Conference PaperPDF Available

The relation of prior IT usage, IT skills and field of study: A multiple correspondence analysis of first-year students at a University of Technology

Authors:

Abstract and Figures

Starting from the question of whether the students in the different fields of study differ in terms of several variables related to IT and learning with IT, namely their general extent of IT usage, the extent of IT usage for learning, the IT skills of the students, their participation in online courses, if they learned coding basics at school and the extent of IT usage in class at school, a multiple correspondence analysis (MCA) was chosen to find answers. The MCA resulted in four clusters of fields of study. The inclusion of the variables in the biplot shows that for one dimension the IT skills of the students are crucial, and for the second dimension a further variable, the type of school added as supplementary points, explains the differences between subjects. Finally, the article discusses that these insights might change drastically-especially due to the changed conditions of first-year students with regard to their experiences with the use of IT in learning as a result of the school closures during COVID-19 pandemic. The data for this analysis come from the survey of first-year students at Graz University of Technology (TU Graz) in 2020 (N=955) which is subjected to a secondary analysis in this article.
Content may be subject to copyright.
Preliminary Version (Preprint) of:
Mair, B., Ebner, M., Nagler, W., Edelsbrunner, S. & Schön, S. (2021). The relation of prior IT usage, IT
skills and field of study: A multiple correspondence analysis of first-year students at a University of
Technology. In T. Bastiaens (Ed.), Proceedings of EdMedia + Innovate Learning (pp. 304-312). United
States: Association for the Advancement of Computing in Education (AACE). Retrieved August 15,
2021 from https://www.learntechlib.org/primary/p/219673/
The relation of prior IT usage, IT skills and field of study: A multiple
correspondence analysis of first-year students at a University of
Technology
Bettina Mair
Educational Technology, Graz University of Technology
Austria
office@bettina-mair.at
Martin Ebner
Educational Technology, Graz University of Technology
Austria
martin.ebner@tugraz.at
Walther Nagler
Educational Technology, Graz University of Technology
Austria
walther.nagler@tugraz.at
Sarah Edelsbrunner
Educational Technology, Graz University of Technology
Austria
sarah.edelsbrunner@tugraz.at
Sandra Schön
Universitas Negeri Malang
Indonesia
sandra.schoen.fs@um.ac.aid
Abstract:
Starting from the question of whether the students in the different fields of
study differ in terms of several variables related to IT and learning with IT, namely their
general extent of IT usage, the extent of IT usage for learning, the IT skills of the
students, their participation in online courses, if they learned coding basics at school and
the extent of IT usage in class at school, a multiple correspondence analysis (MCA) was
chosen to find answers. The MCA resulted in four clusters of fields of study. The inclusion
of the variables in the biplot shows that for one dimension the IT skills of the students are
crucial, and for the second dimension a further variable, the type of school added as
supplementary points, explains the differences between subjects. Finally, the article
discusses that these insights might change drastically - especially due to the changed
conditions of first-year students with regard to their experiences with the use of IT in
learning as a result of the school closures during COVID-19 pandemic. The data for this
analysis come from the survey of first-year students at Graz University of Technology (TU
Graz) in 2020 (N=955) which is subjected to a secondary analysis in this article.
Introduction
The service department “Educational Technology” is responsible for all e-learning activities at Graz
University of Technology (TU Graz) in Austria. Again and again, differences in needs and previous
knowledge of students depending on their field of study are perceived. This paper aims to find out
exploratively whether there are indeed such groups of subjects where first-year students differ in terms
of their IT skills and equipment.
More generally, the choice of major is perceived as the result of numerous factors, such as socio-
economic backgrounds, milieus, parental expectations, etc. Although research on existing literature and
insights concerning IT related factors and study choice / fields of study was carried out, it was not
possible to find similar studies from the last few years. With regard to factors influencing the choice of a
major in IT, there is for example a contribution from Canada that shows that school dropouts from rural
areas are less likely to choose studies in STEM than those from urban areas (Hango et al., 2019).
Humburg (2017) analyzed how the factor-5 personality test results are related to the choice of degree.
