1
Measuring Attitudes in Science: What Exactly are we Measuring and Why?
Jane Johnston
Department of Primary Education
The Nottingham Trent University
Paper presented at B.E.R.A. 1997.
ABSTRACT:
This paper analyses traditional tools used to measure scientific attitudes in secondary school children and considers their use with younger children and adults. It then describes the development of alternative tools to elicit and measure attitudes, focusing on five affective dimensions; looking at individuals' beliefs about,
* their adoption of scientific attitudes,
* their attitude to scientific inquiry,
* the social implication of science,
* the relevance of science
* the nature of science and scientists.
Once rigorously compared with more traditional measures and their validity thus considered, possible future developments can be discussed in answer to the question 'Why are we measuring attitudes?'.
Introduction
Study of the complex and nebulous area of attitudes and science has been on the increase in recent years, although hindered by important research questions such as 'What is an attitude?' and 'How can we measure attitude?'.
Attitudes can be described as postures or positions adopted or expressions of views or thoughts that have an effect on behaviour, ideas or emotions. This has led to a tripartite view of attitude, arising out of Hovland's Learning Theory Model (Hovland, Irving & Kelly, 1963) which separates affective, cognitive and behavioural aspects of attitude. This division has been both useful and confusing and there does appear to be as much overlap as division, with interaction occuring between the way an individual feels and the way they think and act. Indeed, we can argue that affective attitudes are the root of both cognitive and behavioural attitudes, so that how we behave is a result of how we think and an inter-relation of how we feel and think.
In science and science education, the major division has been in terms of scientific attitudes, that is cognitive and behavioural attitudes necessary to undertake scientific inquiry, to be scientific, and attitudes towards science or affective attitudes (Gardner 1975, Johnston 1996).
Education has generally focused on changing scientific attitudes which are rooted in behaviour and cognition (e.g. Harlen, 1977a & b). Research, on the other hand, has concentrated on identifying aspects of affective attitudes (Fraser 1978) and the effects of affective attitudes on behaviour and cognition (Fraser 1982, Shrigley 1990), often based on Ajzen & Fishbein's (1980) theory of reasoned action.
The importance of affective attitudes is generally accepted. They are thought to influence personal and social decision-making (Millar, 1997) as well as affect performance in school science (NCC 1989, Johnston, 1996, Harlen 1996) and especially uptake of science at higher levels (Sears 1992, Havard 1996). In the UK, the recent interest in developing the public understanding of science, through the work of the Royal Society and the British Association for the Advancement of Science and most particularly the National Week of Science, Engineering and Technology, has been criticized for focusing on 'flag waving' (Thomas, 1997); addressing superficial affective attitudes rather than the root causes of attitude.
Both education and research has highlighted the continued need for the development of affective attitudes to science, not only in children but teachers and initial teacher training students (Johnston, Ahtee & Hayes 1996, Watters & Ginns, 1994), as well as parents (ASE, 1992 Johnston, 1995). Some projects e.g. parental interactive workshops, described by Eland et al. (1995) and SEARCH (Science Education and Research for CHildren) described by Bruce et al (1997), advocate collaborative methods of attitude change. This is in recognition that development of attitude will occur through early interactions with family, teachers and peers, being influenced by unconscious prejudices, interests and ideas.
Measuring Attitudes.
Measurement of affective attitudes has been problematic as there are questions about what attitudes are being measured and how they are measured but questions about the questioning context and the validity of the response.
A number of researchers have addressed the problem of measuring attitudes to science by devising semantic differential tests (e.g. Ahtee & Rikkinen 1995) or subject preference studies (e.g. Havard 1996). Whilst these tests are easy to administer problems of analysis and interpretation can occur, as they are often decontextualized and do not measure attitude change. More common measurements involve the use of Likert-type scales (1932), which measure attitude within identified dimensions (e.g. NBEET 1993). These too, are often decontexualized, but even if set within a specific context they can pose difficulties because generalizations cannot be made (Driver et. al 1996). Such scales can also pose analytical difficulties because they are not necessarily uni-dimensional and internally consistent (Gardner 1995), but even those statistical sound scales have been found to be of limited use, particularly with groups other than secondary school children. Likert-type scales with a large number of statements may not be carefully considered by adults who will veer towards the middle ground in responding, or indicate espoused views, thus giving little indication of their implicit views or how these views affect decisions or actions. The attitudes young children hold towards science and scientists are often gauged using a 'draw-a-scientist' test (Chambers 1983). More recent studies (A.S.E. 1997) have focused on the espoused views of young children in answer to set questions, although the focus remains primarily on views of science and scientists.
There seems to be a need for formative measures, of use with a wide age range of individuals from child through to adult which explain attitudes and also indicate experiences which are significant in the development of science-related attitudes. The use of the draw-a-scientist test and picture interaction has been of use in raising awareness of views of scientists and relevance of science (Johnston, 1995) in adults and consideration of individual science biographies (Gray 1996) has been of use in explaining held attitudes as a result of science-related incidents.
