Performance and attitudes of students taught using traditional and online methods.

Emilio A. Laca, Barbara Sommer, Harry Matthews.

Abstract

Institutions and teachers are challenged by the need to provide hands-on laboratory training to more students, without increasing the physical laboratory space and while maintaining quality of instruction. We evaluated online tools for teaching and testing in relation to the traditional highly interactive face-to-face instruction in an introductory course about microcomputers in agriculture.

The 3-unit course is organized into a lecture and a laboratory component. The laboratory is the main component of the course, and it is characterized by frequent one-to-one interactions between students and teachers. The student/teacher ration is 6-8:1. The online alternative was developed and tested during two quarters, Fall 2000 and Winter 2001. During Fall 2000, two online Access modules were developed and used in a test involving a large proportion of the enrolled students. In Winter 2001, all laboratory modules were offered exclusively online to 12% of the students, and all students were tested before and after they received instruction on two Access modules. In Fall 2000 there were 267 students involved in the study, whereas in Winter 2001 the number increased to 291 due to the addition of the online section. Online laboratory modules were delivered using WebCT. Online testing was done with WebCT and SkillCheck.

No differences in performance due to teaching method were detected in any of the performance measures used in Fall 2000. In 2001, online and traditional groups did not differ in performance when students were taught and tested in exactly the same traditional manner (midterm), or when they were taught in different manner (online vs. traditional) but tested in the same traditional (final test) or online (Access). However, the traditional group performed better than the online group when both teaching and testing methods were different between groups (homework and quizzes, P<0.01).

On average, students clearly preferred the traditional instruction method. Yet, the majority of students ranked the methods equally, and a significant proportion of students always ranked the online method higher than the traditional method. Efforts should focus on improving the presentation and attitude elicited by online instruction by incorporating the aspects of face-to-face learning that cause the more positive attitude. Online testing methods must be modified so students can focus on the content, instead on the medium and format of the test.

Introduction

Online teaching and testing methods have the potential to improve the efficiency of teachers and reduce the need for costly laboratory and classroom space on campus. Because of the critical role of computer skills in today’s society, both teaching institutions and business corporations have a great demand for courses in basic computer skills. Several publishing and software companies have developed online systems to teach and test skills on use of computers. These systems differ significantly in format and medium from the traditional face-to-face teaching method, and they differ from other online teaching situations in that the teaching tool and medium is itself the subject of the learning. Students are expected to learn about computers by using computers.

The demand for education about computing is large in the field of agriculture. Johnson, Ferguson, and Lester (2000) reported that students enrolled in upper-division agriculture courses at a land-grant university tended to possess a low level of computer knowledge. At the same time, a survey of faculty members in the college of agriculture showed that faculty intended to maintain or increase the number of required computer tasks in ensuing years (Johnson, Ferguson, Vokins, & Lester, 2000). In a course titled “Computer Applications in Agriculture” Wingenbach (2000) found that online examinations were associated with a decline in performance, relative to that shown in traditional paper-and-pencil tests. The effect was attributed in part to lack of familiarity with the e-mail medium and resultant anxiety. These studies, specific to the agricultural context, mirror those of other academic departments in that an increasing reliance on computers is clearly part of the future. Students need to learn how to use computers efficiently and effectively.

Paralleling the curricular need for computer education is a dramatic increase in enrollment at many institutions of higher education. Computing is not a subject readily taught in a large lecture environment. In most cases, it requires a more intensive hands-on laboratory setting. Institutions and teachers are challenged by the need to provide hands-on laboratory training to more students, without increasing the physical laboratory space and while maintaining quality of instruction. This paper describes the process and outcome of creating online modules for teaching basic computer skills within an agriculture-based curriculum. Our goal was to evaluate and compare online teaching and testing as tools used in combination with highly interactive face-to-face instruction.

Methods

Traditional course

Application of Microcomputers in Agriculture presents the fundamental applications of computing in an agricultural context. It enrolls approximately 270 students per quarter, three quarters per year. Despite its size, each quarter, dozens of students are turned away because of limited laboratory space. Although the course is designed for lower division students, most of the students are juniors and seniors due to the increasing enrollments and a backlog in those who can be accommodated.

The 3-unit course is organized into a lecture and a laboratory component. A faculty member lectures to the entire class once or twice a week (15 lectures per quarter), and covers the basic aspects of how the hardware and software components of computers work, along with a discussion of the history of computers, basic aspects of software, and personal and social implications of computing.

