Cardiff University  


Booking form Evaluation form Contact RCBN


A report on the research development seminar
'Designing and analysing interviews with representational probes'
led by Professor M. Reiss, Dr C. Boulter, Dr S. Dale Tunnicliffe
Institute of Education, University of London, on 25th March 2004

Prof. John K. Gilbert, Institute of Education, The University of Reading



The UK-based section of the ESRC-funded project 'Exploring children's understandings of objects in the natural world' addressed three research questions:

  1. 1. To what extent do the probes used in interviews affect the models of various objects expressed at different ages?
  2. 2. In what ways do pupils at different ages express their models of various objects orally in interview?
  3. 3. To what extent do out-of-school contexts affect the models of various objects expressed at different ages?
  4. That section of the project which undertook a comparison between UK and Brazil addressed the following two questions:
  5. 4. What are the characteristics of the expressed understandings of pupils in suburban Rio de Janeiro and southern England?
  6. 5. How can the nature of these children's expressed understandings of the natural world be modelled using the methodology and data of the project?

The aims of the seminar were:

a) to use the process of and the data from the project for an advanced seminar on interview design;

b) to use the analysis of the materials of the project to discuss the use of systems and modelling theory in analysing the content of interview data.

The selection, balance and timing of the component parts of the seminar were all admirable. Whilst a good pace was maintained, there was ample opportunity for participants (largely Ph.D. students) to pose questions and debate issues. The arrangements for registration were effective whilst the catering arrangements were unobtrusive and of excellent quality. This brief report is drawn from the presentations made by the Discussants, Professor Kerst Boersma and myself, the seminar leaders, and the participants. It complements the papers 'Understanding the natural world: The significance of methodology and context' and the 'Seminar Programme' prepared by the seminar leaders. It consists of a series of additional points to be borne in mind when designing, conducting, analysing, and reporting interview-based studies of this type.


It is a legend that the launch of Sputnik in 1957 lead to an explosion in the development of the school science curriculum in Europe and North America. This work was generously funded, led by scientists, and completely unfounded on research-justified theories of teaching and learning. When the programme began to meet problems in the late 1960s, largely associated with the rapid rise in the proportion of school populations that was being required to study science, a theoretical underpinning was sought. This was initially based on the work of Jean Piaget and was led in the UK by Philip Adey and Michael Shayer at King's College. By the mid-1970s this, in turn, was meeting problems in that pupils' performances were not always as predicted by Piaget. The notion that the context in which ideas were learnt and expressed was very influential and took root, producing the 'alternative conceptions' movement in which the late Rosalind Driver played such an important role and which continues to this day. If 'context' is important, then one would have expected that projects that compared the consequences of various 'contexts' would have taken place. It is a matter of considerable amazement, with hindsight, that this was not the case. The project which was the basis of this seminar addresses could and should have taken place 20 years ago. Professor Reiss and his colleagues are to be congratulated on putting a vitally important missing piece in place in the jigsaw.

Towards better interview-based studies

For this to take place, attention need to be paid to:

* The idea of 'context' itself

It can be argued that what a learner confronts in a lesson or informal educational setting is a physical situation that has been (one would hope!) selected both because it might interest the learner and because it requires scientific ideas for its understanding. A given situation is converted into a context by the mental activity of the learner who pays attention to specific aspects of the situation in the light of analogies drawn to situations previously met and perceived to the (somewhat) similar. With a context established, a learner then links it to events in personal or social life to establish a narrative. This done, the learner then has an educational experience framed by the narrative associated with the specific situation. In short, a 'context' is a personal creation and is not an externally-provided 'given'. Researchers need to find out as much as they can about what context is created by an interviewee for a given situation.

* the notion of 'cultural background'

The ever-growing heterogeneity of populations means that clearly ascribing 'cultural background' to a large group of interviewees must be of ever-declining validity. In that socially-related narratives are an important way of interpreting situations and that such narratives are heavily culturally determined, it does mean that cultural background must be evaluated at an individual level. This will require clear models for the design of samples and the use of sensitive methods for the collection of such data.

* the purpose of the enquiry as perceived by the interviewees

The seeking of valid answers to questions does require that the interviewees are aware of the purpose of the investigation. This also means that the subject matter of the probes must be capable of having personal significance for the interviewees and that they must have an overview of all those probes used at some time, preferably at the outset of the enquiry.

* the appropriateness of explanations provided by interviewees

An appropriate explanation is one where the respondent provides an answer to a specific question. It must be of the type that is logically associated with that question type (e.g. 'why' questions produce 'because' answers). It must be relevant, in that they meet the needs of the questioner. It must be of adequate quality, based on the relevant biological model. In order to obtain valid answers, the questioner must enable the interviewee to judge carefully the appropriateness of the explanation expected. Scientifically inappropriate explanations can be produced where the interviewee wrongly diagnoses the purposes of the question posed, irrespective of the scope of the biological knowledge that is known.

* the relative use of drawing and photographs

Drawings provide weak 'contextual cues' and are therefore flexible across different cultural backgrounds. This cultural sensitivity does allow interviewees to draw their own specific inferences about the nature of the situations presented to them, but they must know the conventions associated with the form of representation being used.

Photographs, on the other hand, provide strong 'cultural cues' and may therefore be effective when there is a clear perspective held on the cultural background of the interviewee. This certainty enables the interviewees' knowledge of situations, as opposed to their making of inferences about them, to be probed. In particular, this certainty enables the interviewees’ knowledge of the precise structural and functional relationships between lower levels of biological representation (e.g. cell or organelle) and upper levels of biological representation (e.g. community), which are related as 'white box' to 'black box', to be explored. Similarly, they enable all the aspects of biological thinking to be systematically explored, i.e. form-function thinking, ecological thinking, systems thinking, and biological research methodology. This great load can only be carried if the photographs used are carefully analysed to ensure that only the information wanted, and none other, is elicited.

The use of pre-prepared drawings and photographs does depend on the interviewees being capable of verbally presenting their ideas. An approach for the less verbally gifted may be to let them take and explain their own photographs and /or to draw their answers.

* clarity of address to the research questions

As such enquiries tend to be complex, it is important that there is a clear analysis of the demands of specific research questions in terms of the situations presented, the age of the interviewees, and the precise probe used.

* the exploration of specific situations

The data produced by the project showed that there is distinction between the understandings demonstrated in in-school situations and in out-of-school situations. In short, the social purposes of being in school are different from being out-of-school. This does mean that a fine-grained exploration of several out-of- school situations may not be particularly rewarding. If it can be done without boring the interviewees, the use of a pre-visit / any visit / post-visit design with the same probe may be both efficient and effective.

* the value of social knowledge to interviewees

The large category of 'human influences' data, that may be re-categorised into 'the physical location of the object', 'the acquisition of knowledge by the respondent', and 'perception of consequences for the respondent', shows the value attached to the social construction of knowledge by interviewees. There may be gender differences here. In any event, studies that concentrate too closely on the acquisition of knowledge of cognitive knowledge may fail to grasp the value of this affective knowledge.

Home > Activities > Developing the use of in-depth datasets
Return to top of page
This page was last updated 27th April 2004