Attitudes to e-learning, learning style and achievement in learning neuroanatomy by medical students

Abstract

Background: Two main learning approaches adopted by students have been identified by research: deep (seeking for meaning motivated by interest in the subject matter) and surface (rote-learning motivated by fear of failure). There is evidence that learning approach is influenced by learning environment (e.g. Trigwell et al. ). Online courses pose the challenge of designing software that will encourage the more desirable approach to learning.

Aims: The aims were to evaluate how successful an online course is at encouraging deep approach to learning, which factors might influence the approach adopted towards it, and whether the approach adopted is related to academic performance.

Method: Using 205 second-year pre-clinical medical students, we compared their approach to learning, as measured by Biggs et al. () 2F-SPQ-R, for a computer-aided learning (CAL) course in Neuroanatomy with that for their studies in general. We then examined student attitudes towards the CAL course and the ratings of the course Web pages in terms of the learning approach they encourage (done by 18 independent raters).

Results: The students reported using significantly less deep approach to learning for the CAL course. However, their approach for the course was not related to results on a neuroanatomy assessment based on it.

Enjoyment of the course, assessment of the amount of information in it as appropriate, and ease of understanding the course were all associated with a deeper approach.

The only agreement between the raters of the CAL course was for some pages that included patient case studies, which were unanimously given a very high deep rating. Assessment marks for questions referring to these pages were higher than for the rest of the assessment.

Conclusions: The study suggests that maximizing the use of clinical relevance should increase the interest and enjoyableness of the course and thereby aid deep learning and retention of information.

Approach to learning

The concept of approach to learning originated with the work of Marton and colleagues (e.g. Marton & Saljo 1976a, b) and has been investigated in a variety of learning contexts by a number of educational researchers (Entwistle & Kozeki 1985; Hambleton et al. 1998; Biggs et al. 2001; Fox et al. 2001).

There is reasonable consensus that there are two contrasting approaches: deep and surface. Deep learning involves seeking the meaning of the information being studied, trying to incorporate the new information into existing knowledge and hence form a coherent understanding of the material. It is likely to be motivated by interest in and enjoyment of the subject matter. Surface learning involves rote-learning and focuses on discrete elements of the studied material without integrating them into existing knowledge and does not lead to the student going beyond the course requirements. A third approach, achieving, has also been proposed (Biggs 1985) but this can be a part of either deep or surface approach, and relates to the selection of a strategy to maximize performance.

Table

Practice points

Second-year pre-clinical medical students at Oxford University show a more surface approach to learning from a computer-aided learning (CAL) neuroanatomy course than to the rest of their studies. Students who enjoyed the (CAL) course had higher deep approach scores on it. Students who perceived an information overload on the course scored lower on deep and higher on surface approach to it. There was no relation between surface and deep scores and marks on a formative assessment based on the course.

Interview research has shown that deep approach is associated with better quality of learning (e.g. Marton & Saljo 1976a, b; Svensson 1977). The development of questionnaires to assess learning approaches (e.g. Entwistle & Ramsden 1983; Biggs 1987) has lead to much research relating these to scores in examinations. Higher deep learning scores are associated with better exam performance (Ramsden & Entwistle 1981; Watkins & Hattie 1981; Entwistle & Ramsden 1983; Trigwell & Ashwin 2003), better performance in coursework and project work (Duff, 2003), higher student self-esteem (Watkins 1996), higher course satisfaction (Trigwell & Ashwin 2003), and, in medical students, better clinical experience (McManus et al. 1998). It is self-evident, therefore, that a deep learning style should be encouraged and fostered in university students.

There is consensus in research that approach to learning is not a fixed personality trait. Biggs's (1993) original model of the learning process postulated that learning outcome depends not only on the characteristics of the student (background, intelligence, etc.) but also on the learning context, and that the learning context and the nature of the desired outcome can determine the approach adopted. This has been supported by research using a variety of methods (Biggs 1985; Watkins & Hattie 1985; Newble & Clarke 1986; Tooth et al. 1989; McManus et al. 1999; Trigwell et al. 1999).