With the data for our annual survey amongst first-year students at TU Graz, similar analysis concerning
possible correlations of Web 2.0 behavior and field of studies was conducted (Ebner, Nagler & Schön,
2013) and clusters emphasised as “architecture students hate Twitter and love Dropbox” were found.
Within this contribution, the current data from the September 2020 survey of freshmen at TU
Graz is the basis for analysis. This paper aims to determine whether it might make sense to assume that
students have different prerequisites depending on their chosen field of study at TU Graz and, for
example, to offer different targeted measures for first-year students.
Research questions and approach
Our research questions are the following:
Are there differences and similarities between students of various fields of study at TU Graz in
terms of prior knowledge, behaviour and equipment around IT and digital learning?
If this is the case, does it make sense to adapt the strategies of the Educational Technology
Department targeting first-year students based on their study subjects or groups of study
subjects?
The data used for the following investigations stem from an annual survey amongst first-year
students of TU Graz. The most recent data was collected in September 2020 during the introductory
event for first-year students at TU Graz, called “Welcome Days”, using a questionnaire. The
questionnaire focuses on possession of digital devices, use of IT in general and for learning purposes, as
well as IT competence first-year students have gained in secondary school including aspects such as risk-
awareness and communicational behavior. The questionnaire contains mostly questions in a closed
format. The survey has taken place annually since 2007.
The basis for the analysis presented in this paper are the answers of 955 students on questions
such as their chosen major, their extent of IT usage in general and for learning, IT skills, participation in
online courses and other (described below). Interpreting associations of several complex categorical
variables solely based on contingency tables might be challenging, sometimes even impossible.
Therefore, a method that gives the opportunity to graphically discover similarities and dissimilarities
between different comparison groups using multiple other categorical variables was used: For the
analysis we have chosen the multiple correspondence analysis (in short MCA, see Abdi & Valentin, 2007;
Lam, 2016; Kassambra, 2017). MCA is an exploratory multivariate technique that visualizes cross-tabular
data. Based on relative frequencies (the so-called column profiles and row profiles), the associations
between variables are visualized as points in a low-dimension plot (usually two or three dimensions). In
this process, the column and row profiles are mapped along axes (dimensions) that later can be
interpreted in terms of their content. The intersection of the axes (the centroid) represents the average
values. The more a profile diverges from the centroid, the more it differs from these average values.
Additionally, column points that are located close to one another tend to share similar profiles (the same
applies to row points).
The variable to be explained in the MCA is the field of study at TU Graz. To possibly reveal
similarities and dissimilarities between the students of the different bachelor’s degree programs at TU
Graz, the following variables were included in the MCA:
the general extent of IT usage (1) as well as the extent of IT usage for learning (2) regarding
office applications, gaming, communication and social media, cloud technologies and online
encyclopaedias
the IT skills (3) of the students to the extent to which they consciously exert IT-related activities
the frequency of participation in online courses (4)
the answer to the question whether they did or did not learn elementary coding basics at school
(5)
the extent of IT usage in class (6) during their middle school days
While variables 4 and 5 were asked directly in the questionnaire, the remaining variables (1, 2, 3 and 6)
were calculated in the form of a mean value index and then divided into the three levels low, mid and
high. For general IT use (1) and IT use for learning (2), the frequency of use of 33 different IT applications
in the areas of office applications, gaming, communication and social media, cloud technologies and
online encyclopaedias served as the basis. With regard to IT skills (3), 8 items were used that ask about
the frequency of IT-related activities (e.g. programming own applications or the use of encryption
techniques), while IT use in teaching (6) is based on 9 different items such as a question about the use of
digital textbooks or Moodle platforms at school.
Of the 20 subjects of study at TU Graz, only the data of students from 14 subjects could be
used: To avoid empirically unsubstantiated statements and interpretations, the MCA only includes
studies with a sample of more than ten students.
Results
In the following section, the results of the analysis are presented, first the results of the
independence test of the variables and then the results and content interpretation of the multiple
correspondence analysis.
Chi-square test shows dependence of variables and model
To determine whether the chosen variables are related and to evaluate the overall model, we
performed a chi-square test of independence. The test shows a statistically highly significant result (chi²
= 491.61, p < 0.001), so it can be concluded that the variables of the MCA are indeed related. Their
distributions differ from one another, so we can assume they statistically depend on each other.
