Studying Programming Students Motivation using Association Rules

Paula Correia Tavares

1,2

, Elsa Ferreira Gomes

1,3

and Pedro Rangel Henriques

Departamento de Informática, Instituto Superior de Engenharia do Porto, Porto, Portugal

Centro Algoritmi & Departamento de Informática, Universidade do Minho, Braga, Portugal

INESC TEC, Portugal

Keywords: Motivation, Educational Data Mining, Association Rules.

Abstract: For Programming teachers it is of utter most importance to understand the factors that impact on students’

motivation to improve their ability to become good computer programmers. To understand a problem, to

develop an algorithm for its solution, and to write the corresponding program is a challenging and arduous

task, demanding time and self-confidence. In previous work we studied computer based technics to engage

students in the learning activity; visualization, animation, automatic program assessment were some

approaches that we combined. To support that work we studied carefully students’ motivation and

complemented that study with an inquiry to a group of students of Algorithm and Programming course of

the first year of an Engineering degree. In this paper we show how Association Rules can be used to mine

the data gathered in the inquiry to discover relationships among factors influencing extrinsic motivation.

1 INTRODUCTION

According to Hundhausen and Douglas (2000) or

Proulx (2000) and many other authors, including the

various notes included in the Computer Science

Curricula of ACM/IEEE Guideline of 2013

(ACM/IEEE, 2013), confirmed by our professional

experience teaching courses on introduction to

computer programming, learn to program is an

arduous and complex task that raises many

challenges both to teachers and students. The lack of

motivation is one of the main reasons for the

students’ failure in programming courses (Santos et

al., 2006; Ramos, 2013). This has a direct impact on

their ability to acquire knowledge and consequently

affects the results of the teaching / learning process,

which translates into a strong frustration for students

and teachers. It is important to understand the causes

of this disinterest, to combat them. In fact, there are

several reasons why students fail to learn

programming (Proulx, 2000); Faced with less

difficulty in understanding the statement, in the

development of an algorithm or in the use of a

programming language, students become

discouraged and give up (Tavares et al., 2017).

1.1 Objectives of the Study

Several theories have been developed to explain the

motivation from the beginning of the history of

psychology as a science. Because it is a complex

phenomenon, the subject has been studied under

different prisms (Williams et al., 2011; Almeida,

2012). Some of them claim that people are

motivated by material rewards, others by increasing

their power and prestige in the world, or by an

interesting work, enriched environments,

recognition, or to be respected as an individual. The

fact is that humans in general have very complex

needs and desires. Motivation is one of the keys to

understand the human behaviour; it acts on the

thought, attention, emotion and action of the Human

Being, involving desire, effort, dreams and hope

(Williams et al., 2011; Almeida, 2012).

People are driven by very different factors, with

varied experiences and respective involvement. The

motivation leads to an action directed to a particular

goal, being regulated by biological or cognitive,

factors of each person. This action is enabled by the

needs, emotions, values, goals and personal

expectations, constituting a single intentional and

multifaceted phenomena (Ryan et al., 2000).

Motivate students is one of the biggest

challenges that in general teachers have to face. In

514

Correia Tavares, P., Ferreira Gomes, E. and Rangel Henriques, P.

Studying Programming Students Motivation using Association Rules.

DOI: 10.5220/0006816805140520

In Proceedings of the 10th International Conference on Computer Supported Education (CSEDU 2018), pages 514-520

ISBN: 978-989-758-291-2

programming courses (in the context of higher

education institutions) this task is particularly

difficult due to the skills needed and the complexity

of the teaching topic; actually programming

demands a new thinking paradigm and a high level

of abstraction capabilities.

In order to assess the real difficulties, students

feel in the teaching/learning process of

programming, we have designed and conducted in

our Engineering schools a survey which, in the first

instance, was applied to 1.st year students attending

Programming, or Algorithms & Data Structures

courses. This was applied to 160 students

anonymously on the day of their exam; students

were asked to answer with the utmost sincerity and

without hesitation.

The work reported here, after the students’

responses to an inquiry, aims to: understand the

reasons for the real difficulties that arise in the

teaching / learning process of Computer

Programming and to study how these reasons are

associated with Data Mining techniques in order to

understand what kind of association rules could be

inferred between the different factors and the

motivation.

The paper is organized as follows. In Section 2

we provide a short background on Association

Rules, our working tool. Then in Section 3 we

discuss related concerning the application of Data

Mining technics and tools to the analysis of

educational phenomena. In Section 4 we review the

basics on students motivation. Section 5 is the core

of the paper: we discuss the application of Apriori

algorithm to the data gathered from a motivation

inquiry involving Programming Students in order to

extract relations among factors that impact on

motivation. Section 6 closes the paper.