Approach to learning and online materials

The introduction of online-teaching in university courses poses the challenge of designing software that will encourage deep approach to learning. Just like the other teaching methods such as lectures and laboratory work, online courses might or might not be successful at this. For example, online courses could incorporate links between different Web pages, and thus might help the students to see connections between different topics. Integration of information is one of the key aspects of the deep approach (e.g. Trigwell & Ashwin 2003). On the other hand, if the information that the students require is ‘all there’ on each Web page, it might encourage rote memorisation without any effort on the part of the learner, which is the defining characteristic of the surface approach (e.g. Biggs 1985). It is therefore important to evaluate online courses in terms of the approach to learning they promote and to examine which of their features are particularly successful at encouraging deep approach.

Evaluation of online courses

Recently, a number of researchers have attempted to evaluate some computer-based courses in terms of student attitudes towards them and have reported both positive and negative responses. Positive effects, compared with traditional teaching, have been found with respect to sustaining interest (Kerfoot et al. 2006) and reference ratings (Gelb 2001; Williams et al. 2001). However, such positive responses were not necessarily related to improved performance. Other studies (e.g. Vogel & Wood 2002; Hahne et al. 2005) found that the use of computer-based courses lead to a more negative attitude to them. Relevantly Hahne et al. (2005) found that increased surface learning scores could account for 9% of the variance in attitude. A recent study by Chapman & Calhoun (2006) showed that individuals who were more field-independent and more given to an abstract approach were most likely to benefit from a computer-based course. These studies illustrate that student attitudes, learning approaches and academic performance need to be considered together in order to evaluate fully an online course's effectiveness.

Current study

The research reported here is part of a larger long-term investigation into the factors influencing approach to learning and exam performance in pre-clinical medicine undergraduates at Oxford University. In the present study, in order to explore some of the issues concerning online courses and learning style, we have used the Neuroanatomy Computer-Aided Learning (CAL) course that is undertaken by second year pre-clinical medical students at Oxford University. The second year students learn their neuroanatomy both via the CAL course and through lectures, tutorials and classes, reading books and research papers, and dissection. The CAL course presents the information in a variety of ways that vary in complexity from simple labelled diagrams to interactive demonstrations of the relation between neuroanatomy and neurological disease.

Our first aim was to evaluate how successful the neuroanatomy CAL course was at promoting deep approach to learning. In order to do this, we decided to compare the approach to learning that the students adopted towards the neuroanatomy CAL course with their general tendency to adopt a particular approach (also referred to as learning orientation or learning style, see Lonka et al. 2004). If students show a lower deep approach for the online course than their general learning style, then there is definitely room for improvement in the course. Little research has been done attempting to adapt the existing questionnaires to measure learning approach to specific courses or course aspects. No widely accepted scales exist. Therefore, we decided not to change the learning approach scale ourselves (risking reduction in its reliability and validity) but to convey to participants what we aimed to measure through instructions. We also wanted to see whether learning style for the CAL course is related to academic assessment of the information contained within the course. In consideration of research mentioned above we predict that it would be.

Our second aim was to examine what it is about the CAL course that might be particularly successful or unsuccessful at encouraging deep approach to learning.

We considered this from (a) the point of view of the medical students taking the CAL course and (b) the point of view of independent assessors not previously familiar with the CAL course.