The first two dimensions (axes) of the model that are visualized in the following plots explain
64.9% of the total variance of the contingency table (the so-called “total inertia”, see Lam 2016, p.305).
Dimension 1 accounts for 49.9% of the total inertia, dimension 2 for 15.5%.
Column points of the MCA: Four clusters of field of studies
To identify indications of similarities between the individual fields of study, we first examine the
plot of the column points. The closer two profiles are located to each other, the more similar they are
with respect to the latent characteristics represented by the two axes.
Figure 1: Plot of column points: Spatial distribution of students in different fields of study based on their
IT skills and usage
The first dimension is mainly characterized by the contrast of two clusters. The fields of study
“Architecture” and “Civil Engineering Sciences” seem to be similar and at the same time in strong
contrast to the fields of study “Software Engineering”, “Computer Sciences” and “Information and
Computer Engineering”. These five fields of study are also the main contributors to the alignment of the
first axis (see Tab. 1, column "contribution to axis 1"). The remaining fields of study are more in line with
the average with respect to the first dimension (close to the centroid) and differ more with respect to the
second latent dimension. Especially the subjects “Mechanical Engineering”, “Mechanical Engineering
and Business Economics” as well as “Electrical Engineering” are likely to show similar characteristics,
while these contrast with the fields of studies in the upper section of the diagram. Especially the subjects
“Electrical Engineering and Audio Engineering” and “Biomedical Engineering” contribute significantly
to the alignment of the second axis and differ from the previously mentioned fields of study (see Tab. 1,
column "contribution to axis 2").
Within the plot above, we call the four clusters of field of studies, marked with ellipses in Figure
1, “
Building Cluster
” on the left (“Architecture” and “Civil Engineering Sciences”), “
Science Engineering
Cluster
” at the top (“Biomedical Engineering”, “Environmental Systems Sciences”, “Chemical and
Process Engineering”) and “
IT cluster
” on the right (“Software Engineering”, “Computer Sciences” and
“Information and Computer Engineering”) and below, the “
Electrical/Mechanical Engineering Cluster
”)
(“Electrical Engineering”, “Mechanical Engineering”, “Mechanical Engin. & Business Economics).
Table 1: Parameters of the column points
Note: The highlighted cells indicate that these variables contribute more to the dimension than
expected.
column profiles
mass
inertia
contribution to axis
correlation with axis
1
2
1
2
Architecture
0.125
0.160
0.279
0.008
0.864
0.007
Civil Engin. Sciences & Constr.
Management
0.074
0.121
0.125
0.005
0.512
0.006
Electrical Engineering & Audio Engineering
0.019
0.036
0.002
0.113
0.034
0.486
Biomedical Engineering
0.114
0.036
0.003
0.168
0.041
0.727
Environmental Systems Sciences
0.047
0.034
0.004
0.081
0.052
0.365
Chemical and Process Engineering
0.029
0.037
0.001
0.050
0.016
0.213
Computer Science
0.108
0.150
0.254
0.000
0.835
0.000
Software Engineering & Management
0.090
0.104
0.142
0.026
0.671
0.039
Information & Computer Engineering
0.074
0.095
0.153
0.009
0.801
0.014
Mathematics
0.034
0.049
0.010
0.027
0.100
0.087
Chemistry
0.040
0.040
0.012
0.005
0.149
0.020
Electrical Engineering
0.072
0.053
0.001
0.254
0.009
0.740
Mechanical Engineering
0.097
0.047
0.014
0.104
0.142
0.341
Mechanical Engin. & Business Economics
0.076
0.038
0.000
0.151
0.002
0.611
Row points of the MCA: IT Skills and the role of the students' former school type
The content of the two dimensions mapped in the MCA can be further interpreted by looking at
the plot of row points in our model.