The Development of New Tools.
This study involves the modification and use of techniques and strategies, previously successful in raising awareness of and explaining science-related attitudes, such as picture interaction and critical incident analysis. It focuses on five dimensions reflecting the inter-relationship between affective, behavioural and cognitive attitudes to science. These dimensions are taken from Fraser's Test of Science-Related Attitudes (1981) which has been substantially used and validated (Schibeci & McGraw 1981). They were chosen as the most appropriate dimensions for use with a wide range of individuals.
Statements from TOSRA which related to these five dimensions were used to create a Likert-type questionnaire. Statements were chosen because of their suitability for use with our chosen target group, adults attending primary science workshops.
* Views of Science and Scientists
Scientists are more committed to science than their families
Scientists are less friendly than other people
Scientists have diverse hobbies and interests
You can always tell a scientist by their appearance
Scientists are concerned about their working environment
* Attitude to Scientific Inquiry,
I would prefer to find out why something happens by doing an experiment than by being told
I like to question scientific ideas because science does not know everything
I can watch the television to find out all I need about science
I am more likely to understand science if I experience it myself rather than being told
I would prefer to read about science than do it
* Adoption of Scientific Attitude
I am curious about the world in which we live
We don't need to repeat experiments to check that we've got the correct results
I enjoy reading about things which disagree with my previous ideas
In science experiments I like to use new methods which I have not used before
I am unwilling to change my ideas
* Relevance of Science,
Science lessons are a waste of time
Science is important so there should be more science lessons.
A career in science would be boring
I get bored watching science programmes on the television
I use my science knowledge in some of my leisure activities
* Social Implications of Science
Scientific discoveries are doing more harm than good
Too much money is being spent on science which could be put to a better use
Money spent on science is well worth spending
Science helps to make life better
More money should be spent on scientific research
Scoring on this questionnaire was in common with other likert-type scales, with strong agreement to negative statements scoring 1 mark and to positive statements 5 marks.
Negative responses
Strong agreement =1, Agreement=2, Not sure=3, Disagree=4, Strong disagreement=5
Positive responses
Strong agreement =5, Agreement=4, Not sure=3, Disagree=2, Strong disagreement=1
Thus a high score (<25) would denote a positive attitude and a low score (5>)a negative attitude.
The aim was to develop alternative tools to measure attitude which would compare significantly with the questionnaire, but also be of future use with younger children. Additionally, by being interactive, it could raise possible explanations for those attitudes measured. Initially the devised tools were used as part of a two hour interactive primary science workshop with adults and the resulting scores compared with the questionnaire responses by the same people.
A 'Draw and Describe' a scientist poster aimed to collect data on views of science and scientists. These had previously been used successfully with adults and children and were adapted to enable scoring and comparison with the scores on likert questionnaire items. Marks were deducted from an initial score of 13 for every negative and stereotypical attribute and added for every non-stereotypical attribute.
Attitude to scientific inquiry, adoption of scientific attitudes and relevance of science were assessed using a picture interaction sheet. To ascertain attitude to scientific inquiry the respondent was asked to identify statements they agreed with, such as 'I don't need to find out new things 'and 'I use science ideas'. Marks were again deducted for each negative statement identified and added for each positive statement; a high score thus indicating someone who had a positive attitude to scientific inquiry. To ascertain the adoption of scientific attitudes respondents were asked to identify from a given bank of words ones which described themselves as individuals. Words included precise, logical, open-minded, objective. Each circled word was given 2 marks, so the higher score indicated an individual with more positive views of their scientific qualities. Relevance of science was determined by circling pictures of people who were using scientific knowledge or skills in an activity they were undertaking (e.g. a snooker player, priest, a weather forecaster and a policeman). Marks were given for each picture circled. Having completed this sheet respondents were encouraged to discuss their views with others, raising awareness of their own ideas and differences between their ideas and the ideas of others.
Participants at the workshops were asked to consider statements related to the social implications of science and to indicate which statement, if any, they agreed with. Again they could discuss these statements with others, which would aid the setting of the views within a social context.
In addition to using these tools, some workshop participants were interviewed with particular reference to the five dimensions and with an additional intention of illuminating the experiences which had been significant in the development of their attitudes to science. In all, responses to the questionnaire and picture interactions were collected from over 100 workshop participants in 3 workshops.
Preliminary Findings.
It was obvious at a fairly early stage of research and analysis that we had some problems with the tools and that we were not necessarily measuring the same dimensions. The interactions and interviews did however illuminate aspects of attitude and provide indications of how we could improve the tools.