1. We examined medical students’ attitudes towards the course, specifically course enjoyment, user-friendliness, amount of information, and also the students’ computer anxiety. It is now an acknowledged view that affect, as well as motivation and cognition, play an important part in learning (e.g. Ainley 2006; Eynde & Turner 2006, Linnenbrink 2006). Enjoyment of learning might encourage learning for its own sake, i.e. increase intrinsic motivation, an aspect of deep approach. On the other hand, anxiety and perceived overload in academic work have been linked with a more surface approach to learning (Fransson 1977; Tooth et al. 1989; Trigwell & Ashwin 2003). With the introduction of online courses there is a new issue to consider: computer anxiety. The influence of perceived information overload is likely to be just as true for online courses, especially in medicine where courses tend to be loaded with material regardless of the medium. Finally, user-friendliness of the software is the basic requirement for online courses (Forman et al. 2002). If students find the course difficult to follow, they might lose any interest in the material presented and not be able to employ any of the course's important features, such as links to other topics.

2. We instructed a group of third year psychology undergraduates, who were studying approaches to learning, to develop criteria for assessing the pages of the CAL course in terms of the approach they encourage and to then use their criteria to evaluate the course. We examined their criteria and their ratings to determine whether there was any consensus regarding which features of the CAL course promote deep approach. We also explored whether the information contained within the pages judged to encourage deep learning would actually be better recalled in an academic assessment and whether those students with the most deep approach would recall these pages best.

Method

Participants

There were 86 male and 119 female second year pre-clinical medical students who completed the approach to learning questionnaires, 190 of them also completed the Neuroanatomy Formative Assessment. The study was run for two consecutive years, so these students were from two successive cohorts. All 205 students had obtained at least three A grades in A-levels (not including General Studies) before coming to Oxford, and more than 60% of them had obtained four or more grade As. (A-levels are courses, usually three or four, completed by the UK students in the two years preceding university. The final grade is awarded based on coursework and examination results. 11.6% of all A-level candidates in England attained three or more A grades in 2006, Emery (2007).)

Materials and measures

Learning approach and attitudes towards the CAL course

Learning approach was assessed using Revised Two-Factor Study Process Questionnaire (R-SPQ-2F; Biggs et al. 2001) a modified and shortened version of Biggs's (1985) Study Process Questionnaire (SPQ). R-SPQ-2F consists of ten Deep and ten Surface questions. Responses are made on a five-point scale. The results are added up to produce Deep Approach and Surface Approach score. The maximum for each is 50. Biggs et al. (2001) report R-SPQ-2F to have satisfactory reliability and a good fit to the intended two-factor (Surface Approach and Deep Approach) structure.

Participants were asked to answer R-SPQ-2F twice, once in relation to the CAL course and once in relation to their general academic work (the order of administration was counter-balanced). In addition to R-SPQ-2F they filled in 9 items from the shortened SPQ (McManus et al. 1998) and an anxiety item (which are not discussed here), a computer anxiety item (‘I find using a computer makes me feel anxious’: rated on a five-point scale from Rarely True to Usually True) and the CAL Evaluation Questionnaire.

The CAL Evaluation Questionnaire was designed to obtain useful information for both the researchers and the CAL course developers. Relevant for this study are the items: ‘I found the CAL course easy to follow’ and ‘I enjoyed the CAL course’: rated on a five-point scale from Strongly Disagree to Strongly Agree), and ‘I would like the neuroanatomy CAL to contain: 1 Less information; 2 About the same amount of information; 3 More information’.

The questionnaires were administered by computer in the classroom under supervision. Care was taken to ensure that the students filled them in without someone else watching and since the results were to be anonymized they were particularly exhorted to fill them in truthfully.

Web-based neuroanatomy course

The computer based course under investigation here was Neuroanatomy Computer Aided Learning (CAL) course designed by Zoltan Molnar, Jeremy Taylor, Susanna Blackshaw, Irene Tracey, Jo Begbie, Damion Young and Kelly Smith of Department of Physiology, Anatomy and Genetics, University of Oxford.

