The Approaches to Study Inventory is a 52 item questionnaire that measures the following dimensions Deep, Strategic and Surface learning. Entwhisle et al 2000
Use the ASI questionnaire to measure Deep, Strategic and Surface learning. If you want to examine any other variables relating to your particular learning environment you must include these on the questionnaire.
Distribute the questionnaires to the students
|
Create relevant variables in SPSS |
|
Input numeric value for each case into each variable, set variables to numeric. Some variables normally expressed as character strings, can be recoded into numeric or dummy variables. |
|
Use the compute command to calculate the score for each of the sub scales. Each subscale will have four separate questions relating to it, the subscale will become the sum of the score for each of the questions. |
Factor analysis20 is a multivariate analysis technique that aims to reduce a large number of variables to a few more manageable ones called factors. In this case the aim was to obtain three factors, representing 'deep', 'strategic' and 'surface' from their component subscales. In the case of deep learning the primary question being answered by the factor analysis is whether the concepts seeking meaning, relating ideas, use of evidence and interest in ideas, can be combined with confidence to form a central concept called deep learning. It is important to validate the underlying structure of the questionnaire when applying it to a new context. Factor analysis 20 was carried out on all of the thirteen subscales for the entire sample (n =215). In this instance maximum likelihood analysis, with delta set to zero and varimax rotation was carried out, following the approach used by Tait et al 17.
 |
The next stage is to validate the previously reported structure for the questionnaire using Factor Analysis. Select, Statistics - data reduction - Factor within SPSS. Select all the subscales (e.g. Seeking meaning, Lack of Purpose etc) for analysis. Within the same dialogue box select extraction and chose maximum likelihood from the pull down menu. |
|
Under
rotation select varimax |
|
Select continue and then ok |
|
you will get the following result……. |
|
Three factors emerge, values below 0.3 are not considered relevant to the factor. |
Cronbach alpha coefficients 17
were extracted using SPSS 8.0 to test the internal reliability of the three main
scales and the thirteen subscales. This
procedure is applied to test the extent to which items within a scale are
measuring the same dimension.
In the case if the ASI questionnaire for example four questions in the
questionnaire are supposed to measure the concept time management. The
cronbach alpha coefficient indicates the extent of this.
|
Statistics-Scale -Reliability analysis |
|
Select variables of interest for each subscale, set
model to alpha |
|
Click Ok |
|
Result |
Loglinear analysis 21,22 was employed in this study to
investigate the associations between the approaches and situational factors. In
this case it was used to answer three questions, (1) whether or not there is a
significant association and (2) whether this association fits a linear trend,
e.g. does an increase in one variable lead to an increase in the other? (3) the
direction and relative strength of this association. Loglinear analysis
was employed in this instance because upon investigation, the data collected in
this study violated many of the assumptions of more familiar and simpler types
of statistical analysis such as Pearson's correlation or regression analysis, in
that the data were not interval in nature and for a significant number of the
variables were not normally distributed. Loglinear analysis allows significant
relationships between two (or more) categorical variables to be identified, and
was in this instance seen to be the most robust statistical method. Loglinear
analysis also allowed an explanatory linear model to be applied to the data.
For
the purpose of loglinear analysis the main scales were recoded into, 'Low'
'Medium' and 'High' and these representations of the data were assigned dummy
variables 1,2 and 3, respectively. General loglinear analysis was carried out
using SPSS 8.0.
The general loglinear method models the counts of observations falling
into each cell in a cross tabulation or contingency table. Each categorical
variable is an independent variable e.g. deep learning, grade etc and the
dependent variable is the number of cases (frequency) in a cell of the
contingency table. Each cell count comes from a population whose distribution is
assumed to be Poisson, where the mean parameter (m)
is modelled on the log scale as a function of the important categorical
variables. For two categorical variables, the most general (saturated) model is:
log(mij)=intercept + αi + βj +αβij
where the subscript denotes the level of the categorical variables α and
β. The saturated model will always fit the data. The independence model is
represented by the notation log(mij)=intercept + αi
+ βj, i.e. the interaction term αβij
is removed from the equation.
The linear x linear model is represented by log(mij)=intercept + αi + βj+bxy, where x and y are non-zero integer-valued covariates (i.e.
1,2,3,4) representing the ordinal level of the categorical variables α and
β, respectively. Imposing a linear model simply says that one variable has
an effect which increases proportionally to the other. The b
parameter or 'estimate' indicates the quality (+ or -) and strength of the
association between the two categorical variables. The general loglinear
analysis was carried out using SPSS 8.0, which estimates parameters by
maximum-likelihood using the Newton-Raphson method.
Goodness of fit in loglinear models can be assessed using the Likelihood
Ratio (LR). Non significance of the likelihood ratio indicates that the reduced
or modified model e.g. independence or linear model respectively does not
deviate significantly from the saturated model, therefore indicating a good fit.
|
Next you should calculate using the compute command you add up the sub scales to make the main scales |
|
These should be recoded using the Recode command into categorical variables i.e. Deep, Strategic and Surface into categorical variables i.e. High, medium and Low. |
|
Ensure that there are equal numbers of people in each of the categories. |
To test the independence model
|
Select Statistics - Loglinear Analysis - General |
Result!!
|
Generate a Covariate which will be the product of the each approach variable and contextual variable. For example Strategic X Grades. |
|
Make the covariate a new variable, this has to be done for every approach and every variable. (Note how you code them). |
|
Rerun a custom model this time including the covariate into the model. - this covariate represents a linear interaction term which we will include in this model to try and explain the data. |
Result !!
|
A non significant value tells us that a linear model explains the data. |
|
You must also look at the estimate for the covariate. This will tell you the direction and the relative strenght of the association. |
1.