The highly structured laboratory is the main component of the course. There are nine laboratory groups. Each group is comprised of 30 students and meets once per week. Lab instruction covers Microsoft Office 2000, Access 2000, web-based electronic mail, Netscape, HTML and Windows OS.

The staff that delivers laboratory instruction is extensive and hierarchical. The staff is comprised solely of undergraduate students who are organized into categories: interns, junior teachers, teachers, and head teachers. They are the key to the success of the laboratory.

Interns are selected from the pool of students who took the course in the previous two quarters. The faculty member in charge, head teachers, teachers, and junior teachers participate in the selection process. They look for personable students who can explain the material learned and who will probably be around for more than a year. Interns are unpaid volunteers, and get experience in teaching by helping teachers during labs and office hours, and by grading assignments. They have considerable one-to-one interaction with students in the lab. There are 10-12 interns every quarter.

Junior teachers are selected from the previous quarter’s interns by interviews that involve a simulated teaching session and several general questions about teaching and interacting with students. Teachers, including head teachers, and the faculty member in charge make the selection based on the results of the interview and the performance of the candidate as an intern. The 8-9 teachers and junior teachers are employed by the university and have a 25% appointment as Readers.

Head teachers are nominated by the previous head teachers and selected by the faculty in charge, usually one quarter ahead of time so the future head teachers have an entire 10-week quarter for training before taking the position. Head teachers are employed by the university and have a 50% appointment as Readers. In addition to teaching as the other teachers, head teachers plan and manage all activities of the interns and teachers, in consultation with the faculty member in charge.

The lab material is organized into eight modules, each following the same structured set of steps. Each lab section has a head teacher, who presents the materials, and 4-5 assistants who answer the questions of individual students on a one-to-one basis during the presentation of the material. First, students turn in the assignments from the previous week. Second, students take a quiz on the material covered the previous week by answering questions on paper and by performing actions on their computer and showing the results to a teacher or assistant, who immediately assigns a grade. Third, each student works individually at a computer station and performs all steps of the procedure being demonstrated. Finally, students complete the day’s exercise and receive an assignment for the following week. A student/reader ratio of 6-8:1 allowes each individual student to take the quiz while having frequent direct interactions with and feedback from a reader. All testing and grading takes place by hand. Even when the questions involve tasks to be completed on a computer, the answers are checked and graded directly by a teacher or intern, by going to a student’s station upon being called and checking the results on the screen.

Homework, and quizzes are almost exclusively based on information taught during the labs, and they require intensive exchange of information between students and instructors. Homework is assigned in printed form and reinforced verbally at the end of each meeting. Most homework is graded during one-to-one interactions between a reader and each student. Quiz instructions are given in printed form immediately before homework is graded. Students answer questions and perform tasks on the computer. Then, they call a reader to check and grade their work.

Only undergraduate students are present in the lab during instruction to create a non-threatening environment that fosters interaction between teacher and student. The presence of several teachers and interns provides a high teacher-to-student ratio. Grading for the course is based on 7 quizzes, 11 assignments, a midterm, and a lab final exam.

Online alternative

As an alternative to the in-class laboratory, the first author designed a series of web-based presentations and used WebCT to:

1. take the students out of the physical laboratory space;

2. mirror the face-to-face traditional labs such that no changes in structure or content would be needed;

3. deliver the material through a web browser;

4. provide an automated interface for students to obtain and submit assignments;

5. give rich visual and auditory information; and

6. manage student and grade records in a web database.

The online alternative was developed and tested during two quarters, Fall 2000 and Winter 2001. During Fall 2000, only two online modules were developed and used in a preliminary test involving a large proportion of the enrolled students. In Winter 2001, all laboratory modules were offered exclusively online to 31 students.

Fall 2000

The instructor and one assistant recorded presentations for the two lab modules that cover MS Access (workshops 7 and 8), the last two modules for the quarter. The lab modules were partitioned into smaller logical subunits based on the written lab manual. The instructor explained each subunit of each lab as if in front of an audience and using the computer screen as the demonstration screen. The session was recorded with Camtasia software (http://www.techsmith.com/) and then exported to QuickTime 4 Pro (QT4) where it was edited and briefly cleaned. Each subunit was then compressed with QT4 and inserted in a web page using Dreamweaver. An additional web page was made with the text corresponding to the movie and a link to the page that contained the movie including instructions to open the movie in a different window. Then, the set of pages were organized into the two labs by including them into the WebCT course into a Content Module. Each lab was one module and contained 8-20 subunits. Each subunit contained one web page and one movie that ranged in length between one and four minutes. The first file compression codec used made files that were 20-30 MB in size. Better compression was achieved later when the UCD media unit produced the final movies for the full test.