The course runs for eight weeks, during which time the students have one three-hour session per week in the CAL laboratory where there is staff available with explanations; they can also access the programme away from the lab. Each week's work is devoted to a different aspect of the nervous system (see Appendix A), and consists of 17–23 Web-pages with various amounts of information (see Appendix B, Figures 2, 3, and 4). Apart from textual information almost all pages contain diagrams, e.g. of parts of the brain or spinal cord and their efferent and afferent nerves. The mean number of diagrams per page is three. In some cases the pictures are interactive, e.g. the student can click on a particular label and the relevant part then lights up. Twenty six percent of the diagrams within the course are interactive in some way. Twenty of the 155 pages include patient case studies. Sixteen pages contain links to other related pages within the course and/or relevant outside websites. At the end of each week's work there is a voluntary quiz with feedback.

The CAL laboratory sessions are a compulsory part of the Neuroanatomy course, which constitute half of one of the three papers that the medical students complete during the second year.

Neuroanatomy formative assessment

This is a computer-based assessment of the neuroanatomy information contained in the CAL course. It consists of 40 multiple-choice questions. The questions are of four types: identifying labels for parts of the brain or spinal cord, identifying origin or destination of neuronal pathways, verifying whether a statement is true or false and identifying symptoms for disorders or disorders from symptoms. (Examples can not be given for copyright reasons.) The assessment is completed in the CAL lab under exam conditions. It is voluntary but the majority of students (about 75%) complete it. Feedback is provided, so the students have the opportunity to improve before the Neuroanatomy Examination, which is part of their pre-clinical medical qualification.

Assessment of the CAL course

Third year Psychology undergraduates at Oxford University who had been studying Learning Styles (18 in total) were asked first to develop criteria for rating the Web pages of the CAL course on a 1–10 scale according to whether these pages were encouraging deep or surface approach (10 = high deep; 1 = high surface). They then had to use the criteria to rate each page of the CAL course.

This was done over two three-hour sessions in the CAL laboratory. The raters worked in pairs. This task was part of a series of practical classes for Advanced Option in Education and Psychology. (More details on this will be reported elsewhere.)

Procedure

The neuroanatomy CAL course for the medical students ran for the 8 consecutive weeks of the Autumn Term. R-SPQ-2F and additions were administered at the start of the eighth (and last) week of term. Formative Neuroanatomy Assessment was administered seven weeks later, at the start of the Spring Term, before any further teaching had taken place. This was done for two consecutive academic years.

Statistical analyses

Attitudes and computer anxiety data were not normally distributed, so were analysed using Spearman's correlations. The relationship between learning approach and assessment performance was analysed using parametric correlations. For other analyses paired t-tests were used unless stated otherwise.

SPSS version 12 software was used to perform the analyses.

Results

Approach to learning and neuroanatomy CAL course

On average the students reported lower Deep Approach scores (25.55 vs. 27.44) and higher Surface Approach scores (25.77 vs. 24.02) for the CAL course than for their studies in general. Two paired t-tests showed that the difference for both Deep Approach (t = 6.01, df = 204, p < 0.001) and Surface Approach (t = 5.05, df = 204, p < 0.001) was significant. These results are illustrated in figure 1.

Figure 1. A comparison of the students’ learning approach with respect to the CAL course and their studies in general.

Attitudes towards neuroanatomy CAL course

Most students reported that the CAL course was easy to follow (61.1%), however, 20.7% disagreed or strongly disagreed that it was. Only 26.0% of students agreed or strongly agreed that they enjoyed the CAL course, 40.2% disagreed or strongly disagreed.

Participants were also asked whether the CAL course should contain less information, more information or about the same amount; 74.5% of participants were happy with the current amount of information in the CAL course. Only 10.3% thought that it should contain more information and 15.2% thought it should contain less information. For further analyses participants were divided into two groups: those who wanted the course to contain less information versus others.

Attitudes and approach to learning

As can be seen from Table 1, enjoyment of the CAL course, finding it easy to follow and not wanting it to contain less information were significantly positively correlated with Deep Approach scores. The CAL course enjoyment and not wanting it to contain less information were also significantly negatively correlated with Surface Approach scores.