Figure 2: Plot of row points: Spatial distribution of the students IT skills and usage
Of the variables listed above (“general IT usage”, “IT usage for learning”, “IT skills,” “online
courses”, “coding basics at school” and “IT usage in school”), some contribute more clearly to the
alignment of the two axes. The first dimension of the MCA is mainly characterized by the variable IT
skills; more precisely by the contrast between low and high IT skills (see also Tab. 2, column
"contribution to axis 1"). The items "Online Courses - often", "Online Courses - never" and whether one
has acquired basic programming skills at school also contribute to the alignment of the first axis. The
left-hand side is therefore characterized to an above-average extent by students who have low IT skills,
did not gain programming skills at school, and never take online courses, while the right-hand side of the
first axis is characterized to an above-average extent by students with high IT skills (accompanied by
existing programming skills and comparatively frequent use of online courses).
Table 2: Parameters of the row points
Note: The highlighted cells indicate that these variables contribute more to the dimension than
expected.
row profiles
mass
inertia
contribution to axis
correlation with axis
1
2
1
2
General IT Usage
low
0.060
0.031
0.010
0.011
0.165
0.057
mid
0.059
0.029
0.002
0.033
0.033
0.181
high
0.052
0.039
0.004
0.118
0.055
0.470
IT Usage for Learning
low
0.056
0.024
0.001
0.005
0.016
0.035
mid
0.057
0.030
0.000
0.000
0.000
0.000
high
0.057
0.031
0.001
0.013
0.015
0.066
IT Skills
low
0.063
0.152
0.263
0.080
0.852
0.081
mid
0.051
0.032
0.008
0.021
0.124
0.102
high
0.056
0.124
0.213
0.039
0.850
0.048
Online Courses
never
0.088
0.064
0.097
0.000
0.750
0.000
rarely
0.050
0.072
0.044
0.082
0.300
0.178
often
0.020
0.079
0.087
0.099
0.544
0.194
Coding Basics at School
yes
0.089
0.094
0.115
0.178
0.608
0.295
no
0.069
0.112
0.150
0.173
0.664
0.239
IT Usage in School
low
0.048
0.028
0.000
0.097
0.000
0.530
mid
0.066
0.022
0.002
0.035
0.042
0.244
high
0.056
0.036
0.002
0.014
0.030
0.061
The second dimension of the MCA is somewhat more difficult to interpret; there is no emerging
pattern as clear as in the first dimension. The second axis also explains less of the total variance (total
variance of the contingency table). The variable “coding basics at school” contributes most strongly to
the alignment of the second axis (see Tab. 2, column contribution to axis 2). The lower part of the plot
is more likely to contain students who did not acquire basic programming skills in school, while the
upper part contains students with programming skills.
Since previous programming experiences could be related to the type of school the students
attended, in a next step another variable, the type of school students attended, was added to the model
as supplementary points. Supplementary points in MCA are variables subsequently projected into the
space that were not included in the calculation of the geometric centroid of the MCA. They can be used
supportively in the interpretation of the latent dimensions and the characterization of the column points.
If these supplementary points are included in the interpretation of the second axis, it becomes apparent
that the second dimension is primarily characterized by the contrast between grammar school and
federal secondary colleges of engineering (HTL) alumni. The upper part of the axis is dominated by
students from grammar schools, the lower part by students from federal secondary colleges of
engineering (HTL). Table 2 shows that additional items contribute to the orientation of the second axis.
Grammar school students tend not to learn programming at school, but seem to be characterized to a
certain extent by an above-average general IT use, while federal secondary colleges of engineering
(HTL) students often acquired above-average programming knowledge at school, but otherwise tend to
be characterized by a low IT use in class.
Biplot: Spatial distribution illustrating the low contribution of IT usage and higher contribution of IT skills
and school type to the differentiation of the fields of study
Figure 3: Biplot of the MCA with column points, row points and supplementary points: Spatial
distribution of students in different fields of study based on their IT skills and usage
The biplot of the MCA (Fig. 3), which contains both the column and row points, shows that the
fields of study previously identified as similar can be distinguished regarding the first dimension,
primarily in terms of IT skills. Students in the
Building Cluster
” are characterized more than average by
low IT skills, do not take online courses, and have not been taught programming skills in school. In
contrast, the “
IT cluster
” is characterized by high IT skills, frequent use of online courses and pre-existing
programming skills. Regarding the remaining fields of study, IT skills do not seem to be a differentiating
factor. Regarding IT skills, the remaining subjects correspond more closely to the average than the two
previously mentioned clusters. The second axis shows that certain fields of study are similar in that
students are recruited from similar types of schools. While the “
Electrical/Mechanical Engineering
Cluster
” seems to attract mainly students from federal secondary colleges of engineering (HTL), students
in the fields of study represented in the upper part of the biplot, the “
Science Engineering Cluster
”, are
mainly recruited from grammar schools. In the first two clusters with high and low IT skills, however, no
one school type emerges with an above-average frequency.
