of students’ scientific attitude and research skills. Both instruments will be further explained in the following section.
Basic Psychological Needs of Students
Perceived Autonomy Students were asked to fill in 4 items (perception of autonomy in the
academic domain) from the Perceived Autonomy in Life Domains Scale (PALDS-16) based on Blais, Vallerand and Lachance (1990). Examples of questions are: ‘At school, I feel as if I were in jail’ and ‘I feel obligated to go to school’. Examples of how these questions were adapted to science education are: ‘During the science lessons, I feel as if I were in jail’ and ‘I feel obligated to attend the science lessons’. The PALDS-16 was used in the study of Lavigne et al. (2007), who found a relatively moderate reliability (α = .74). All items are scored on a 5-point Likert scale, ranging from (1) “definitively not true” to (5) “definitively true”.
14
Perceived Competence The scale for perceived competence (4 items, perceptions of
academic competence) was based on Losier, Vallerand & Blais’s (1993) Perception of Competence in Life Domains Scale (PCLDS). These 4 items are reported to have a relatively high reliability (α = .84; Losier et al., 1993). Examples of questions are: ‘In general, I have difficulty doing my school work well’ and ‘I have developed very good abilities as a student’. Examples of how these questions were adapted to science education are: ‘In general, I have difficulty doing my science-work well’ and ‘I have developed very good science-abilities as a student’ All items are scored on a 5-point Likert scale, ranging from (1) “definitively not true” to (5) “definitively true”.
Perceived Relatedness The scale for perceived relatedness (8 items) was originally based
on the shortened version of the Teacher as Social Context Questionnaire (TASCQ, Belmont, Skinner, Wellborn & Connell, 1988; Connell & Wellborn, 1991). The TASCQ was reported in the study of Lietaert, Roorda, Laevers, Verschueren and De Fraine (2015) to have relatively high reliability (α = .86). Examples of questions are: ‘My teacher tells me what he expects
Teachers: SMID Students: questionnaire Posttest Teachers: SMID Students: questionnaire Weekly meetings to evaluate progress and discussion on preparation for the next week provided by the first author.
Mentor support: students conduct own research according to empirical cycle (wondering, collecting answers, processing data, drawing conclusions and presenting findings) together with mentors. Result are shared with community. Magnetism Eight predetermined science lessons: Science containers with experiments. Intervention group Control group Prepatory meeting Teachers: SMID Students: questionnaire Teachers: SMID Students: questionnaire Pretest Selection of topic, instruction on empirical cycle and the SMID. Instruction on content of Science containers and the SMID.
Figure 1. Visualization of Research Design.
Note. SMID refers to the Skills Monitor Inquiry & Design which measures teachers’ perceptions of
15 from me in school’ and ‘My teacher tests whether I am done before he continues with
something new’. Examples of how these questions were adapted to science education are: ‘My teacher tells me what he expects from me during the science lesson’ and ‘My teacher tests whether I am done during the science lesson, before he continues with something new’. All items were scored on a 5-point Likert scale, ranging from (1) “definitively not true” to (5) “definitively true”.
Psychometric properties of the Basic Psychological Needs
A confirmatory factor analysis (CFA) using maximum likelihood estimation was used to test the factorial validity of the basic psychological needs instruments. Fit indices of the baseline model indicate poor model fit, χ2(101) = 211.99, p < .0001, RMSEA = .085 (90% confidence
interval [CI] [0.069, 0.101), CFI = .74, SRMR = .090. Not only does the CFI shows to be below the conventional threshold of .90 for satisfactory fit, the RMSEA and SRMR also indicate poor goodness of fit according to a cutoff value of .08 (Bentler, 1992; Browne & Cudeck, 1993; Hu & Bentler, 1999; Kline, 2011). Modification indices suggested significant improvement of model fit, by adding a correlation between the residuals of relatedness item 3 (‘My teacher paid a lot of time on me during the project’) and item 4 (‘My teacher really cared about me’). These two questions share the underlying concept of the attention of the teacher. In addition, model improvement was reached after the deletion of one question of perceived competence (‘In general, I am having difficulty to get my science-work done’), one question of perceived autonomy (‘I do not feel forced to attend the science-lessons’) and one question of perceived relatedness (‘My teacher just did not understand me during the
project’). Fit indices of the final model show to have satisfactory model fit, χ2(61) = 92.52, p < .0001, RMSEA = .058 (90% confidence interval [CI] [0.320, 0.081), CFI = .91, SRMR = .060. The factor loadings for the basic psychological needs were adequate in the final model, ranging from .56 to .72 for perceived autonomy, from .37 to .70 for perceived relatedness and from .27 to .87 for perceived competence. The latter showed one inadequate factor loading (.27; ‘I do not think I work very quickly on my science-work’), but deletion of this question did not improve model fit. In addition, the perceived relatedness, autonomy and competence subscales have a moderate reliability after deletion of the aforementioned questions on the pre- and posttest, except for the relatively low reliability of competence on the post-test (see Table 2).