Students, all of whom were taking the traditional course, were offered extra credit for completing either workshop 7 or 8 online. At the end of the quarter, two to three weeks after being taught modules 7 and 8, students completed a multiple-choice written evaluation and survey that was administered with the final exam.

In the online labs, assignments were given in written form only, both in a printed booklet and on the web site. Students downloaded the necessary files, did the homework, and then uploaded and submitted the assignment through the computer. Quizzes were administered in a similar manner. The online group answered multiple-choice questions and performed tasks on the computer using SkillCheck, a computer-based testing program that simulates the applications taught in the course. SkillCheck automatically recorded which questions were correctly or incorrectly answered and assigned a percentage correct. All grades during the course were given as a percentage, and a single letter grade was assigned at the end of the quarter.

Winter 2001

In the winter quarter of 2001, all eight workshops were professionally produced by the UCD media unit, following the same structure described above for Fall 2000. Online labs were organized as above, but the grouping of students was different.

In the winter 2001, class enrollment was increased by about 12% through the creation of a new online lab section with 31 students. These 31 students had access to the online resources and workshops and had exactly the same opportunity to contact and interact with the teachers as the rest of the students in the traditional course, with the exception that they were not allowed to attend any face-to-face lab sessions. The online lab section did not have any scheduled meetings with the teachers other than lectures and one mid-term meeting with the faculty in charge. Two undergraduate readers were assigned to interact with online students by email and by having access to their records in the WebCT database.

We offered all students the opportunity to get extra credit by taking a test before and after taking workshops 7 and 8. The test was administered using SkillCheck, and students determined themselves when they took the tests and the workshops. The actual temporal sequence of tests and instruction followed by each student was determined based on web and laboratory attendance records.

Student performance in 2001 was evaluated for four different parts of the course. First, the midterm (MT) tested their knowledge about history of computers and hardware. Both the online and the traditional groups received the same instruction through common lectures and reading assignments, and they were both tested in the same manner with a traditional multiple-choice written test. Second, quizzes and homework assignments (QZ&HW) evaluated computer use skills. The online group was taught and tested on the use of computers by interacting with computers through a web browser or SkillCheck, whereas the traditional group received face-to face instruction and testing by readers. Third, performance in workshops 7 and 8 was evaluated by administering tests before and after the students went to the laboratory or viewed the online materials. The online group was taught online and tested by computer using SkillCheck, whereas the traditional group received face-to-face instruction in the traditional labs and was tested by computer with SkillCheck. Fourth, in the final exam (FNL), readers tested computer skills of students from both the online and traditional groups in the usual face-to-face interactive fashion, regardless of the teaching method students had experienced.

Student attitudes

In Fall 2000 we asked the students to indicate their agreement with the following statement as an indicator of their assessment or perception of the teaching method:

Regardless of whether you did any workshops online, give the ratings based on your experience in workshop ___. Select the best option based on the statement: The material was presented at a good pace and I was able to follow it.
A. Strongly agree.
B. Agree.
C. Not sure.
D. Disagree.
E. Strongly disagree.

Based on the degree of agreement with the statement that the workshop was presented at a good pace and that the student was able to follow it, we ranked the response from 5 (strongly agree) to 1 (strongly disagree). The difference between ranks given to workshops 8 and 7 was analyzed as a function of the workshop that was studied online using a contingency table and a c² statistic.

In Winter 2001 we evaluated students’ perception of the method of teaching by surveying their agreement with the following statement:

The contents and materials of workshop ___ were presented in a manner that was accessible and understandable.
A. Strongly agree.
B. Agree.
C. Not sure.
D. Disagree.
E. Strongly disagree.

The ranks given by students to each workshop were compared between the online and traditional groups by Wilcoxon non-parametric tests. In each quarter we asked only about the last two workshops. These questions were interspersed among other questions of the test.

Analyses

Performance scores for workshops 7 and 8 in Fall 2000 were analyzed by ANOVA as a function of teaching method and using performance in the lecture part of the course as covariate. Scores for the midterm test, combined quizzes and homework, and final test in Winter 2001, were transformed to achieve normality of residuals and analyzed by ANOVA as a function of teaching method, class, cumulative GPA in the previous quarter and interactions. Scores for workshops 7 and 8 in Winter 2001 were analyzed by ANOVA as a function of the method used for learning the material (online vs. traditional). We analyzed the scores before the exposure to the material and the difference in score (“after” minus “before”) separately.