Table 1.  Correlations between approach to learning and student attitudes towards the CAL course

	1	2	3	4
1 Deep approach	–
2 Surface approach	−0.28**	–
3 Course enjoyment	0.32**	−0.17*	–
4 User-friendliness	0.25**	−0.12	0.64**	–
5 Information load	0.19**	−0.30**	0.33**	0.42**

*Correlation is significant at the 0.05 level (2-tailed).

**Correlation is significant at the 0.01 level (2-tailed).

The attitudes variables were all significantly positively correlated with each other, in particular the ‘enjoyment’ and ‘easy to follow’ items. However, even if the 20.7% of students who did not find the CAL course easy to follow were removed from the analysis, there was still a significant positive correlation between the CAL course enjoyment and Deep Approach scores (Spearman's r = 0.27, N = 162, p < 0.001).

Computer anxiety

The majority of students (63.9%) reported that they rarely find that using a computer makes them anxious. Only 3.9% said that using a computer usually makes them feel anxious.

Computer anxiety was not significantly correlated with approach to learning or Formative Neuroanatomy Assessment scores.

Formative neuroanatomy assessment

Neither Deep Approach nor Surface Approach was significantly correlated with Formative Neuroanatomy Assessment scores. Also there was no significant correlation between the assessment and student attitudes or computer anxiety.

Rating of CAL-course web pages

Some criteria developed by the student raters were more detailed than others, however, there was good agreement between them on which particular features would encourage deep and which surface approach. See Table 2 for one set of the developed criteria, which is a representative example of what the different raters produced. According to all of them, the features that deserve highest deep rating are integration of different information, including prior knowledge and inclusion of clinical implications.

Table 2.  Criteria developed for rating of deep or surface learning intention of parts of the Neuroanatomy CAL course

Score	Criteria
1 = only surface	Rote learning of simple written information
2	Rote learning of labels for diagrams
3	Inspecting diagrams using information from captions (information goes beyond simple labels)
4	Rote learning of labels but also requiring consideration of how these relate to one another within the content currently presented
5	Students asked to formulate an answer and then test their knowledge
6	As for 5 but in addition, students asked to relate different bits of information within that section of the CAL course
7	Instructing student to relate information to previous learning (i.e. on earlier pages of the CAL course) within the subject area
8	Students asked to integrate information within the subject area in an engaging exercise requiring use of information not explicitly given
9	Students asked to integrate information across subject areas in an engaging exercise
10 = only deep	As for 9 but also asking them to consider wider clinical implications

However, there was a lot of disagreement in the actual page ratings, even though similar criteria were used. For 80.0% of pages the ratings ranged by 4 points or more. The most common difference between the ratings for the same page was 5.00 and the average was 4.64 (1.46).

For only one page was there a 1 point difference between the ratings, everyone agreed that it encouraged deep approach (9–10). This page presented a case study. For 15 pages (9.7%) the ratings varied by only 2 points. One of them was considered by all to encourage surface approach (1–3), four were given medium ratings and ten were given ratings at the deep end (7–9 or 8–10). All ten ‘deep’ pages included in their content at least one patient case study.

In order to relate the learning approach ratings of these study materials to learning performance, we identified the Web pages that contained information needed to answer the questions of Formative Neuroanatomy Assessment. For four questions the information was contained within the pages that received very deep rating from all reviewers (8–10). We will refer to them as Deep Questions. For each participant we calculated the percentage score on Deep Questions and the percentage score for the rest of the assessment questions.

The percentage scores on Deep Questions or on the rest of the questions did not significantly correlate with either Deep Approach or Surface Approach scores of the students. However, the percentage scores for Deep Questions were significantly higher on average than for the rest of the assessment (79.71 vs. 73.71; t = 4.83, df = 188, p < 0.001). ANCOVA showed that this difference was significant even after controlling for the students’ Deep Approach scores towards the CAL course (F(1,187) = 6.77; p = 0.01).