2 ASSOCIATION RULES

In a dataset composed by binary variables V_1, V_2,

... V_n, it is often useful to find frequent associations

between sets of these variables. These associations

allow us to understand the interactions between any

variables and can take the form of frequent set of

positive co-occurrences of variables (V_i = 1, also

called items) or association rule. An association rule

has the form A => B where A and B are frequent

sets of items and expresses a tendency to observe B

whenever we observe A. The confidence measure of

a rule reflects the probability of observing B in a

transaction of the dataset knowing that A is also

observed in the same transaction. Confidence is

calculated empirically from the dataset itself and

estimates the aposteriori probability of B. Another

important measure that characterizes an association

rule A => B is the support that is defined as the

proportion of transactions that simultaneously

contain all items in A and B. Association rules are

automatically discovered from a dataset using

algorithms such as APRIORI (Agrawal et al., 1994)

or FP-Growth (Han et al., 2004).

Support = (freq (A U B)) / N

(1)

Where N is the number of transactions in the

dataset and freq(S) represents the number of

transactions that contain all the items in S.

confidence = (freq (A U B)) / (freq (A))

(2)

Given minimal support and minimal confidence,

these algorithms efficiently discover all rules with

support and confidence equal to or above the

required values. APRIORI works in two steps. In the

first step it discovers all the frequent itemsets (with

support at least equal to the minimum support) and

in a second step extracts from these frequent

itemsets the rules with sufficient confidence.

These algorithms easily produce a large number

of rules, so other measures of interest of the rules,

such as lift, conviction or Chi-square, are used in

addition to confidence and support measures. In

general, these measures determine to what extent the

joint observation of antecedent and consequent

happened by chance. The lift measure calculates the

quotient between the aposteriori probability of the

consequent of the rule and its apriori probability

(Bayardo et al., 1999).

Lift (A=>B) = confidence(A=>B)/support (B)

(3)

When the lift value is close to 1, it means that A

does not bring information about B and the rule is

not interesting. Even if the rule has high support or

confidence.

3 ASSOCIATION RULES IN

EDUCATIONAL RESEARCH

Data mining uncovers hidden information from data

and has been applied in a range of fields including

educational research (Educational Data Mining)

(Karkhanis et al., 2015; Mohamad et al., 2013).

Educational data mining can be a useful tool for

helping teachers to modify their teaching strategies

Studying Programming Students Motivation using Association Rules

515

and solve different educational problems and

objectives. Association Rules mining is a widely

used data mining technique. (García et.al, 2011;

Ayala, 2014). In this paper we focus on the

understanding of factors that may affect student’s

motivation. For that purpose, we use association rule

mining. The Apriori algorithm is also used in

educational data mining applications (Buldu et al.,

2010; Abdullah et al., 2011; Kularbphettong et al.,

2012).

4 STUDENTS MOTIVATION

Resulting from the study that has been made, and

published in the article (Tavares et al., 2017), we

briefly recall in this section the basic concepts on

student motivation.

For the teacher to play an important role in the

learning process that occurs in the classroom, the

teacher must have control over the external factors

that influence the behaviour and involvement of

students (Callahan, 2010). The level of motivation

needed to involve each student in a given task is

determined by his expectation for success and the

value that the student gives to that particular task.

This theory suggests that students can succeed if

they dedicate with effort and appreciate the activities

in which they enrolled. As Almeida (2012) stated,

it’s important to understand why students do not

have motivation. Many students attribute this

problem to the behaviour of the teachers and the

school in general, with the expectation that they are

active elements in their learning. To verify this

statement, we designed a questionnaire to survey

students’ actual opinion; as soon as we finish the

analysis of the collected answers we will publish the

study. On the other hand, the teacher assigns the

difficulties to the students, with the expectation that

they are interested, auto-regulated, with energy to

search for knowledge, and responsible for their own

motivation. In this way, there is a conflict between

students’ expectations, and teachers, who expect a

general behaviour distinct from that, manifested by

students (Almeida, 2012). The motivation is not only

a unitary phenomenon, which refers to the concept

of quantity. More than a lot of motivation, there are

variations in levels and motivational guidelines. In

this way, it is possible to ask what is the reason that

leads to a more or less motivated behaviour. To

reason about motivational quality it is crucial to

consider the attitudes and goals that move people

towards an action. A good example is the motivation

that compels a student to do his homework. He can

do it without any curiosity or interest, simply

looking for the approval of the teacher or parent;

but, in the other way around, he can be motivated to

acquire new knowledge, or face new challenges

because he understands that his attitude brings

advantage and values; or he can still be motivated

because the knowledge acquired will give him a

position to attain better grades or a better social life.