16 Table 2
Cronbach’s α for the Basic Psychological Needs Scale on the Pre- and Post-test
Test Scales Cronbach’s α
Pretest Competence .63 Autonomy .66 Relatedness .71 Posttest Competence .56 Autonomy .72 Relatedness .63
Teachers’ perceptions of students’ scientific attitude and research skills
To study whether teachers perceive that mentor support promotes students’ scientific attitude and research skills, the Skills Monitor Inquiry & Design was used as pre- and posttest (Van Keulen & Slot, 2015). Teachers rated the performance of their students on a 7-point Likert scale, ranging from (1) “low performance” to (7) “excellent performance”. This was done for scientific attitudes (divided into subscales: motivation, initiative and self-regulation, social attitude, and creativity and innovativeness) and research skills based on the empirical cycle for students (divided into subscales: wondering, collecting answers, processing data, drawing conclusions and presenting findings). Every subscale consisted of one question, which resulted in a relatively high overall reliability for the pretest (α = .96) and posttest (α = .96). Examples of questions are: ‘rate a student’s ability to formulate a research question’ and ‘rate student’s motivation for science’.
Results Preliminary analyses
Little’s MCAR test was conducted and assumed missingness completely at random (χ2 =
318.78, df = 294, p = .153). Two cases were deleted because one student decided not to fill in the questionnaire and another student moved to another city after the intervention. This resulted in absence of data on the post-test measure. Remaining missingness was negligible and therefore one case was deleted list-wise (N = 152, with N = 81 in the intervention group and N = 71 in the control group). Mean scores were calculated separately for each basic psychological need, for the pre-test as well as the posttest. A score of one represents the lowest mean score and 5 the highest mean score. A number of z-values for skewness and
17 kurtosis of these mean variables appeared to be higher than the critical z-value of 3.29 (see Table 3; Field, 2009). In addition, a number of variables of teachers’ perceptions of Scientific Attitude and Research Skills showed the same pattern. The assumption of normally
distributed data is therefore supposed to be violated. However, it can be assumed that the ANCOVAs used in the following analyses are relatively robust against the violation of this assumption (Tabachnick & Fidell, 2014). No multivariate outliers were detected through the use of the Mahalanobis distance with p < .001.
Table 3
Z-values of Skewness and Kurtosis on the Pre- and Posttest for Basic Psychological Needs, Scientific Attitude and Research Skills.
Note. * skewed distribution compared to a critical z-value of 3.29.