Results and Discussion

Student characteristics

Table 1. Percentage of students in each major and total number of students in each quarter of the study.

Major	Fall 00	Winter 01
Agric. & Managerial Economics	30	18
Animal Science & Management	14	18
Wildlife, Fish. & Conservation Biol.	5	6
Biotechnology	4	6
Biochemistry	1	4
Biological Sciences	4	2
Exploratory Program	1	3
Undeclared-Life Sciences	3	2
Economics	1	3
Environ. Biology & Management	1	3
Community and Regional Develop	1	2
Textiles & Clothing	2	2
Communication	0	3
Human Development	1	2
Psychology	1	2
Environmental Toxicology	2	1
Undeclared-Physical Sciences	2	1
Other	25	22
Total number of students	267	291

Fifty-two majors were represented in Fall 2000 and 55 in Winter 2001. Most of the students (more than 55%) in each term were in animal science, agricultural economics, biology, or closely related fields.

Most of the students who took the course in Fall 2000 were junior and seniors. The class distribution was not different (P>0.60) among the three groups that took workshop 7 or 8 by computer-delivered instruction (online), or both by face-to-face (traditional) interaction with teachers in a classroom (Table 2).

In Winter 2001, a higher proportion of students were freshmen and sophomore in the online than in the traditional group (P<0.01). This is relevant, because student performance in the homework, midterm and final tests was significantly lower for freshmen. However, correction for cumulative GPA by analysis of covariance also corrected for class differences.

Table 2. Percentage of students exposed to each teaching method in each class in Fall 2000. The column and row labeled count show the total number of students.

Method for workshops 7 & 8	Freshmen	Sophomores	Juniors	Seniors	Count
7 online - 8 traditional	6	23	38	33	52
7 traditional - 8 online	2	23	48	27	52
Both traditional	4	15	49	32	162
Count	10	49	124	83	266

Table 3. Percentage of students exposed to each teaching method in each class in Winter 2001. The column and row labeled count show the total number of students.

	Freshmen	Sophomores	Juniors	Seniors	Count
online	16	26	39	19	31
traditional	4	10	49	37	259
Count	15	34	140	101	290

Student performance

Fall 2000 Access workshops

No differences in performance were detected among groups of students in any of the performance measures used for workshops 7 and 8 in Fall 2000 (Figure 1, P>0.40). Cumulative GPA, midterm performance and overall performance in the course did not differ significantly among groups, regardless the method used to learn workshops 7 and 8. In addition, their performance in the material covered in workshops 7 and 8 did not differ either. This indicates that the groups were similar in pre-test performance, and performed equally well with the different teaching methods.

Figure 1. Effect of teaching method on performance in two different workshops about Microsoft Access during Fall quarter of 2000. Students were divided into three self-selected groups: online-traditional learned workshop 7 by the online method and workshop 8 face-to-face (traditional); the sequences for the other two groups is indicated using the same terms. Dark bars are scores for the test of performance in workshop 7; light bars are scores for performance in workshop 8. Maximum scores possible were 11 for workshop 7 and 5 for workshop 8. Thin lines represent the standard errors.

Winter 2001 performance

Online and traditional groups did not differ in performance (Figure 2) neither when they were taught and tested in exactly the same traditional manner (midterm) nor when they were taught in different manner (online vs. traditional) but tested in the same traditional way (final test) (P>0.05). However, the traditional group performed significantly better than the online group when students were both taught and tested in different ways, as was the case in the homework and quizzes (P<0.01). Effects on performance of class by GPA, teaching method by GPA, class by teaching method, and the 3-way interaction were not significant, so they were dropped from the model.

Figure 2. Effects of teaching and testing methods on student performance as indicated by the percentage of correct answers. White bars are for the group of students who received al instruction and testing in the traditional way. Dark bars are for students who received online instruction and testing for the quizzes and homework (QZ&HW), online instruction and traditional testing for the final test (FNL), and traditional instruction and testing for the midterm (MT) materials. Thin lines represent standard errors.

Winter 2001 Access workshops

Because of logistics, students were free to decide when to take the tests relative to the workshops, which resulted in three self-selected groups. Twelve students took the online “before” test, studied workshops 7 and 8 online, and then took the “after” online test. Eighteen students took the online “before” test, studied workshops 7 and 8 in the traditional intensive face-to-face form, and then took the “after” online test; 122 students studied workshop 7 in the traditional manner first, took the “before” online test, studied workshop 8 in the traditional manner, and then took the “after” online test (Figure 3). All students were evaluated using the same method, but unlike the midterm and final tests, in this case the method was computer testing with SkillCheck, the same method used by students in the online group for quizzes during the whole term.

