Skills in problem solving, including finding and applying the appropriate knowledge to a problem, are important learning outcomes from the completion of a Physics degree at University. This thesis investigates the characteristics of successful and unsuccessful novice University students solving problems in Physics in various contexts. Gaining an insight into student behaviour can clarify areas of weakness and potentially provide research based instructional strategies in these contexts. Access to external information during problem solving, such as the Internet, is becoming an increasingly relevant research area, as students use resources for homework questions and then in employment after University. Three chapters (Chapters 3-5) investigate individual novice problem solving with and without resources, such as a textbook. Participants were from introductory years one and two of Undergraduate study at University. The results from this chapter show successful and unsuccessful approaches by students to multi-step problems. One notable result is that unsuccessful students demonstrated an inability to apply the appropriate physics concepts, with or without the availability of resources. These results have implications for the skills required in closed and open-book exams. Three chapters of the thesis focus on the analysis of Peer Instruction (Chapters 6-8), an instructional method designed to improve conceptual understanding. Peer Instruction was used with a first year Introductory University class. Technical word use was not associated with success on Peer Instruction questions. Conversations were also analysed qualitatively. The results reflect diversity in reasoning regardless of correctness on the question. Some recommendations for the implementation of Peer Instruction are presented. The thesis is organised as follows. A literature review was conducted in relevant areas of study and is presented to set the context of the work. Three chapters report the study with novice individuals solving multi-step problems with and without resources. Three further chapters investigate successful and unsuccessful Peer Instruction discussions in Physics. The final results chapter (Chapter 9) presents a study of a group of experts solving physics problems. Overall successful and unsuccessful problem solving strategies were compared, as well as preliminary comparisons between expert and novice behaviour when solving physics problems.

The phrases like ‘scientific thinking’, ‘scientific method’ and ‘scientific attitude’ are all widely used and frequently appear in school curriculum guides but the meaning of such phrases is much less clear. In addition, there is little about how such skills might be taught or assessed. In the light of this, this thesis is a study which focusses on several related areas: the meaning of scientific thinking will be explored and the features of scientific thinking which make it uniquely different from other kinds of thinking will be analysed, set in the context of what is known about how conceptual learning takes place; the measurement of scientific thinking skills will be attempted and ways by which scientific thinking can be taught in the context of physics will be developed. There are two possible hypotheses which arise in this study: genuine scientific thinking is not accessible until learners have matured developmentally and have sufficient experience of the sciences. The way the sciences are taught will encourage or hinder the development of such skills. The empirical work was conducted in three stages to explore these hypotheses. Overall, 1838 students were involved in the study. The first experimental study was carried out with students (boys and girls) aged 15-18 from various schools in the Emirates and seeks to explore the extent to which they are thinking scientifically as well as making several other measurements of their abilities and attitudes. A test for measuring scientific thinking, based on physics, was developed and used along with an established test of working memory capacity, known to be a rate determining factor in much learning. In addition, a test to measure understanding of ideas in physics was constructed and used and the national examination marks for these students in the three sciences and mathematics were considered. It was found that the test of scientific thinking, the test of understanding physics and the national examination marks measured very different outcomes which are likely to be: scientific thinking, understanding and recall, respectively. In the second stage, some of the measurements completed in the first stage were repeated to confirm the outcomes. However, the main part was the development and use of five teaching units which, together, aimed to teach the key skills which had been defined as scientific thinking. The success of this was measured by using the same test of scientific thinking and comparing the outcomes to those obtained in the previous experiment. In addition, the results from the use of two of the items in the test of scientific thinking were compared to the outcomes compared in a previous study (using the same items) which had been based on large samples of younger students (aged 12-15). A survey was also used to see how the students saw themselves in relation to their study in physics. It was found that the use of the units had improved scientific thinking significantly with the younger two groups (age 15-16 and 16-17) but no improvement was observed with the oldest group (age 17-18). It was also found that the older groups of students were significantly better in the skills measured by the items used by the previous study when compared with younger students. The outcomes of the survey showed that their self-perceptions related poorly to their abilities in thinking scientifically while the interests of boys and girls were remarkably similar, suggesting that physics could appeal equally to both genders. In addition, there is clear evidence that all students want their studies in physics to relate to the real issues of life which are important for them and that boys are less willing to memorise than girls. The third phase employed the academic game known as Eloosis (which is considered to be an excellent model of scientific thinking) with three groups: one group had completed studies in one or more of the sciences and were about to leave school; one group were studying for a degree in an arts subject and were unlikely to have had much experience in the sciences; the third group had all graduated in a science discipline recently. While all groups played the game excellently, the group who had little or no science background did not appreciate the significance of the game as it illustrated the way science works in exploring the world around while both the senior school students and the science graduates, without prompting, could express a clear conception of the way science works although the graduate group, understandably, used more sophisticated language. The overall conclusions are that the test of scientific thinking certainly measures something completely different from the other measurements and, linked to the outcomes of the academic game, it does appear that it measured something close to scientific thinking. If this is true, then such thinking can be taught but is not accessible to those younger than aged about 15-16. All of this is consistent with the type of observations made by Piaget many decades ago and suggests that any attempts to develop scientific thinking with young adolescents will be unlikely to be successful. However, with older adolescents, for the skills to develop, there needs to be some teaching of this way of thinking. With the very large sample sizes and good cross-section of the population, there is reasonable confidence that the conclusions are generalisable and can inform future practice.