Data Screening and Descriptive Statistics
Three repeated measures 2 (intervention phase, pre-posttest) x 2 (group, control/mentor support) ANCOVAs, with Age and Gender as covariates were conducted to assess the contribution of mentor support to any changes in the mean variables of Autonomy, Competence and Relatedness, controlled for any possible effect of Age and Gender (i.e. Anderson, 2006; Chang et al., 2009; Griethuijsen et al., 2014; Murphy & Beggs, 2003). Furthermore, homogeneity of regression slopes and homogeneity of variance were assumed to be satisfactory for all three repeated measures ANCOVAs. The assumption of linearity is
Variables Skewness Pretest Skewness Posttest Kurtosis Pretest Kurtosis Posttest Basic Psychological Needs
Autonomy Competence Relatedness Scientific Attitude
Motivation
Initiative and self- regulation Social attitude Creativity and innovativeness Research Skills Wondering Collecting answers Processing data Drawing conclusions Presenting findings -11.60* -2.47 -1.77 -3.56* -4.85* -6.09* -2.42 -3.58* -2.72 -1.89 -3.09 -2.83 -10.12* -.64 -1.43 -4.43* -3.69* -4.42* -4.33* -5.03* -4.35* -3.81* -4.48* -3.16 16.27* -.01 -.28 1.04 1.72 4.86* .96 -.28 -.01 -.36 .01 .91 10.67* -.08 -1.38 .95 .72 3.70* 4.13* 1.55 1.91 .85 .91 .25
18 assumed to be violated, based on visual inspection of scatterplots in which the mean variables of the Basic Psychological Needs and covariates are plotted against the two groups. However, because of a sample larger than N = 30, repeated ANCOVAs are assumed to be relatively robust against this violation (Tabachnick & Fidell, 2014). Descriptive statistics can be found in Table 4.
The second set of repeated measures 2 (intervention phase, pre-posttest) x 2 (group, control/mentor support) ANCOVAs, with Age and Gender as covariates is devoted to teachers’ perceptions of students’ scientific attitude (Motivation, Initiative and
Self-Regulation, Social Attitude, and Creativity and Innovativeness) and research skills (divided into Wondering, Collecting Answers, Processing Data, Drawing Conclusions and Presenting Findings). Mean scores for pre- and posttest are displayed in Table 5, in which 1 represents the lowest mean score and 7 the highest mean. These mean scores show to be relatively high (above 4) on the pre- and posttest. The same holds for the mean scores reported by students of their Basic Psychological Needs (above 3). The former might indicate that teachers are on average already satisfied with their students’ scientific attitude and research skills The latter shows that students might generally already had satisfied basic psychological needs before the intervention started. In comparison to the majority of primary schools in the Netherlands (Dutch Inspection of Education, 2005), the school was already providing their students with structural science education, which might explain these high mean values.
The assumption of homogeneity of regression slopes appeared to be satisfactory for all repeated measures ANCOVAs, except for the statistically significant interaction terms of Presenting Findings and Gender, F(1, 148) = 5.766, p = .018, and Motivation and Gender, F(1, 148) = 4.873, p = .029. These analyses were also conducted without the covariates, providing similar results. Therefore, it is still assumed that the test-statistics of the repeated measures ANCOVAs of Presenting Findings and Motivation can be trusted. The assumptions of linearity and homogeneity of variance were assumed to be satisfactory for all repeated measures ANCOVAs. Partial Eta Squared was used as an effect size, in which .01 or higher refers to a small effect, .06 or higher to a medium effect and .14 or higher to a large effect (Richardson, 2011).
19 Table 4
Descriptive Statistics Repeated Measures ANOVAs of The Basic Psychological Needs
Students’ Perceived Autonomy, Competence and Relatedness
Results of the three repeated measures ANCOVAs with Autonomy, Competence and Relatedness as outcome variables, indicate no significant interaction effect between mentor support and Autonomy after controlling for Age and Gender, Wilks’ Λ = .99, F(1, 148) = .10, p = .759, partial η2 = .001, and no significant main effect for Autonomy, Wilks’ Λ = .994, F(1,
148) = .85, p = .358, partial η2 = .006. In addition, no significant interaction effect was found
between mentor support and Competence after controlling for Age and Gender, Wilks’ Λ = 1.00, F(1, 148) = .064, p = .801, partial η2 = .000, with no main effect for Competence, Wilks’
Λ = .999, F(1, 148) = .12, p = .728, partial η2 = .001. Finally, there was no significant
interaction effect between mentor support and Relatedness after controlling for Age and Gender, Wilks’ Λ = .999, F(1, 148) = .08, p = .779, partial η2 = .001. No main effect for
Relatedness was found, Wilks’ Λ = 1.00, F(1, 148) = .48, p = .488, partial η2 = .003. In
summary, mentor support was not found to contribute significantly to Autonomy,
Competence and Relatedness reported by the students in comparison to the group of students without mentor support.
