ARE DRILL-and-practice programs for everyone?

Piret Luik, Triin Marandi

University of Tartu, Department of General Education, Estonia

Paper presented at the Association for Teacher Education in Europe (ATEE) conference, Changing education in a changing society, University of Riga (Latvia), 2-3 May 2003

Abstract

This study explores appropriateness of drill-and-practice programs for several students. We wanted to determine how students’ learning styles, gender and attitudes affect on effectiveness of drill-programs. The results are based on the experiment, which was carried out with 8-9 year old students in Estonian school. Coefficients of Spearman rank correlation between the characteristics of students and test scores were calculated. Main results of 2 months lasted experiment were that with drill-programs girls benefit more than boys, this kind of software is more suitable for visual learners and for students who do not take learning with computer program as a competition. Also we found out that learning outcome in experimental group was related with students’ computer skills. But there was not any evidence that learning outcome was related with computer anxiety.

Key words: math, elementary school, drill-and-practice programs, learning styles, gender

Introduction

Drill-and-practice programs are widely used for differentiation and individualisation of learning. This kind of software is mostly used in schools, especially in elementary level obtaining basic skills (Rattanapian & Gibbs, 1995). Positive aspects of drill-programs are patience; intensification; individualisation and corrective and immediate feedback. High ability students do not feel bored and do not lost attention because low ability students need considerably more time for obtaining learning skills. Low ability students do not disturb because the others are more progressive and they do not feel themselves inefficiently (Behrmann, 1984).

Different studies have shown the usefulness of drill and practice programs for learning basic math skills, foreign languages (Behrmann, 1984; Clements, 1999; Van Scoter et al., 2001). Even if many researches have shown that computerised drill is better than traditional repetition, the effect of using electronic drills depends on different factors.

In our study the subject taught with computers was math. Mathematics is always been domain of men. Several studies have found out that boys outperform girls in mathematics (Hood & Togo, 1993/94). Studies about educational software also confirm that with educational programs about mathematics (Hativa & Shorer cit by McCoy, 1996) boys benefit more than girls. Probably this is connected with their usage of computers. Different researchers have said that boys use computers more than girls (Kay, 1992; Durndell, et al., 1995; Rattanapian & Gibbs, 1995; Young, 2000). Also boys are more interested in computers than girls in both home and school (Rattanapian & Gibbs, 1995) and have therefore higher computer skill levels than girls (Kay, 1992; Durndell, et al., 1995). Some researches have shown that in pre-school age girls are as much interested in computers as boys are, the difference arises late, in school years (NAEYC, 1996). Kay (1992), giving survey about gender differences in behaviour towards computers, reports that there are fewer differences in computer attitudes and use among pre-schoolers and primary school students than among older students.

All people gathering information by seeing, hearing and touching. But different people prefer one style of learning to other. There are several different learning style categories and scales (VAK, Gregorc, Kolb, etc). According to VAK learning styles there are three main sensory receivers – vision, auditory and kinesthetic. In different ages people can use different learning style: under 4 – kinesthetic, grades 4 to 8 – visual, up to 9 grades – auditory. Learning styles tends to move to greater abstraction (Clark, 2000). Some of scales intended for identifying learning styles of adults (the Gregorc Style Inventory) distinguish four basic learning style abilities: abstract random, abstract sequential, concrete random and concrete sequential (Ross, 1997). Kolb’s model brings out four learning styles: reflectors, theorists, pragmatics and activists (Kolb, 1984). Researches have showed significant difference between the cognitive styles on learning outcomes (Pillay, 1998, Davidson, et al., 1992). When teaching consider with individual learning style the students will learn and remember better and enjoy learning more (Dunn, cit by Cohen, 1997). Different types of learning programs support different learning styles. Studies have showed that rich hypermedia environments are accommodating various learning styles (Reed & Spuck, 1996). There are no studies answering the question which learning styles are supported by drill-programs.

We wanted to determine how students’ learning styles, gender and attitudes affect on using drill-programs. Do multimedia-based drill-programs also suitable for all learning styles and accommodate individual learner differences?