The first two bars of Figure 3 show that there was no difference in performance between the online and traditional groups before or after studying the materials. Both groups exhibited a significant increase in performance (28±2.4) after the learning session, regardless of the teaching method. The last bar shows that students who studied workshop 7 before the initial test did better in this test than the other groups, but did not differ in the score achieved in the second test. This indicates that test scores went up because of learning and not just because of experience with the testing procedure. In summary, performance of students in the online and traditional groups did not differ when tested on workshops 7 and 8 using a computer-based testing method familiar to the online group, and presumably less familiar to the traditional group.

Figure 3. Effect of teaching method on performance in a test delivered online before (white bars) and after (white plus dark bars) learning, during Winter 2001. The dark segments represent the effect of studying the information. TLT Online: students who took an online test, studied online and then took the online test again; TLT Traditional: students who took an online test, studied in the traditional intensive face-to-face manner, and then took the online test again. LTLT-Traditional: students who studied half of the material in the traditional manner, took an online test, studied the rest of the material in the traditional manner, and then took the second online test. Bars segments with no common letters are significantly different at the 5% level by Tukey’s HSD. Only comparisons among bars of the same color are valid. Thin lines represent standard errors.

Results of the study in Winter 2001 can be interpreted more clearly if one considers both the teaching and testing methods used for each aspect of the course. The following interpretation suggests that the only difference in performance observed (quizzes and homework) was caused by the testing and not the teaching method.

The four main measures of performance used in Winter 2001, midterm, quizzes and homework, Access test, and final exam scores, assessed how students did in distinct aspects of the course (Table 4). The midterm evaluated their performance in the lecture topics. Lecture topics were taught and tested in the same manner in both the online and traditional groups, thus, no direct effect of lab teaching method (online vs. traditional) was expected on midterm scores and none was observed. The final test evaluated almost exclusively the information taught in the labs, and it was administered in exactly the same face-to-face manner to both online and traditional groups. No difference due to teaching method in final performance was observed here either. Access workshops 7 and 8 were taught either online or in the traditional manner, depending on the group, but all students were tested in the same online manner and no performance differences were detected. Finally, homework and quiz topics were both taught and tested either online or in the traditional face-to-face manner, depending on treatment group. In this case, the traditional group performed better.

Table 4. Teaching and testing methods used for different topics in each group of students. The “online” group received instruction on the use of computers through the web, whereas the “traditional” group attended laboratories where they had intensive interaction with teachers.

		Topics
Group	History, Hardware (Midterm)	Use of computer (Quizzes & Homework)	Use of computer (MS Access)	Use of computer (Comprehensive Final)
online	Traditional lectures & written test	Online instruction & computer testing	Online instruction & computer testing	Online instruction; face-to-face testing by reader
traditional		Face-to-face instruction & testing by readers	Face-to-face instruction & computer testing	Face-to-face instruction & testing by readers
Difference in performance	0	Traditional > online	0	0

The integration of all results allows us to derive a qualitative evaluation of the potential interaction between teaching and testing methods. Differences in performance were detected only when both the testing and teaching methods differed between groups. No differences were observed when only the teaching method was different, regardless of the testing method. Therefore, the difference detected in performance is related to the testing method. This interpretation is consistent with the fact that we experienced problems implementing the testing software on the network; students complained about the speed of the client software, and they were generally displeased with this form of testing relative to the face-to-face, personalized testing provided in the regular labs and experienced by students in the traditional group. We surmise that the lower performance in the online testing group was related to the facts that students were frustrated and intimidated by taking tests on a computer, and that computer testing did not have the flexibility to ameliorate unexpected situations (e.g. computer crash) provided by the traditional testing. Students performed better when tested in the classroom with a high level of interaction with, and immediate feedback from readers. Students tested by readers were able to concentrate on answering and showing how much they knew, whereas students tested by computer had to devote a great deal of attention to the testing process.

Overall, it is our assessment that testing was the weakest part of the online method. The instructor and readers spent more time troubleshooting and managing the testing server and software than in any other activity related to the online group. Effects of testing and teaching method should be studied separately when considering online courses. Different combinations of testing and teaching methods may yield the benefits of online instruction without the negative impact of computer testing.