Variables Pretest
Mean SD
Posttest
Mean SD
Basic Psychological Needs Autonomy Intervention group Control group Competence Intervention group Control group Relatedness Intervention group Control group 4.56 4.56 3.72 3.68 3.75 3.53 .73 .76 .94 .91 .68 .69 4.45 4.51 3.53 3.47 3.62 3.45 .82 .83 .91 .81 .67 .61
20
Table 5
Descriptive Statistics and Results of Repeated Measures ANCOVAs for Teacher’s Perceptions of Research Skills and Scientific Attitude.
Note. *p < .05, **p < .01, ***p < .001. Variables Pretest Mean SD Posttest Mean SD Fpre/post Partial η2 Fgroup Partial η2 Fpre/post* Fgroup Partial η2 Scientific Attitude Motivation Intervention group Control group
Initiative and self-regulation Intervention group Control group Social attitude
Intervention group Control group
Creativity and innovativeness Intervention group Control group Research Skills Wondering Intervention group Control group Collecting answers Intervention group Control group Processing data Intervention group Control group Drawing conclusions Intervention group Control group Presenting findings Intervention group Control group 5.47 5.08 4.96 4.93 5.01 4.93 5.11 4.96 5.28 4.78 4.80 4.62 4.53 4.66 4.72 4.60 4.77 4.75 1.04 1.03 1.21 1.16 1.19 .98 1.01 .95 1.20 1.11 1.25 1.25 1.30 1.28 1.26 1.28 1.03 1.22 5.79 5.03 5.24 4.75 5.35 4.93 5.36 4.97 5.40 5.20 5.09 4.90 4.83 4.83 5.09 4.61 5.04 4.89 1.06 1.12 1.20 1.20 1.22 .95 1.00 .89 .97 1.14 1.04 1.24 1.18 1.18 1.11 1.36 1.12 1.19 .40 .07 2.15 .41 .20 .03 .59 .29 3.02 .00 .00 .01 .00 .00 .00 .00 .00 .02 14.46*** 1.88 2.27 3.32 4.87* .98 .21 2.22 .19 .09 .01 .02 .02 .03 .01 .00 .02 .00 7.78** 13.17*** 5.84* 3.51 7.15** .00 1.22 7.19** 1.78 .05 .08 .04 .02 .05 .00 .01 .05 .01
21 Teachers’ Perceptions of Scientific Attitude
The results of four repeated measures ANCOVAs with Motivation, Initiative and Self-Regulation, Social Attitude and Creativity and Innovativeness as outcome variables are displayed in Table 5. There was a statistically significant interaction effect between Motivation and Group, so the main effects cannot be interpreted unambiguously. The
interaction was probed by testing the effect of Wondering on Group. Sidak-adjusted p-values were used to control inflation of the Type I error rate for multiple pairwise comparisons. Pairwise comparison showed that the mean of Motivation for the control group was significantly lower than the mean of the mentor support group on the pretest F(1, 148) = 5.476, p = .021, and also on the posttest, F(1, 148) = 20.652, p = .000. This indicates that mentor support contributed to an increase in Motivation in comparison to the control group.
A statistically significant interaction effect was also found between Initiative and Regulation and Group. The interaction was probed by testing the effect of Initiative and Self-Regulation on Group, with Sidak-adjusted p-values. Pairwise comparison showed that the means of Initiative and Self-Regulation on the pretest did not significantly differ between the two groups, F(1, 148) = .016, p = .900, while the mean of Initiative and Self-Regulation was significantly higher in the mentor support group on the posttest in comparison to the control group, F(1, 148) = 6.040, p = .015. This indicates that mentor support contributed to an increase in Initiative and Self-Regulation in comparison to the control group.