Motivation is like an impulse, a feeling that

moves people to act to obtain their goals. Is what

makes the individuals do their best, do what they can

to get what they want. According to the theory of

self-determination, the motivation can be intrinsic or

extrinsic. The intrinsic motivation does not need any

external factor. It derives from the student himself as

the dedication, competence, willingness and ability

to accomplish a task. Extrinsic motivation is the

result of external factors, such as the resources that

the student has, the rewards, and the environment

where it develops his tasks (Silva et al., 2014). Both

work together and the result will set the student's

behaviour, as shown in the Figure 1 (Tavares et al.,

2017).

Figure 1: Student/Teacher and Motivation.

5 ASSESSING MOTIVATION

An experiment was conducted to assess the

motivation of students in the process of learning

programming. For that purpose, we have designed a

survey (questions in appendix) and applied it to 237

first-year students of a Programming course from

different institutions.

5.1 Using Association Rules

Besides conducting a descriptive statistical analysis

of the data collected for the 237 students, we have

A2E 2018 - Special Session on Analytics in Educational Environments

516

used Association Rule mining. A preliminary study

using Apriori algorithm was carried out to

understand what kind of associations could be

inferred between the different driving factors and the

motivation. We used the 'arules' package of R (a

language and environment for statistical computing),

with the parameters support = 0.1 and confidence =

0.9. These algorithms easily produce a huge number

of discovered rules. To select strong rules that apply

to at least 10% of the students and have a confidence

close to one (Han et al., 2000). We used the lift

interest measure in addition to confidence and

support. We obtained a large set of association rules

from which we selected some that are listed in Tab.

Table 1: Results of some selected rules.

Discovered Rules

Measures

{Facing.difficulty.of.the.subject.I.fee

l.demotivated.and.I.give.up,Need.extern

al.incentives} =>

{Feel.that.motivation.decreased.

throughout.the.semester}

s=0.14

c=0.92

l=1.99

{Facing.difficulty.of.the.subject.I.fee

l.demotivated.and.I.give.up,I.would.lik

e.to.be.able.to.use.an.exercise.support

.platform.that.gives.feedback.shows.sol

utions.or.guides.resolution,Need.extern

al.incentives}=>

{Feel.that.motivation.decreased.through

out.the.semester}

s=0.14

c=0.91

l=1.98

{attribute.motivation.to.interest/

challenge.of.the.subject,Don’t.

feel.demotivated.because.didn’t.have.

feedback.from.the.exercise.marking,The.

main.reason.why.I.study.is.acquire

new.knowledge,Don’t.feel.that.motivatio

n.has.decreased.throughout.the.semester

}

=>{feel.intrinsically.motivated.to.stud

y.this.subject}

s=0.12

c=0.93

l=1.72

{attribute.motivation.to.the.importance

.of.subject.for.professional.life,

Don’t.feel.demotivated.because.didn’t.h

ave.feedback.from.the.exercise.marking,

The.main.reason.why.I.study.is.

acquire.new.knowledge,feel.intrinsicall

y.motivated.to.study.this.subject}

=>{Feel.that.motivation.don’t.decreased

.throughout.the.semester}

s=0.11

c=0.93

l=1.79

As we can see in Table 1, the difficulty of the

subject and the need for support in the study seems

to be related to the students' lack of motivation

throughout the semester (0.92 of confidence). On the

other hand, students who are aware that the subject

may be useful for their professional life, who do not

depend on feedback and who seek new knowledge

seem to be more motivated to learn (0.93 of

confidence).

In this case, the application of association rule

mining to the student’s survey data allowed us to

identify important factors that have an impact in the

level and the dynamics of the motivation of the

students throughout a semester.

5.2 Statistics

To assess the real difficulties that students feel in the

teaching / learning process of programming, we

elaborated a survey (in appendix) to 160 students of

a first-year of programming course from one

institution. Some of the questions asked and their

results can be appreciated in Table 2.

Table 2: Percentage results of some questions asked to 160

students.

Yes

Did not

answer

Do you feel that your motivation has

decreased throughout the semester?

45,6%

53,8%

0,6%

I have difficulty in understanding exactly

what is intended by the exercises and test

questions.

49,4%

50,0%

0,6%

I feel unmotivated because I do not have

feedback from the exercise marking.