Discussion

Approach to learning and student attitudes towards the neuroanatomy CAL course

This study showed that the medical students’ scores on the Biggs et al. (2001) study process questionnaire for the CAL course were significantly less deep and more surface than their usual approach to learning. The differences, though statistically highly significant, were small. However, since a deep approach is desirable in university learning, this suggests that either the CAL course needs further refining to elicit a deeper approach or at least that it may be important to ensure that the course is introduced in such a way that this is achieved.

As was expected, the user-friendliness of the course was related to approach to learning. Students who did not find the CAL course easy to follow had significantly lower deep learning scores with respect to CAL. Also, understandably, they tended to find the course less enjoyable. The finding that 60% of the students did find it easy to follow is encouraging.

A striking finding of the study was that only about a quarter of the students responded that they enjoyed the CAL course and 40% reported not enjoying it. Since we found that greater course enjoyment was associated with higher deep approach scores (even when we only look at those who did find the course easy to follow), it seems likely that part of the reason that the students are more likely to adopt a less favourable approach for CAL course is that they did not find it as enjoyable or interesting as the material presented in other ways.

Neuroanatomy is a very fact-based subject: many anatomical names of structures and pathways have to be memorized and the CAL course is a dense way to present information. This could be the reason that the students did not find the CAL course particularly enjoyable and felt that a surface/less deep approach might be the most successful way of dealing with the large amount of material. Supporting this proposition, responding that the course contained too much information was associated with lower deep and higher surface approach scores and also with lower enjoyment of the CAL course and finding the course difficult to follow. This finding is in agreement with the report by Trigwell & Ashwin (2003) on Oxford undergraduates that a perceived overload in academic work was associated with a more surface approach to learning (and poorer examination results).

The problem is that with subjects such as anatomy or neuroanatomy any course would be heavily loaded with information. The important task for the course designers is to reduce the perception of heavy information load. The CAL course is only partly successful in this.

One way of making the anatomical details more interesting is to relate them to clinical information. Most of the CAL pages that did this were unanimously judged by raters to be successful in promoting deep approach to learning. It would probably also increase interest of the course and increase deep processing if in future the CAL course integrated the neuroanatomy with other aspects of the pre-clinical neuroscience course, particularly by exploring links between genetics, neurological diseases and the neurochemistry of different neuronal pathways. At the moment only 10% of pages provide any sort of links and the majority of those are referring to pages within the neuroanatomy CAL course. Yet all the raters saw integration of information as the factor that would most certainly encourage deep learning. Making connections with other parts of the pre-clinical neuroscience course is likely to make CAL more varied and enjoyable. Also research has shown that putting information in context makes it easier to remember (e.g. Bower et al. 1969; Kaplan & Murphy, 2000), so doing this could actually reduce the perceived information load of the course. Chou (2003) reviews a number of features that could be used to interconnect material within one or more online courses and increase interactivity.

It was encouraging and perhaps not unexpected that the students did not report high levels of anxiety with respect to computer-based learning; the current student generation is mostly highly familiar with the use of computers. From our results we can conclude that, in the current undergraduates, computer-related anxiety is unlikely to be a threat to the academic performance or deep approach to learning from on-line courses.

Our major problem with interpreting the interrelations between student attitudes, affective responses and approach to learning was that we only have correlational data. We presented here our results and gave our interpretation of what is likely to be happening. However, we fully acknowledge that there might be alternative interpretations and that detailed longitudinal data is needed for any stronger claims.

It is extremely rare for papers to report average R-SPQ-2F scores. Therefore, unfortunately, we cannot determine how the approach to learning of the students in this study compares to that of students in other courses/universities/countries.

Assessment of the learning approach encouraged by different parts of the CAL course

Despite being well acquainted with the concepts of deep and surface approach and agreeing broadly on the criteria that should be used to decide whether a deep or surface approach was likely to be engendered, our class of Psychology students were unable to reach consensus on many of the pages of the course as to whether they would promote deep or surface learning. This suggests that individual medical students faced with the same pages will also be likely to use different strategies to learn the material. In future it would be interesting to assess the learning approach of the assessors and see if this related to the way in which they score the pages of the CAL course.