A final statistically significant interaction effect was found between Social Attitude and Group. The interaction was probed by testing the effect of Social Attitude on Group, with Sidak-adjusted p-values. Pairwise comparison showed that the means of Social Attitude on the pretest did not significantly differ between the two groups, F(1, 148) = .155, p = .695, while the mean of Social Attitude was significantly higher in the mentor support group on the posttest in comparison to the control group, F(1, 148) = 5.555, p = .020. This indicates that mentor support contributed to an increase in Social Attitude in comparison to the control group. The other interaction and main effects did not indicate to be statistically significant, indicating that there are no significant differences between the control and mentor support group for Creativity and Innovativeness. In summary, Motivation, Initiative and Self-Regulation and Social Attitude increased significantly in the mentor support group in comparison to the control group.
22 Teachers’ Perceptions of Research Skills
The results of five repeated measures ANCOVAs with Wondering, Collecting Answers, Processing Data, Drawing Conclusions and Presenting Findings as outcome variables, are displayed in Table 5. There was a statistically significant interaction effect between Wondering and Group, so the main effects cannot be interpreted unambiguously. The
interaction was probed by testing the effect of Wondering on Group. Sidak-adjusted p-values were used to control inflation of the Type I error rate for multiple pairwise comparisons. Pairwise comparison showed that the mean of Wondering for the control group was significantly lower than the mean of the mentor support group on the pretest F(1, 148) = 7.892, p = .006, but not on the posttest, F(1, 148) = 1.686, p = .196. These results indicate that Wondering increased in the control group in comparison to the mentor support group.
There was also a statistically significant interaction effect between Drawing Conclusions and Group. The interaction was probed by testing the effect of Drawing Conclusions on Group, with Sidak-adjusted p-values. Pairwise comparison showed that the means of Drawing Conclusions on the pretest did not significantly differ between the two groups, F(1, 148) = .277, p = .599, while the mean of Drawing Conclusion was significantly higher in the mentor support group on the posttest in comparison to the control group, F(1, 148) = 5.325, p = .022. These results indicate that Drawing Conclusions increased in the mentor support group in comparison to the control group. There were no other significant interactions and significant main effects, indicating that there are no significant differences between the control and mentor support group for Collecting Answers, Processing Data and Presenting Findings. In summary, Wondering increased significantly more in the control group, while Drawing Conclusions increased statistically more in the mentor support group.
Discussion
This study explored how effective a mentor support intervention was for satisfying basic psychological needs of third and fourth grade students. In addition, teachers’ perceptions of students’ scientific attitude and research skills were investigated. Important conclusion of this study is that a discrepancy was found with regard to what students reported about their needs satisfaction and what teachers reported about motivation towards science.
23 Students’ Needs Satisfaction
Inconsistent with expectations, students noted that the mentor support intervention did not contribute to their feelings of autonomy, competence or relatedness in comparison to students in the control group. According to the Self-Determination Theory (SDT, Lavigne et al, 2007; Ryan & Deci, 2002) and solely based on these results, it can be assumed that the mentor support intervention did not help teachers to change their behavior. SDT describes that students’ needs satisfaction mediates between a supportive science context and motivation towards science education (Lavigne et al, 2007). Based on this theory and the present study’s results, it is doubtful that science motivation increased with mentor support. It is therefore also unlikely that students with mentor support had higher intentions to become a beta
scientist. These findings do not align with what was expected based on the findings of Forbes and McCloughan (2010), and Forbes and Skamp (2013; 2014), who describe based on
interviews with teachers, that mentor support increased students’ motivation.
However, the findings of the present study are not as unsatisfactory as they may seem. Students already reported that they had satisfied basic needs before the mentor support intervention started. These findings might be explained by the school’s effective approach in providing students with structural science education. This has probably made it more
challenging for teachers to increase students’ feelings of autonomy, competence or
relatedness. Moreover, students explained that when mentor support was offered their feelings of autonomy, competence and relatedness did not decline in comparison to the control group. Instead, mentor support was able to maintain their already very supportive science context. This shows that mentor support is still a very promising way of providing students with structural science education. In addition, teachers’ perceptions showed to be very different from their students’ perceptions which underlines the notion that mentor support is indeed promising.