Integration of computers in our schools requires from students more and more using this kind of technology. When some students do it with enthusiasm, the other feel anxiety. Different researchers have got contradictory results compared computer anxiety of boys and girls (Temple & Lips, Rosen et al., Rohner & Simonson cit by Todman & Lawrenson, 1992). Todman & Lawrenson have also find that primary schoolchildren are less anxious about computers than the older students. According to his findings general predisposition to anxiety is not related to computers, but predicts anxieties specifically related to math (Todman & Lawrenson, 1992).

Hereby were our hypotheses:

H1: With drill-programs benefit boys more than girls.

H2: With drill-programs benefit more students whose computer skills are higher.

H3: With drill-programs benefit more students whose computer anxiety is lower.

H4: With drill-programs benefit more students with visual learning style than students with other styles (auditory and tactile styles).

H5: With drill-programs benefit more self-dependent learners.

H6: With drill-programs benefit more students for who learning with computer is not as a competition.

H7: With drill-programs benefit more students who like more computer assisted learning than traditional learning.

Methods

Sample of our experiment was children of the 3rd grade (age 8-9), because development of basic skills is most important just in elementary level. Previous researches have shown that drill and practice programs are more beneficial for low ability students and therefore we used in our research ordinary class and levelling class (accordingly A and B class). Experiment was carried out with 36 students (24 from A class and 12 from B class). Experimental and control group equated by using average marks of mathematics (in both groups 3.8). For addition we used pre-tests for controlling equality of the groups. There was not any statistically significant difference between the groups and between the boys and girls of the experimental group.

There were 10 boys and 8 girls in both groups. All participating students have previously used computer, so every student had knowledge and skills of using this kind of technology.

Instruments

We used drill-programs about math because previous researches have shown the greatest effect in sciences, and mathematics skills do not depend on language. We used 15 drill-and-practice programs. In choosing drill programs we originate from these criteria:

1. the theme should accord to 3rd grade of Estonian curricula;

2. teaching in programs should be similar to teaching in our traditions;

3. lesson should be not longer than 15 minutes for high ability students;

We used 15 different drill and practice programs about next themes: addition and subtraction (mental and in writing arithmetic), multiplying and dividing (multiplication table) and clock. For carrying out precise instructions were elaborated for teachers.

Problems for pre- and post-tests composed active teachers. Before and after the experiment students filled the questionnaires. With these instruments we tried to evaluate motivation, self-confidence and learning style of the student, feasibility and pleasantness of mathematics and math lessons, computer anxiety and computer skills.

Some researches (Davidson et al., 1992; Ross, 1997; Chuang, 1999) have shown that computer assisted learning is more suitable for students with specific learning styles, we used instrument for evaluating students learning styles. For that we used Learning Style Inventory (URL: http://rrcc-online.com/~psych/LSInvenory.html). Mostly is used for determining learning styles Gregorc Style Inventory, but it is suitable only for adults. Subjects of our study were 8-9 year old children, so we chose this instrument because it was suitable for elementary students and it was not depend on culture. Also teachers gave evaluations for two students’ characteristics: ability to learn oneself and level of co-operation.

All instruments were tested in pilot project in the May 2002. Corrections and improvements for being correctly understandable for students of elementary level were taken into account. Reliability of the tests (Cronbach’s ) was .67-.85 and validation of tests and questionnaires were checked by experts.

In addition the teacher monitored students during the experiment for fixing events connected with the treatment and changes in behaviour of the students (aggressiveness, self-confidence, relationships between students etc.). For that person who was with the students in the computer lab filled diary of the experiment.

Procedure

Before the experiment, students of experimental and control group filled pre-questionnaire, where they asked to evaluate their attitudes toward math, computer skills and anxiety. Also they completed pre-test to measure their initial knowledge.

Experimental group worked 2-3 times in the week independently with computerised drill-programs at the same time when control group practised same skills in the class with the teacher. The first stage of this experiment lasted 6 weeks (from September to October 2002). After the first stage executed both group tests and questionnaires.