Summary
In summary, it can be stated that in 2020 students in certain fields of study or in clusters of fields
of study at TU Graz show similarities in terms of their
IT skills and the types of school attended
. The
extent of personal use of IT applications - whether in general, for learning purposes or during school
lessons -
cannot be used to differentiate
between the fields of study.
However, if we look at the position of the categories “low”, “mid” and “high” of the usage
variables (“general IT usage”, “IT usage for learning” and “IT usage in school”) in the biplot, we can see
that for all three areas mentioned, the categories “high” tend to be located in the upper part of the plot
while the categories “low” are located in the lower part. A high level of IT usage in the three above-
mentioned areas seems to be increasingly occurring together. This tendency is confirmed in the bivariate
cross-tabulations and by calculating Spearman's rank correlation coefficient. In particular, high general IT
usage tends to be associated with high IT usage for learning (p=.0442**) and to a lesser extent with IT
usage in school (p=.204**).
Discussion
The strength of MCA is, as mentioned at the beginning, that one can examine a multitude of
categorical variables for possible correlations and that the analysis is supported by a graphical
presentation. Otherwise, it is hardly manageable on the basis of cross-tabulations alone. The weakness
of MCAs in general is that they are, of course, highly dependent on the author's interpretation, for
example which variables are included in the model and which are omitted. One more variable might
already cause the results to look different. Therefore, the analysis and results presented here need to be
seen as explorative. Before we consider our results empirical facts, additional research, such as IT skills
tests within samples of the students in the named field of studies, are necessary.
Looking at the variable with the most importance in our analysis, the “IT skills” variable, it is of
interest as well to have a closer look at what it aims to measure. Practically, as described in the chapter
2, the “IT skills” variable is an aggregation of the results of 8 items about the frequency of IT-related
activities. As described, these activities are for example programming of own applications or the use of
encryption techniques. These are already clearly advanced IT skills; they would not necessarily be
expected from people who are considered to otherwise have very good IT skills. It is hardly surprising
that students in IT-related BA degree programs are particularly strong here. Therefore, the fact that the
IT skills of first-year students of architecture and civil engineering seem rather low according to the
survey should not, conversely, lead to the statement or impression that they only have low IT skills. Fact
is that these students answered that they rather rarely develop their own programs or use encryption
technologies. Nonetheless, their digital literacy or application-oriented IT skills might be, and are
potentially on a good level. The questionnaire was not specifically designed to meet the needs of our
evaluation; another reason for this is that the items have been asked in a similar way for several years for
reasons of comparability.
Outlook and potential consequences
The MCA has shown that students of different subjects at a TU can differ significantly from each
other. There are therefore seemingly good reasons to assume that students also have different
experiences with the use of technologies in learning. At least from the perspective of the educational
technology team, this would be an important basis for possibly different approaches to the introduction
of e-learning technologies for first-semester students depending on the subjects or the identified
clusters. However, this was not shown to be the case with regard to this particular variable. Nevertheless,
the MCA shows clear differences between first-year students in four clusters, which, however, differ
significantly in terms of their general IT skills and the school they attended. It is then less surprising that
first-year students in the field of IT-related degree programs have better prerequisites concerning their
IT skills. What is probably more surprising is that the subjects “Architecture” and “Civil Engineering
Sciences & Construction Management” are positioned in the MCA in such a way that they stand out as
subjects where students possess rather low (self-assessed) IT skills. However, it was shown in the
discussion section that the items addressed advanced activities were. From the perspective of
educational technology, the analysis has revealed exciting insights into the prerequisites of first-year
students in the various BA degree programs, but practically, due to the nature of the analysis, it is not
possible to formulate concrete conclusions for different strategies with regard to e-learning introductions
or the like.