Teachers’ Perceptions of Scientific Attitude
Students participating in this study were very young (age: 8 to 11 years). Even though it was assumed beforehand that they could report well on their basic psychological needs, within-subject correlations show that teachers (r = .66 to .88) reported much more reliable than their students (r = .44 to 48). This could explain the discrepancy in what students and teachers reported. Teachers observed that with mentor support their students’ science motivation increased in comparison to what teachers observed in the control group. This resonates well with expectations and the literature (Forbes & McCloughan, 2010; Forbes & Skamp, 2013;
24 2014). According to SDT, these perceptions of teachers indicate that students experienced that mentor support increased their needs satisfaction (Lavigne et al, 2007; Ryan & Deci, 2002). As seen, students did not report that this was the case. It might therefore be that mentor support does have a direct effect on motivation of students, which is not mediated by an increase in needs satisfaction. Although teachers already thought that their students were highly motivated for science education before the intervention, teachers still saw a growth in students’ motivation with mentor support. It is assumed that if science motivation was reported lower by teachers before the intervention, the effect of mentor support would even perceived to be larger by teachers. These results show that mentor support can be successful in helping students to get motivated for science education, even when students’ needs satisfaction did not improve.
Teachers noted that their students in the mentor support group showed more initiative and self-regulation, and had better social attitude in comparison to students in the control group. This is in accordance with the literature (Forbes & McCloughan, 2010; Forbes & Skamp, 2013; 2014), which showed that teachers thought that their students took more responsibility, worked more student-centered and increasingly collaborative when mentor support was offered. Again, teachers explained that initiative, self-regulation and social attitude were already high in both groups before the intervention started. As previously discussed, this particular primary school was already providing students with motivating structural science education. In addition, it was also a Montessori school in which initiative, self-regulation and social attitude were already hugely important for the daily activities of every child. Although this was the case, the present study was still able to find improvements of these three
important aspects of a scientific attitude. It demonstrates that the presence of a mentor can lead to more initiative and self-regulation and a better social attitude.
Creativity and innovativeness were not reported to be higher by teachers in the mentor support group in comparison to what teachers reported in the control group. This is not in accordance with the study of Forbes and Skamp (2013). They interviewed teachers who described that science education requires creativity from students which was stimulated by MyScience. An explanation for the lack of improvement in the present study might be that creativity has to be developed by students with mentor support over a longer period of time than the eight weeks of the intervention. However, teachers also noted that students’ creative and innovative attitude was already high before the intervention started. In addition, it remained high over the course of the intervention. This showed that mentor support did at least not negatively influence students’ creativity and innovativeness.
25 Teachers’ Perceptions of Research Skills
Inconsistent with expectations, teachers did not observe that their students with mentor support showed to be better in collecting answers, processing data and presenting findings than students in the control group. This is not in accordance with findings of Forbes and Skamp (2014), who reported an increase in students’ understanding and application of the nature of science attributes. It should again be addressed that the short period of time in which mentor support was offered might be not enough to stimulate these fairly complex skills. This might however not be the only explanation. Teachers in the mentor support group were responsible for instructing the empirical cycle to their students. This cycle was new to them. The lack of experience with the empirical cycle might have resulted in teachers who
perceived no improvement in research skills of their students.
Against expectations, teachers in the control group noted that their students were better at wondering than students with mentor support. This was unexpected because it was assumed that students in the mentor support group had more room for wondering because they conducted their own research. A possible explanation might be that mentors already helped students in a certain direction in the early stages of their research, which left less room for students to wonder in comparison to the regular method. In contrast, teachers with students in the mentor support group observed that their students were better at drawing conclusions in comparison to students in the control group. This is consistent with expectations and in accordance with findings of Forbes and Skamp (2014). It is assumed that this increase in the ability to draw conclusions was caused by the emphasis placed by teachers and mentors on the last phase of the empirical cycle (presenting findings). In preparation for this phase, students had to think carefully of their conclusions to be able to present their research. This might have resulted in students who are perceived by their teachers to be better at drawing conclusions than the students in the control group.