Methods of Statistics

For data analysis was used statistical package StatSoft STATISTICA 6.0. To find out characteristics of students, which have influence on learning outcomes in experimental group, we calculated coefficients of Spearman rank-correlation between the evaluations of students and their learning results.

We assigned identifier 1 to girls and identifier 2 to boys.

Results

The average result of the pre-test in experimental group was 62% and average score of the test after the first stage was 68%. Correlation coefficient between the two tests was r=.72 (p<.01). With t-test we found out that there was statistically significant difference between these two tests (p<.001).

There was a significant relationship between gender and the test of the first stage (r= -.50; p<.05). Girls benefited more than boys with the drill-programs.

Students’ computer skills were separated into subgroups: familiarity of the keyboard, typing skills, understanding computer messages, mouse using skills, and ability of start-up programs. Students’ computer anxiety was also evaluated with questionnaires. Correlation coefficients with these characteristics are given in Table 1. None of these skills depended on gender.

Table 1. Correlation coefficients between computer skills and test scores of experimental group.

Characteristic / Correlation coefficient with the test of the first stage
Familiarity of the keyboard / .51 *
Typing skills / .55 *
Understanding computer messages / .49 *
Mouse using skills / -.09
Ability of start-up programs / .56 *
Computer anxiety / -.42

* Statistically significant at the .05 level

With learning styles scale we assessed each students’ learning style: visually, auditory or tactile. Also we assigned is the student independent learner or needs somebody’s help and support. Correlation coefficients with these characteristics are given in Table 2. We found out the relationships between gender and learning styles. Boys were more with the tactile learning style (r=.48; p<.05) and girls used more auditory learning style (r= -.48; p<.05).

Table 2. Correlation coefficients between learning styles and test scores of experimental group.

Characteristic / Correlation coefficient with the test of the first stage
Visual learning style / .52*
Auditory learning style / .28
Tactile learning style / -.79 **
Self-dependency of students / .80 **

* Statistically significant at the .05 level

** Statistically significant at the .01 level

With questionnaires we evaluated also did the students take the learning with computer as a competition. There was not any statistically significant difference between this characteristic and gender. Spearman correlation coefficient between the taking drill-program as a competition and test score was -.53 (p<.05).

Because some students like more traditional learning and some students like innovational learning like computer assisted learning, we asked students to evaluate which learning method is better for them. This characteristic did not depend on gender and it had no influence on test outcome (p<.05).

Discussion and conclusions

The most important result was that computerised drills were more effective for students who do not take computer program as a competition. Computer assisted instruction is learning, not game or competition with someone. Apparently students, who took this kind of learning as a competition paid too much attention to score. Their goal was not learning, but winning.

Also we found out that learning outcome in experimental group was related with students’ computer skills, especially familiarity with keyboard. When student entered accidentally the wrong answer, she/he would get frustrated because she/he knows the right answer, but computer says that the answer was wrong. Used software required mostly typing skills and thereby the mouse using skills were not so important. But important skills were also understanding of computer messages and ability to start-up programs.

We found out that there was not any evidence that computerised drill suits better for boys like some researches (McCoy, 1996) have shown. In our research we found out that drill-programs were more suitable for girls. Maybe girls are more dutiable and they work with computer program more intensively than boys. Also in teachers’ dairies we found out that some programs’ features attracted boys and lead their attention away from learning goal.

Also there was not any poof that learning outcome was related with computer anxiety and CAI preference. Some researches (Todman & Lawrenson, 1992) have found out that computer anxiety is a problem for adults and for older students. Participants of our experiment were 9-10 years old children and apparently they feel themselves with computers confident. In pre-questionnaires none of our students wrote that she/he feels oneself less or more unconfident.

Mostly, a computer is a visual medium. Thereby in our experiment benefited more visual learners. Some used drill-programs allowed also sound, but none of the programs was appropriate for tactile learners. Previous researches have studied mostly computer based instructional materials and environments. These results (Pillay, 1998; McKay, 1999) have shown that text-based materials are more suitable for verbalizers (auditory learning style). Our experiment with drill-programs did not support this result.