Classification and similarity of teachers reflections using multidimensional scaling
1
Introduction
Classification & similarity
“Classification is basic to all our intellectual activities”
“Similarity is at the very heart of a theory of cognition and memory”
“Similarity is at the very centre of a theory of concept and the theory of similarity would why people have the concepts that they do”
Method
Qualitative or quantitative?
People understand, believe and behave.
Educational research should try to work out what and why people understand as they do, how their beliefs are formed and how both inform activity....
We can evaluate understandings, beliefs and behaviours and they can generate huge quantities of data - how to analyse such data becomes problematic, hence the need for data reduction, and traditionally this has been done by factor analysis.
Multidimensional Scaling
Similar to cluster analysis and factor analysis
MDS, FA, and PCA involve a continuous coordinate space.
MDS assumes that most tasks vary on several features and all are taken into account: that individuals vary along each dimension according to their ideal point
factor analysis limited at the level of the inidividual
assumption of a small number of factors that can represent differecne
Exemplars
What do the axes represent?
Feeling Stressed?
Kruskal’s stress function (1964) is the most commonlyused measure in determining a model’s goodness of fit
Although there is no strict rule regarding how much stress is tolerable, the rule of thumb is that a value ≤0.1 is excellent and anything ≥0.15 is not tolerable (Kruskal & Wish, 1978).
To conclude
interesting questions in ongoing work
Participants at different times may be included in the same data set. Thus, progression can be monitored in one visual representation e.g., pre/during/post
A hypothetical 'ideal' can be included in the data
to see if there has been 'movement' towards, the
'ideal' or not - the question of effect of the ideal on the data is relevant - a case of uncertainty principle?
Finally, when 'gathering' plot points into clusters
where to make the 'cut' with respect to proximity to an ideal is relevant
Classification and similarity of teachers’ reflections using multidimensional scaling
Thomas McCloughlin,
Education Department,
St. Patrick’s College (a college of DCU),
Drumcondra, Dublin 9, Ireland
Plot of multi-group evaluations of science tasks
Plot of students' views on reflections following a science lesson
Plot of perceptions against a 'norm' or 'ideal' (VAR1