Limitations
The findings of the present study need to be read with several limitations in mind. The first limitation is that this study did not test for mediation effects that would have illustrated the complexity of the SDT framework (Deci & Ryan, 2002; Lavigne et al. 2007). The present study purely tested for direct effects of the intervention on students’ and teachers’
perceptions. However, the results showed that a mediation model would most likely not have provided more information. The relation between the intervention and teachers’ perceptions was much stronger, than the relation between the intervention and students’ needs
26 satisfaction. Hence the intervention seemed to have directly contributed to motivation
according to teachers, but not specifically to the satisfaction of students’ basic psychological needs. This showed that the relation between a supportive science context, the needs
satisfaction of students and the motivation towards science education are much more complex than expected. The present study showed that it is fairly simple to motivate students directly with a science intervention, but to satisfy students’ basic needs is a different and more complex story. Future research could focus on how we could aim a mentor support intervention more towards the needs satisfactions of students.
This is just one side of the coin. On the other side of the coin are the instruments which have been used to measure students’ needs satisfaction (PALDS-16, PCLDS and TASCQ). These instruments showed to be more difficult in practice for the very young age group than was expected. It seemed very hard for students to report consistently on their basic needs satisfaction. This may have resulted in relatively low and moderate reliability. The test statistics of students’ basic psychological needs therefore need to be interpreted with care. However, up to this point there are not many other options to choose from in order to measure a large group of young students’ needs satisfaction. In future research, the development of such instruments would be helpful. In addition, to complement the results of this study, interviews and observations would provide more insight into how teachers’ behavior can influence students’ needs satisfaction. It could be really interesting and exciting to try this in a follow-up study.
A second limitation is that gender and age were used as covariates to exclude their influence on any possible relations between mentor support and the outcome variables (Cheryan et al., 2015; Lavigne et al., 2007; Murphy and Beggs, 2003, Osborne et al., 2003). However, more effects are thought to have influenced motivation towards science education or other aspects of a scientific attitude. For example, the interaction between mentors and teachers might have influenced students’ science motivation. This was also observed by Forbes and Skamp (2013), who describe that some teachers found it easier to work with mentors than others. In addition, the experience of a school with science education might also be of great influence on how students perceive mentor support as seen in the present study. It is therefore advised that future research should dive into the influence of these effects and possible other effects which are not thought of in the present study.
A third limitation is that, as seen with MyScience’s teachers, it takes time to change the classroom from teacher-directed to student-centered and to get used to working with mentors (Forbes & Skamp, 2013). In addition, MyScience’s students became increasingly better over
27 time at conducting their own research. The present study was limited in comparison to
MyScience, in that it was only conducted at one school during a short period of eight weeks. A longer intervention might have rendered different results, which could be addressed by a follow-up study.
A final limitation is that the participating school in this study was already concerned with the importance of teaching science. For example, all students participating in this study were part of a one-off science zoo-project in the same year as the present study. Students were asked to design a shelter for a zoo animal. This might have resulted in higher means on every scale on the pretest and relatively low effect sizes for significant effects. It is assumed that these effect sizes would have been larger in schools where structural science education was less developed. It also shows that students’ needs satisfaction might be subject specific. For most students building a zoo probably appealed more to the imagination than conducting research on magnetism, although these two different subjects were selected to correspond to the different needs of every individual student. Future research could focus on different subjects for science education by collecting longitudinal data in which a wide range of subjects are included. This could provide greater insight into which subjects are motivating for students to improve structural science education.
Conclusion
Even with its limitations, the present study has shown to be of value in providing theoretical and practical insight into whether mentor support was able to satisfy students basic
psychological needs. It also demonstrated how mentor support contributed to certain areas of students’ scientific attitude and research skills. The most important finding is the discrepancy between what students reported about their needs satisfaction and what teachers reported about motivation towards science.
Students did not report an increase in their basic psychological needs when mentor support was offered. This indicates that their already supportive teachers were not able to further increase their basic psychological needs with mentor support. Teachers with students in the mentor support group reported an increase in motivation towards science education. This shows that mentor support did not contribute to a supportive science context as reported by students, but directly to motivation towards science education according to teachers. This discrepancy between what teachers and students report, shows the complexity of the SDT-framework in science education. On top of this conclusion, initiative, self-regulation and
28 social attitude, were reported by teachers with students in the mentor support group to
increase in comparison to what teachers reported in the control group. This leads to the
conclusion that according to teachers’ perceptions mentor support is able to promote students’ scientific attitudes. However, important to note here is that research skills were not found to increase according to teachers’ perceptions, except for the skill ‘drawing conclusions’. The skill “wondering” was even found to increase significantly more than in the intervention group.