44,4%

55,6%

I am in favor of the use of digital platforms

for distance learning (eLearning), Moodle

type or Blackboard.

90,6%

9,4%

I would like to be able to use an exercise

support platform that gives feedback, shows

solutions or guides resolution out-of-class.

92,5%

5,6%

1,9%

Questions focused on student motivation: If it

increased or decreased during the semester; if the

student had difficulty solving the exercises; if the

student need more feedback, if they have difficulty

understanding what the question ask, among other

questions.

It is observed (Table 2) that about 50% of

students say they are discouraged during the

semester. The same percentage says they cannot

easily understand what they are asked to do. Also, a

high percentage (about 45%) consider it to be

demotivating due to the teacher's lack of feedback in

good time.

In addition to these direct questions, it was asked

whether the students felt intrinsically motivated to

study the subject or if they needed external

incentives, and found that 42% said they needed

these incentives. This is a crucial result to justify the

ongoing concern that the teacher has to find new and

more effective ways to improve.

Studying Programming Students Motivation using Association Rules

517

Figure 2: Percentage results to the question "I feel

intrinsically to study this subjects or I need external

incentives" of 160 students.

To the question "The main reason why I study",

48% answered that it would be to acquire new

knowledge, 33% for good grades, which suggests

that students need external rewards, notably grades,

which is consistent with the results already shown in

Figure 2. Through other questions, it was also

possible to understand that students feel motivated

due to the "interest / challenge of the subject" and to

"importance of the contents for the professional

life". The reasons chosen for their demotivation

were the "difficulty of the subject" followed by a

factor "mode of functioning of the classes".

To consolidate the previous study, a further 77

surveys were made to students from three different

courses in different school years from two different

Higher Education Institutions. The results

corroborated the conclusions presented for the first

sample, as can be seen in Table 3 and Figure 3.

After the conventional analysis of all the data

collected for the 237 respondents, a preliminary

study was made with Data Mining techniques in

order to understand what kind of association rules

could be inferred between the different factors and

the motivation.

Table 3: Percentage results of some questions asked to 77

students.

Yes

Did not

answer

Do you feel that your motivation has

decreased throughout the semester?

47%

51%

I have difficulty in understanding exactly

what is intended by the exercises and test

questions.

53%

45%

I feel unmotivated because I do not have

feedback from the exercise marking.

35%

60%

I am in favor of the use of digital platforms

for distance learning (eLearning), Moodle

type or Blackboard.

60%

39%

I would like to be able to use an exercise

support platform that gives feedback, shows

solutions or guides resolution out-of-class.

87%

12%

Figure 3: Percentage results to the question "I feel

intrinsically to study this subjects or I need external

incentives" of 77 students.

6 CONCLUSION

In the context of a project looking for computer-

based strategies to aid on teaching/learning

Programming, we started by identifying some

factors that justify the difficulties felt by students

that lead them to fail in the Programming Courses.

Then we have studied the intrinsic and extrinsic

factors that impact on the Humans’ motivation. That

research led us to proposed two approaches based on

the combination of animation techniques with auto-

evaluation systems to engaged students in the

practical exercises of Programming courses aiming

at the improvement of their learning. To assess the

proposal some experiments were designed and

conducted. In that context we decided to inquiry

students about their own feelings concerning their

motivation and behavior along those courses. This

paper discussed the exploration, through the use of

Association Rules, of the gathered data to quantify

the relationship among the different parameters

involved in our research hypothesis. The results

extracted from that data analysis corroborate our

assumptions: most of the students face big

difficulties in the beginning of the course and as they

are not intrinsically motivated for that area they gave

up if they are not supported by immediate feedback

and tools that make their engagement more fruitful

and appealing.

In the future we will conduct more experiments,

and collect more data concerned with students’

feelings in order to apply again those Association

Rules and other data mining approaches that will

rigorously quantify our thesis about the ways we can

increment students’ motivation to learn more

effectively the science of programming computers.

A2E 2018 - Special Session on Analytics in Educational Environments

518

ACKNOWLEDGEMENTS

This work has been supported by COMPETE:

POCI-01-0145-FEDER-007043 and FCT –

Fundação para a Ciência e Tecnologia within the

Project Scope: UID/CEC/00319/2013.

REFERENCES

Abdullah, Z., Herawan, T., Ahmad, N., Deris, M., 2011.

Mining significant association rules from educational

data using critical relative support approach. In

Procedia - Social and Behavioral Sciences, Volume

28, pp. 97-101, ISSN 1877-0428,

https://doi.org/10.1016/j.sbspro.2011.11.020.

