• No results found

7.3 P ART III: S URVEY OF DROPPED OUT AND PASSED STUDENTS

7.3.2 Difficulty of programming related issues & strategies to cope with them

The respondents estimated the difficulty of the programming related issues by rating each listed item on a four-step fully anchored rating scale. In addition, the respondents were able to choose the option “I do not know”. In that case, the response was regarded as a missing value in the analysis. The results are shown in Table 28. The higher the mean value is, the more difficult the issue at hand was perceived as by the respondents.

The dropped out students experienced all aspects as more difficult than the students who passed the exercises did. The Mann-Whitney U test indicated that the differences were statistically significant in all but one case (using the text editor).

Table 28 Respondents’ estimation on how difficult programming related issues were

Passed Dropped out U

Using text editor 1.4 1.6 17181.00

Mathematics that is needed to solve the problems 1.4 1.7 18569.00 ***

Testing own code 1.6 2.1 18064.00 ***

Discovering a principle solution to the problem 1.9 2.2 21432.00 * Adopting the exactness needed in writing program 1.9 2.4 13347.00 ***

Understanding how given code is executed 2 2.6 16709.00 ***

Adopting the programming style required at the course 1.8 2.6 13347.00 ***

Identifying structures in a given code 1.9 2.7 15806.50 ***

Discovering algorithm that executes the principle solution 2.3 2.8 16960.00 ***

Finding compile time errors 2.5 2.9 17824.00 ***

Transferring own thinking into programming language 2.4 3.1 15782.00 ***

Designing parts of own code 2.4 3.1 15528.50 ***

Designing the functioning of own code 2.3 3.3 14603.00 ***

Finding run time errors 2.6 3.3 17684.50 ***

ional statements (e.g., if-else) 1.3 1.7 17673.50

Loops (e.g., while) 1.7 2.3 15085.50 ***

Methods 1.8 2.5 14664.50 ***

Ideas of OO 2.3 2.6 20058.50 ***

Table 2 2.9 10076.50 *

Exceptions 2.2 3.2 5807.00 ***

Inheritance and abstract class 2.5 3.3 10118.00 ***

Handling files 2.4 3.3 8983.50 ***

* p < 0.05, *** p < 0.001

The respondents were asked to elaborate in an open-ended question on what helped them get over the previously listed difficult issues. 62.8% (n = 243) of the students who got a grade from the exercises answered this question, whereas the response rate of the dropped out students was 37.0% (n = 44). I used a data-driven approach to analyse the open question answers. The procedure was close to what Ryan and Bernard call free list (2003). The written answers were categorised into 19 categories that emerged from the data. The procedure started by reading through the answers several times. The text quotes that expressed a strategy to get over difficult issues were placed into preliminary piles from which the categories emerged, as the process was iterated a couple of times.

Next, I looked for the common factors among the categories and ended up with six strategies (or meta-categories) that the students used to get over a difficult issue (see Table 29) The value f in the table refers to the frequency of how often the strategy at hand was mentioned.

Table 29 Strategies to get over difficult issues

Passed Dropped out

Category f % f %

Ask for help Assistant teachers 62 27 6 14

Friends 45 20 10 23

Total 107 47 16 36 ***

Study & think Educational material provided by the teacher 42 18 6 14

Internet 23 10 2 5

Lectures and slides 21 9 6 14

Literature 11 5 1 2

Code examples found from different sources 8 3 1 2

Course news group 7 3 1 2

Studying, seeking answers (general) 24 10 12 27

Thinking, trying to understand 14 6 0 0

Total 150 66 29 66

Practice Doing exercises, practising 31 14 5 11

Trial and error 7 3 3 7

Total 38 17 8 18

Persistence Persistence & perseverance 38 17 3 7

Compulsory course 5 2 1 2

Total 43 19 4 9

Manage time Allocating more time for studying 18 8 4 9

Talking a pause 1 0 0 0

Total 19 8 4 9

Interest &

experience

Interest towards programming 4 2 0 0

Previous programming experience 5 2 0 0

Total 9 4 0 0

Nothing Nothing helped 2 1 9 20

*** p < .001

Asking for help either from the assistant teacher or a friend was a much-used strategy among all students.

“... My friends were a great help. I did not think that the lectures helped me a lot. Together, we tried to understand things and we helped each other. If someone understood something, he then helped others.

The next category is studying hard using different information sources (like course material, internet, and course newsgroup) and thinking about the difficult issue with the

intention of really understanding. Many respondents stressed that they read the material over and over again until they understood. The following respondent uses his/her old programming exercise as a learning material.

“… sometimes I went through my programming exercises after I had got maximum points from them. I thought over what I had actually done, what happens where and why.”

The third strategy is practicing programming by doing exercises.

“The greater number of the programming exercises I did, the easier the problems I faced at the beginning of the course seemed to become. Difficulties I had at the last part of the course did not seem to be insuperable anymore like the difficulties I faced at the middle of the course. I started to internalise entities and discern Java’s structure. The amount of practice was essential for getting over difficulties.”

“Great programming exercises! For example, a constructor-method thing got cleared up when I did the program that mixed drinks. The instructions for the exercise clearly implied that you use a constructor to create an empty class and you use methods to fill it with different drinks and to drink it.”

A subcategory of this third category is “getting over a difficult issue by doing exercises applying a trial and error approach”.

The fourth category is persistence. The respondents reported that they had stamina and determination to work with the difficulties. They were self-disciplined and worked hard because they did not want to quit or the course was compulsory and they did not want to enrol in the course once again.