These results of the present study may have several implications for educational researchers and educators. First, they show that the road to creating a supportive social science context that can be offered by teachers on a structural basis is long and difficult. As perceived by teachers, mentor support can help to promote students’ scientific attitude. This was also found by Forbes and McCloughan (2010), and Forbes and Skamp (2013; 2014). However, educational researchers and educators should not forget to keep searching for ways to get insight in students’ perceptions. This study shows that discrepancies can exists between students’ and teachers’ perceptions. Future research should be aware of these discrepancies. In addition, it should also focus on how to measure young students’ perceptions in a reliable way. Second, in contrast to the Forbes and Skamp (2013; 2014) study, research skills were not perceived to increase with mentor support in comparison to the regular method. This stresses implications for future research and educators who want to work with mentor support. They are advised to monitor whether students develop their research skills with mentor support over time. In conclusion, educational scientists and educators should pay close attention to
students’ perceptions of mentor support and students’ research skills. However, the
perceptions of teachers that students show an increase in motivation and other aspects of a scientific attitude seems like a solid foundation to build on in designing science education for all students in the future by researchers and educators.
29 References
Adedokun, O. A., Bessenbacher, A. B., Parker, L. C., Kirkham, L. L., & Burgess, W. D. (2013). Research skills and STEM undergraduate research students' aspirations for research careers: Mediating effects of research self‐efficacy. Journal of Research in Science teaching, 50(8), 940-951. doi: 10.1002/tea/21 102
Anderson, I. K. (2006). The relevance of science education: As seen by pupils in Ghanaian junior secondary schools. (Doctoral dissertation, University of the Western Cape). Beker, T., Graft, M., Greven, J., Kemmers, P., Klein Tank, M., & Verheijen, S. (2009). TULE
– Oriëntatie op jezelf en de wereld: inhouden en activiteiten bij de kerndoelen van 2006 [Orientation of the self and the world, content and activities of the 2006 standards]. Nationaal Expertisecentrum Leerplanontwikkeling: Enschede.
Belmont, M., Skinner, E., Wellborn, J., & Connell, J. (1988). Teacher as social context: A measure of student perceptions of teacher provision of involvement, structure, and autonomy support (Tech. Rep. No. 102). University of Rochester, Rochester, NY. Bender, W. N. (2010). Learning Disabilities: Characteristics, Identification, and Teaching
Strategies (6th ed.). Harlow: Pearson Higher Education.
Bentler, P. M. (1992). On the fit of models to covariances and methodology to the Bulletin. Psychological Bulletin, 112, 400–404. doi:10.1037/0033-2909.112.3.400
Blais, M., Vallerand, R., & Lachance, L. (1990). The perceived autonomy in life domains scale (PALDS-16). Recuperado el, 13.
Boggiano, A. K., Flink, C., Shields, A., Seelbach, A., & Barrett, M. (1993). Use of techniques promoting students' self-determination: Effects on students' analytic problem-solving skills. Motivation and Emotion, 17(4), 319-336.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen, & J. S. Long (Eds.), Testing structural equation models (pp. 136–162). Newbury Park, CA: Sage.
Chang, S. N., Yeung, Y. Y., & Cheng, M. H. (2009). Ninth graders’ learning interests, life experiences and attitudes towards science & technology. Journal of science Education and technology, 18, 447-457.doi:10.1007/s10956-009-9162-6
Cheryan, S., Master, A., & Meltzoff, A. N. (2015). Cultural stereotypes as gatekeepers: increasing girls’ interest in computer science and engineering by diversifying stereotypes. Name: Frontiers in Psychology, 6, 49.doi:10.3389/fpsyg.2015.00049 Connell, J. P., & Wellborn, J. G. (1991). Competence, autonomy, and relatedness: A