ACM/IEEE, 2013. Computer Science Curricula 2013 --

Curriculum Guidelines for Undergraduate Degree

Programs in Computer Science, Final Report.

Agrawal, R., Srikant, R., 1994. Fast Algorithms for

Mining Association Rules. In VLDB-94.

Almeida, D., 2012. A motivação do Aluno no Ensino

Superior: um estudo exploratório. Dissertação

apresentada ao programa de Mestrado em Educação da

Universidade Estadual de Londrina.

Bayardo R., Agrawal R., 1999. Mining the Most

Interesting Rules. In Proc. of the 5th International

Conference on Knowledge Discovery and Data

Mining, KDD-99, pp. 145-153.

Buldu, A., Üçgün, K., 2010. Data mining application on

students’ data. In Procedia - Social and Behavioral

Sciences, Volume. 2, Issue 2, pp. 5251-5259, ISSN

1877-0428,

https://doi.org/10.1016/j.sbspro.2010.03.855.

Callahan, M., 2010. How Do I Motivate My Students?

Teaching Resources, Texas Tech University.

García, E., Romero,C., Ventura,S., Castro, C., 2011. A

collaborative educational association rule mining tool.

In The Internet and Higher Education, Volume 14,

Issue 2, pp. 77-88, ISSN 1096-7516.

Han, J., Kamber, M., 2000. Data Mining: Concepts and

Technique. Morgan Kaufmann Publishers, 2000.

Han, J., Pei, J., Yin, Y., Mao, R., 2004. Mining Frequent

Patterns without Candidate Generation: A Frequent-

Pattern Tree Approach. Data Min. Knowl. Discov. 8, 1

(January 2004).

Hundhausen, C., Douglas, S., 2000. Using Visualizations

to Learn Algorithms: Should Students Construct Their

Own, or View an Expert’s? Proceedings 2000 IEEE

International Symposium on Visual Languages IEEE

Computer Society Press, Los Alamitos.

Karkhanis, S. P., Dumbre, S.S., 2015. A Study of

Application of Data Mining and Analytics in

Education Domain. Int. Journal of Computer

Applications,Volume.120, No.22, pp. 23-28.

Kularbphettong, K., Tongsiri, C., 2012. Mining

Educational Data to Analyze the Student Motivation

Behavior. World Acad. Sci. Eng. Technol., Volume. 6,

no. 8, pp. 1036–1040.

Kumar, V., Chadha, A., 2012. Mining association rules in

student’s assessment data. International Journal of

Computer Science Issues, 9(5), pp. 211-216.

Mohamad, S. K., Tasir, Z., 2013. Educational Data

Mining: A Review. In Procedia - Social and

Behavioral Sciences, Volume.97, pp. 320-324,

ISSN1877-0428

Peña-Ayala, A., 2014. Educational data mining: A survey

and a data mining-based analysis of recent works.

Expert Systems with Applications, 41(4), pp. 1432–

1462.

Proulx, V., 2000. Programming patterns and design

patterns in the introductory computer science course.

Proceedings of the thirty-first SIGCSE technical

symposium on Computer science education, pp. 80-84.

New York.

Ramos, S., 2013. Motivação Académica dos Alunos do

Ensino Superior. Psicologia.pt, O portal dos

Psicólogos.

Ryan, R., Deci, E., 2000. Self-Determination Theory and

the Facilitation of Intrinsic Motivation, Social

Development, and Well-Being. American

Psychologist, Vol. 55, pp. 68-78.

Santos, R., Costa, H., 2006. Análise de Metodologias e

Ambientes de Ensino para Algoritmos, Estruturas de

Dados e Programação aos iniciantes em Computação e

Informática. INFOCOMP – Journal of Computer

Science, Lavras/MG – Brasil, volume 5, n. 1, pp. 41-

50.

Silva, T., Mascarenhas, I., Medeiros, C., Sousa, E., 2014.

A Motivação no Ensino Superior: Um Estudo com

Alunos dos Cursos de Administração e Direito.

Journal of Management Analysis, Volume 3, pp.104-

113.Fortaleza.

Tavares, P., Henriques, P., Gomes, E., 2017. A Computer

Platform to Increase Motivation in Programming

Students – PEP. 9th International Conference on

Computer Supported Education (CSEDU 2017).

Porto, Portugal.

Williams, K., Williams, C., 2011. Five Key Ingredients for

Improving Student Motivation. Research in Higher

Education Journal, pp. 104-122.

Studying Programming Students Motivation using Association Rules

519

APPENDIX

A2E 2018 - Special Session on Analytics in Educational Environments

520