The only pages where there was agreement were some of those where anatomy was related to neurological disorders: all the assessors considered that these pages were encouraging deep approach.

Approach to learning, student attitudes and formative assessment results

Since the assessment was formative and voluntary, the medical students are likely to have approached it with a variable degree of seriousness. Hence these results cannot be interpreted very strongly.

It was predicted that the students with a deeper approach to learning would gain better marks in the neuroanatomy assessment. However, there was no correlation between scores on deep or surface approach and the total marks achieved on the computer-based assessment. Nevertheless, the scores on the questions in the assessment that referred to pages of the CAL course which the assessors agreed were encouraging deep learning (that is the pages relating neuroanatomy to clinical information) were significantly higher than the scores on the rest of the questions. This difference was independent of deep approach scores for the individual students.

Our conclusion from these results is that pages that involve integration of neuroanatomical information with clinical applications promoted a deep approach in all students, regardless of individual differences in how deep their approach usually is. This finding implies that increasing the direct clinical relevance of the course is a straightforward way of increasing deep learning.

In view of previous work showing a positive relation between academic achievement and enjoyment and interest (Ainley 2006) and a negative relation with overload (Trigwell & Ashwin 2003) it was surprising that there was no correlation between overall scores in the assessment and any of the other items (enjoyment, finding it easy to follow, finding it contained too much/too little information) in the CAL evaluation questionnaire.

Conclusion

The CAL course presents a neuroanatomy course in a readily accessible form: the students are able to use it both in the laboratory and at home, where they will be able to integrate it with material acquired through other methods of teaching. The present study suggests that maximizing the use of clinical relevance should increase the interest and enjoyableness of the course and thereby aid deep learning and retention of information. In addition it is suggested that a further increase in deep learning would result from exploitation of the links between neuroanatomy and functional aspects of neuroscience at the cellular and molecular level. At the same time this might well reduce the perceived load of information in the course and thus counteract surface approach tendencies.

Acknowledgments

We thank Dr Damion Young and Dr Vivien Sieber for invaluable help with the electronic aspect of the study, and Dr Susanna Blackshaw and Dr Zoltan Molnar for their comments on an earlier draft of this paper. We are also grateful to the psychology students who acted as our CAL course reviewers and the medical students who took part in the study. E.S. is grateful for financial support from the Clarendon Fund, the Hill Foundation, and the Muriel Wise Fund.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Additional information

Notes on contributors

Elena Svirko

ELENA SVIRKO is doing PhD at the Department of Experimental Psychology, University of Oxford and is a lecturer in Statistics for psychology at St Hilda's College, Oxford. She has been involved in the investigation of approaches to learning in medical students for over 3 years.

Jane Mellanby

JANE MELLANBY is an Emeritus Fellow in Psychology and Physiology at St Hilda's College, Oxford and was formerly a University Research Lecturer in the Department of Experimental Psychology where she continues her research. She currently teaches the Psychology of Education and has taught medical students for many years.

Appendix A

Neuoranatomy CAL course contents

Table

Week	Topics covered
1	Gross structure of the CNS
2	The skull, meninges, blood supply and subarachnoid cisterns of the brain
3	The peripheral nervous system and spinal cord
4	Somatosensory and pain pathways from the body and face
5	Cranial nerves
6	The eye and ear
7	Descending motor pathways
8	The cerebellum and basal ganglia

Appendix B

The following figures show the screen shots taken from Neuroanatomy CAL Web pages. A variety of pages was selected to represent the types of features present in the online course.

Figure 2. Neuroanatomy CAL Web page containing pictures and text but no interactive features.

Figure 3. Neuroanatomy CAL Web page containing an interactive diagram. Footnote: Moving the mouse over one of the labels creates arrows pointing to the relevant part on the diagram.

Figure 4. Neuroanatomy CAL Web page containing a revision question.