There is another reason why we hesitate with such course-related measures and considerations:
We anticipate significant changes in the prerequisites of first-year students at TU Graz in the following
years: Due to the school closures caused by the COVID-19 pandemic, it can be assumed that they will all
start at TU Graz with far more experience of using IT for learning, regardless of school type.
References
Abdi, H. & Valentin, D. (2007). Multiple Correspondence Analysis. Encyclopedia of Measurement and
Statistics.
Ebner, M., Nagler, W. & Schön, M. (2013). “Architecture Students Hate Twitter and Love Dropbox” or
Does the Field of Study Correlates with Web 2.0 Behavior?. In Proceedings of World Conference on
Educational Multimedia, Hypermedia and Telecommunications 2013 (pp. 43-53). Chesapeake, VA:
AACE.
Hango, D.; Zarifa, D.; Milian, R.P. & Seward, B. (2019). Roots and STEMS? Examining field of study
choices among northern and rural youth in Canada, Studies in Higher Education, DOI:
10.1080/03075079.2019.1643308
Humburg, M. (2017). Personality and Field of Study Choice in University. In: Education Economics, v25,
n4 366-378.
Kassambara, A. (2017). Practical Guide to Principal Component Methods in R.
Lam, C. (2016). Correspondence Analysis: A Statistical Technique Ripe for Technical and Professional
Communication Researchers. IEEE Transactions on Professional Communication. 59. 1-12.
10.1109/TPC.2016.2583279.
ResearchGate has not been able to resolve any citations for this publication.
Article
Despite several decades of postsecondary expansion, new research finds youth from northern and rural areas in Canada still experience difficulties making the transition to postsecondary education, and those who do attend take longer to do so. Proximity, we argue, may also have a considerable impact on one’s field selection, as many of Canada’s larger universities and colleges, who offer considerably more program and degree options, tend to be concentrated in large, urban centers, and in the southern regions of Canada’s provinces. This study draws on Cycles 1–4 of Statistics Canada’s Youth in Transition Survey – Cohort A to examine regional inequalities in accessing Science, Technology, Engineering and Mathematics (STEM)-related fields at both the university and non-university levels. Indeed, our findings suggest that location of residence does impact field choices, as students from northern and rural areas were less likely to enter STEM as well as non-STEM, university programs.
Article
This paper demonstrates that the Big five personality traits (openness to experience, conscientiousness, extraversion, agreeableness, and emotional stability) measured at age 14 can be linked to field of study choice in university at around age 19. While personality matters less than cognitive skills, such as math ability and verbal ability, for educational attainment, the influence of personality on field of study choice is comparable to that of cognitive skills.
Multiple Correspondence Analysis. Encyclopedia of Measurement and Statistics
  • H Abdi
  • D Valentin
Abdi, H. & Valentin, D. (2007). Multiple Correspondence Analysis. Encyclopedia of Measurement and Statistics.
Architecture Students Hate Twitter and Love Dropbox" or Does the Field of Study Correlates with Web 2.0 Behavior
  • M Ebner
  • W Nagler
  • M Schön
Ebner, M., Nagler, W. & Schön, M. (2013). "Architecture Students Hate Twitter and Love Dropbox" or Does the Field of Study Correlates with Web 2.0 Behavior?. In Proceedings of World Conference on Educational Multimedia, Hypermedia and Telecommunications 2013 (pp. 43-53). Chesapeake, VA: AACE.
Practical Guide to Principal Component Methods in R
  • A Kassambara
Kassambara, A. (2017). Practical Guide to Principal Component Methods in R.
Correspondence Analysis: A Statistical Technique Ripe for Technical and Professional Communication Researchers
  • C Lam
Lam, C. (2016). Correspondence Analysis: A Statistical Technique Ripe for Technical and Professional Communication Researchers. IEEE Transactions on Professional Communication. 59. 1-12. 10.1109/TPC.2016.2583279.