Banks W. Learning and teaching styles: An important component of the
veterinary medical debate.
J Vet Med Educ, 19(4): 1992.
2.
Faculty of Veterinary Medicine, University College Dublin.
<http://www.ucd.ie/vetmed/> Accessed 04/12/03.
3.
Kwan CY. What is Problem-Based Learning (PBL)? Centre for Development of
Teaching and Learning. 3(3): 1-5, 2000.
4. Keefe, J.W. Learning style: An overview. In NASSP's Student learning styles: Diagnosing and prescribing programs Reston, VA: National Association of Secondary School. 1979, p. 1-17.
5.
Myers IB, McCaulley MH, Quenck N and Hammer A.
MBTI Manual: A Guide to the Development and Use of the Myers-Briggs Type
Indicator®, 3rd Edition. Palo Alto, CA: Consulting Psychologists Press, Inc.
1998.
6.
Kolb DA. Learning-Style Inventory (LSI). Boston: Mcber & Co, 1985.
7.
Felder RM. Matters of style, ASEE Prism, 6(4): 18-23, 1996.
8.
Felder, RM. ‘Reaching the second tier': Learning and teaching styles in
college science education.
J Coll Sci Teach. 23, 286-290,
1993.
9.
Biggs BJ. Approaches to the enhancement of tertiary teaching. High Educ
R&D, 8 (1), 7-25, 1989.
10.
Entwistle N, Tait H and McCune
V. Patterns of response to approaches to studying inventory across contrasting
groups and contexts. Eur J Psychol Educ, 15
(1), 33-48, 2000.
11. Stickle JE, Lloyd J, Keller WF, and Cherney E. Learning Styles in Veterinary Medicine: Relation to Progression through the Professional Curriculum and Integration into the profession. Jour Vet Med Educ, 26 (2), 9-12, 1999.
12. Marton F and Saljo R. On Qualitative differences in learning : I Outcome and process. Br J Educ Psychol, 46, 4-11, 1976.
13. Marton, and Saljo, R.. Approaches to Learning. In F. Marton & D. Hounsell & N. Entwistle (Eds.), The Experience of Learning. Edinburgh: Scottish Academic Press. 1997 p39-58.
14. Entwistle, N. and Ramsden, P. Understanding Student Learning, London: Croom Helm. 1983
15.
Entwistle N, Hanley M and Hounsell D. Identifying distinctive approaches
to studying. Higher Education. 8, 365-380, 1979.
16.
Entwistle N and Tait H. Approaches to learning, evaluations of teaching,
and preferences for contrasting academic environments. Higher Education 19,
169-194, 1990.
17.
Tait H, Entwhistle N and McCune. ASSIST: A Reconceptualisation of the
Approaches to Studying Inventory. InC. Rust (Ed), Improving stud. Oxford: Oxford
Brookes University, the Oxford Centre for Staff and Learning Development, 1998,
p262-271.
18.
ELT Project - Publications and Presentations (containing
download of ASSIST questionnaire).
<http://www.ed.ac.uk/etl/publications.html> Accessed 30/09/03
19.
Statistics package for social sciences (SPSS).
<http://www.spss.com/> Accessed 04/09/03.
20.
Kline P. An Easy Guide to Factor Analysis, Routledge, London, 1994.
21. Agresti, A. Categorical Data Analysis, New York: Wiley series in probability and mathematical analysis, 1990.
22.
Fielding J and Gilbert N. Understanding Social
Science Statistics. Sage Publications, 2000 p287-295.
23.
Rhem, J. Deep/Surface Approaches to Learning: An Introduction. The
National Teaching and Learning Forum, 5 (1): 1-5, 1995.
24.
Ramsden, P. and Entwistle NJ. Effects of academic departments on
students' approaches to studying. Br J Educ Psychol 51: 368-383, 1981.
25.
Thomas P and Bain, JD. Contextual dependence of
learning approaches: the effects of assessments. Human Learning,
3: 227-240, 1984
26. Fransson A, On qualitative differences in learning: IV - Effects of intrinsic motivation and extrinsic test anxiety on process and outcome. Br J Educ Psychol 47: 244-257, 1997.
27.
Richardson, J, Mature age students in higher
education: I. A literature survey on approaches to studying. Stud High Educ. 19:
309-325, 1994.
28.
Chambers EA. Workload and the quality of student learning. Stud High Educ
17(2): 141-152, 1992.
29.
Kember, D, Ng
S, Tse H, Wong, ETT, and Pomfret M. An examination of the
interrelationships between workload, study time, learning approaches and
academic outcomes. Stud High Educ 21: 347-358, 1996.
30.
Light G. and Cox R. Learning and Teaching in Higher Education:
The Reflective Professional. London: Sage Publications. 2001
31.
Fresco, F. and Nasser F. Interpreting student
ratings: consultation, instructional modification, and attitudes towards course
evaluation. Studies in Educational Evaluation. 27: 291 -305, 2001
32. Bloom BS. Taxonomy of Educational Objectives: The Classification of Educational Goals. Handbook I. Cognitive Domain. New York: McKay. 1956, p201-207.
33.
Eyre P. Professing Change. The Veterinary
Profession. J Vet Med Educ. Spring; 28(1): 3-9, 2001.
http://www.ed.ac.uk/etl/publications.html
|
Results (PPT) |