From the data it could be clearly seen that respondents with a high score on the questionnaire and thus a more positive attitude towards science also score high on the picture interaction tests. However, the range of scores for the questionnaire were significantly greater than those for the picture interaction and more importantly there were no significant differences between the individual items. It appeared that responses to the picture interaction required a great deal more thought than responses to the questionnaire, which can be completed with little consideration or with the intention of identifying espoused views. As a result, when responding to the picture interactions some participants experienced some confusion or conflict between their espoused views and implicit views; in many cases science was not seen as clearly as it had been previously. It may be that in raising awareness of other views, the interactions exposed the fragility of individual views and began to facilitate a change of ideas. This would correspond to similar findings on the use of picture interactions in changing attitudes to science (Eland et.al 1995) and illustrate the possible power of interaction in attitude shift.
The 'Draw and describe a scientist' test almost always results in a stereotypical white, European, male scientist, unless there is a conscious decision made by the respondent to espouse an alternative view. In such instances where a non stereotypical image is drawn the correlation between the results on the questionnaire and the picture interaction are stronger. Adult respondents, unanimously assign the cause of stereotypical images to the media and whilst the media has an apparently profound influence on views of science and scientists, Jarvis' (1994) responses from children indicates the influence of school experiences. She provides evidence that children are more likely to indicate their implicit views of science and scientists, drawing pictures representing their experiences of school science; themselves as scientists, their teacher as a scientist or the connection between science and art, as science is always recorded by drawing a picture.
The questionnaire and picture interaction responses within the dimensions, adoption of scientific attitudes and attitude to scientific inquiry, consider individual views (affective attitudes) of behavioural and cognitive attitudes. In addition, the practical workshops allow for the actual behavioural and cognitive attitudes of some individuals to be ascertained and the differences between espoused views and actual behaviour to be considered. Using the three methods of collecting data, illustrated the importance of context in attitude study. Responses within the picture interaction for the adoption of scientific attitudes were decontextualized, being words describing scientific attitudes, whilst the questionnaire responses and especially the workshop activities were contextualized and as such produced responses which indicated actual behavioural attitudes. Responses to attitude to scientific inquiry were significantly similar in both the questionnaire and the picture interactions as both involved response to statements and were very similar in nature.
The use of pictures in considering views of the relevance of science has been useful (Johnston 1995), as not only does it firmly set the view within a context, but is also formative in illustrating different views on the nature of science and scientific attitude. Debate has occurred between respondents, as within the scientific community, as to whether informal science knowledge constitutes actual knowledge and subsequently whether there can be scientifically illiterate individuals in society. A recent article in Science and Public Affairs (Richards 1996) has a photograph of two young boys fishing with the caption 'Are these boys applying science or simple common sense to their fishing?' which captures the essence of this argument, although purists will argue, as Wolpert (1992 & 1997) does, that 'common sense science' is not real science. Most adult respondents are not purists and are able through interaction to identify the science knowledge and skills necessary to undertake seemingly irrelevant activities. Lack of relevance of school science was strongly identified through interactions and interview and indeed school science was a continual focus for participants.
There is no correlation between responses in the questionnaire and responses in the picture interaction for either relevance of science or social implications of science, indicating that we are not measuring the same dimensions. In some ways this is surprising for the dimension dealing with the social implications of science, because both tools require similar cognitive responses to statements and both provide some contradictory responses. Within the questionnaire, respondents are likely to agree that 'too much money is spent on science which could be put to better a use' and that 'money spent on science is well worth spending'. The picture interaction sheet for social implication provides some context through the chosen pictures but each statement can be interpreted in a number of different ways. Respondents have indicated that they both agree and disagree with the statements, depending upon the different contexts that they apply to them, for example, some scientific discoveries do more harm than good whilst others are regarded as beneficial.
Discussion
Whilst this study is in its early stages and we are, at present, no nearer the provision of tools of use in diagnostic and formative measurement of science-related attitudes, there are some deeper insights into attitude development occurring from this study.
The importance of interaction in attitude development and change was an important aspect of this study from the outset, as the formative nature of interaction was recognized (Eland et.al. 1995). Where cognitive in nature interaction my have a role in attitude development through identification of differences between implicit and espoused views and, when in conjunction with practical engagement, by identifying conflict between espoused views and behaviour.
The contextualization of situations was particularly powerful in looking at social implication of science and was useful in raising awareness of the relevance of scientific ideas in decision making, for as Driver et.al (1996:134-5) say the 'ability to make sense of scientific controversies and disputes' can be seen as 'an important facet of public understanding of science'. This has implications for a school science devoid of relevant context or opportunities for moral and ethical debate (Johnston 1995b), although even more compelling is the evidence of the profound influence of school science on developing attitudes. Responses to interactions and interviews provided anecdotal stories about school science being irrelevant and decontextualized, backing up research other studies which have illustrated the problems of school science (Haladyna et.al. 1982 & Woolnough 1994). More importantly is the evidence that positive experiences or critical incidents (LaLumia & Baglan 1981, Gray 1996) aid the development of positive science-related attitudes