Assessment of methods by students

Because in Fall 2000 students chose which of the two workshops to take online, we had three self-selected groups for workshops 7 and 8, traditional-traditional, traditional-online, online-traditional (nobody did both online). In Winter 2001 individual students did both workshops according to their original grouping, either traditional or online, so their answers to both questions reflected their evaluation of a single method, either online or traditional.

Among the students that took both workshops in the traditional way in the Fall 2000, 68% (n=104) ranked both workshops equally, 22% ranked workshop 7 higher, and 10% ranked workshop 8 higher (Figure 4). The proportion of students ranking both workshops equally declined when one of the workshops was taken online. The percentage of students ranking workshop 7 higher declined to 18% while the percentage ranking workshop 8 higher increased to 29%, when workshop 7 was taken online. Conversely, when workshop 8 was taken online, 33% ranked workshop 7 higher and only 14% ranked workshop 8 higher.

Students clearly preferred the traditional instruction. Although the majority of students who took the traditional workshops did not show a preference, more of these students preferred workshop 7 than 8. This preference was accentuated in the group that took workshop 8 online and workshop 7 in the traditional format, whereas the preference was reversed in the group that took only workshop 7 online (P=0.02). Yet, the majority of students ranked the methods equally, and a significant proportion of students always ranked the online method higher than the traditional method. This may indicate that different teaching methods are more suitable for different individuals, and that the traditional method varied among teachers, thus giving students a variety of experiences.

In Winter 2001, ranks did not differ between workshops (P=0.07, signed-rank test), but they were significantly greater in the traditional than in the online group for both workshops (P<0.04). Ranks for workshops 7 and 8 were highly correlated with each other in both groups (Spearman’s rank correlation=0.72 and 0.73 for the online and traditional groups), P<0.001), indicating consistency in the responses within students.

We believe that the differences detected do reflect the general way students felt about the way information and instructions were given in each teaching method. Both methods were good, as reflected by the fact that both in the online and traditional groups most students agreed or strongly agreed (Table 5) with the statement that the workshop materials were presented in a manner that was accessible and understandable. However, the traditional method was clearly viewed as better by a significant proportion of students.

Figure 4. Effect of teaching method on ranking by students. Grey sections of bars are the percentage and show the number of students that ranked two workshops equally. White sections indicate the percentage and number of students who ranked workshop 7 higher than 8, and dotted sections represent those who ranked workshop 8 higher than 7.

The learning environment can affect performance indirectly by directly affecting attitude (Newby and Fisher, 2000). In our study, the teaching methods created very different environments, yet, performance did not differ much. Informal discussions with students and records of office visits indicate that several students in the online group became confused and frustrated with the online delivery method. In our opinion, the secure access to the server and the highly structured navigation were the most problematic. Second were issues with system crashes during tests. It is reasonable to assume that these problems negatively affected attitudes, as detected in the study, and that the negative attitude diminished learning performance. Following this line of reasoning, it should be possible to increase the performance in the online method by fixing the source of negative impacts on attitudes.

Our results differ from those in similar studies. Day et al., (1998) compared a traditional with an “online” version of a writing course and concluded that students in the web-based (online) course achieved at higher level than in a traditional classroom approach. In our study, no differences in achievement were found when students were tested in the same manner. Day et al., (1990) also found that students in the online course improved their attitudes toward writing, and did not change their attitudes towards computers or the web. In our study, students preferred the traditional teaching method.

Table 5. Percentage of students in each teaching-method group that agreed or strongly agreed with a statement that workshop materials had been presented in an accessible and understandable manner.

	Workshop 7	Workshop 8
Online	83	80
Traditional	92	91

Conclusion

We studied objective performance and student assessment of two methods to teach an introductory course on the use of computers in agriculture. The testing involved about 550 students in two different quarters. No significant differences in performance measures were detected between the online and the traditional face-to-face teaching methods, but we found an indication that students did better when tested in the traditional written or face-to-face manner than when tested online. Although the majority of students ranked both methods equally, there was a clear indication that they preferred the traditional to the online learning method. This study demonstrates that objective performance and student attitude towards a learning experience can be decoupled, in spite of the potential effects of attitude on performance. We should focus efforts on improving the presentation and attitude elicited by the online teaching method by incorporating the aspects of face-to-face learning that cause the more positive attitude. Online testing methods must be modified so students can focus on the content, instead on the medium and format of the test.

Acknowledgments

Mellon Foundation Grant

UCD Mediaworks (Steve, Bob, Freja, John)

TRC (Tor)

Computer labs (readers, Johnatan Bar-Or, Tim Leamy, Brian Monroe)

Literature Cited