“Trying. I had taken a stand that I will do all the assignments. I forced myself to do them [exercises] and banged my head against a brick wall and I managed to do the exercises.”

The fifth category deals with time. Allocating more time for studying even at the expense of other courses helped the respondents understand difficult issues. One respondent also reported that taking a pause (like a good night's sleep) was helpful in some cases.

The last strategy deals with the personal interest in programming and previous experience in programming. In addition, especially the dropped out students reported that there was nothing that helped them when they faced difficulties and therefore they dropped out. The dropped out students’ low response rate (37%) to this question could also be interpreted as a statement (if the student thought there was nothing that helped, he/she may have left the question unanswered).

The results suggest that the students who passed the programming exercises mentioned getting help, being persistent, and holding interest in the subject as viable strategies to get over difficult issues. However, the differences between the frequencies of mentions are in most cases not statistically significant or the significance was not calculated due to the low number of frequencies. The only exception is the “Asking for help” -strategy, which the students who passed mentioned statistically significantly more frequently to be helpful than the dropped out students (

!

"2 (2) = 3.0E2, p < 0.001). On average, both dropped out and passed students mentioned 1.6 viable strategies to get over a difficult issue during the course.

7.3.3 Reasons for dropping out

The students who dropped out of the course were given a list of the possible reasons for dropping out and they were asked to rate how much each reason affected their decision

to drop out. The four-step rating scale was fully anchored. In addition, the respondents were able to choose the option “I do not know”. In those cases, the response was recoded so that it equalised the neutral response. In Table 30 the low values represent low significance to the decision to drop out (1 = no effect at all) whereas high values represent critical reasons in the decision to drop out (4 = critical effect). The respondents rated the course workload-related issues as the ones affecting their decision to drop out the most. On the other hand, in general, being caught for plagiarism or plans to continue studies at a different university played a minor role in the decision to drop out.

Table 30 Reasons for dropping out

mean

Doing exercises took too much time 3.4

Course required more time than other courses at the same study week quantity 3.2

Course's workload is not in balance with the payback 2.9

I did not know how to do programming exercises 2.8

I had reserved too little time for the 5 study week course 2.6 I did not understand the content that was covered at the course 2.3

I did not get enough help 2.3

Low motivation to study in general 2.2

Programming does not interest me 2.2

I wanted to concentrate on other courses 2.2

Dropping out does not affect my other courses 1.9

I had a lot of other courses at the same time 1.9

Course personnel's actions 1.7

Course arrangements 1.7

Personal reasons 1.7

Work related commitments took time from the course 1.6

Hobby related commitments took time from the course 1.6

This course is not obligatory for me 1.4

I am not going to stay at TKK 1.3

I had no intention to pass the course in the first place 1.2

I decided to start preparing for entrance exams 1.2

I got caught from plagiarism 1.2

The interviews during the previous parts of the research project revealed that students tended to have several concurrent reasons for dropping out; the quantitative survey corroborates that argument. On average, the respondents reported 10 reasons that had contributed to their decision to drop out at some level. Moreover, the respondents reported, on average, four reasons for dropping out that affected their decision critically.

In order to get a clearer picture of the drop-out phenomenon I decided to use factor analysis to reduce the variables. The procedure and pre-tests are explained in chapter 6.1. A Generalized least squares factor analysis followed by Varimax rotation reduced the list of drop-out variables to five factors, which explained 46% of the total variance.

The first factor explained 10.5 % of the variance, the second 10.1%, the third 9.3 %, the fourth 9.0%, and the fifth 7.1% of the variance. The chosen model fit the data well (

!

"2 (115) = 165.62, p < 0.001).

The variables that loaded high on the first factor dealt with course personnel’s actions and students’ perception of not getting enough help. Therefore, F1 was named Course

arrangements & not help. The second factor had high loadings from the variables that considered difficulties understanding the course content. Therefore, the second factor was named Difficulties in understanding course topics. The third factor discusses the Time management issues and students’ preferences for using time. The fourth factor had high loadings from variables like “The course is voluntary for me”, and “Dropping out of the course does not affect other courses”. Thus, the factor was named Dropping out does not have consequences. The fifth factor was named Prefers other courses. All factors and variables with the loadings are summarised in Table 31.

Table 31 Factor analysis’ results of reasons for dropping out

Factors

F1: Course arrangements & not help (Cronbach’s α = 0.83)

• The course personnel’s actions (.86)

• Did not get enough help (.76)

• Course arrangements (.74)

F2: Difficulties in understanding course topics (Cronbach’s α = 0.75)

• Did not understand subjects that were covered in the course (.99)

• Did not know how to do programming exercises (.74)

• Programming is uninteresting (.41)

F3: Time management and preferences (Cronbach’s α = 0.40)

• Programming exercises took too much time (.71)

• The course required more time than anticipated (.64)

• I had reserved too little time for the course (.40)

• Personal/family reasons (-.36)

• Work related commitments (.30)

• Course workload is not balanced with the payback (.34)

F4: Dropping out does not have consequences (Cronbach’s = 0.65)

• The course is voluntary for me (.74)

• Dropping out of the course does not affect other courses (.63)

• There was no intention to pass the course in the first place (.60)

• I am not going to continue studies at TKK (.50)

• I decided to use the time for preparing for entrance exams to continue studying elsewhere (.42)

F5: Prefers other courses (Cronbach’s α = 0.71)

• I had a lot of other courses at the same time (1.0)

• I wanted to concentrate on other course (.57)

7.3.4 Summary of the passed and dropped out students’ questionnaire