Java程序辅导

© 2009 The Authors
 
Teaching Statistics. Volume 31, Number 1, Spring 2009
 
•
 
21
 
Journal compilation © 2009 Teaching Statistics Trust
 
Blackwell Publishing LtdOxford, UKTESTeaching Statistics0141-982X© 2008 The Au orsJournal comp ati n © 2008 Teaching Statistics TrustXXXriginal Article
 
XX XXX
Internet Approach versus Lecture and Lab-Based Approach 
for Teaching an Introductory Statistical Methods Course: 
 
Students’ Opinions
 
KEYWORDS:
 
Teaching;
Online;
Students’ perspectives;
Pedagogical research.
 
H. Dean Johnson, Nairanjana Dasgupta, 
Hao Zhang and Marc A. Evans
 
Washington State University, USA.
e-mail: dean_johnson@wsu.edu
 
Summary
 
The use of the Internet as a teaching tool continues
to grow in popularity at colleges and universities.
We consider, from the students’ perspective, the
use of an Internet approach compared to a lecture
and lab-based approach for teaching an introduc-
tory course in statistical methods. We conducted a
survey of introductory statistics students. Contra-
dictory to what was hypothesized by the authors,
they favoured keeping the lecture and lab-based
approach for teaching the class.
 

 
INTRODUCTION
 

 
I
 
nternet classes are being seen more and more as
alternatives to lecture-based classes. There have
been several studies to examine the effects of the
two approaches on test scores in introductory
statistics classes (Utts et al. 2003; Ward 2004;
Dutton and Dutton 2005). There has been less
research, however, on the opinions of students
regarding the use of an Internet approach com-
pared to a traditional approach for teaching statistics.
If one is considering using an Internet approach
to teaching, it is important for the instructor to
understand the attitudes that students have towards
Internet classes.
This study was conducted to shed some light on
the perspectives of introductory statistics students
regarding the type of teaching approach they
would like to see used for teaching statistics. Being
teachers of statistics, we wanted to learn whether
students would advocate an Internet approach to
teaching an introductory course in statistical
methods over the traditional lecture and lab format
that we currently use. The popularity of the
Internet with students coupled with the perceived
dislike that many students have for lecture classes,
one perhaps would suspect that students would
welcome an Internet class. To investigate whether
students would favour an Internet approach over
a lecture-based approach to teaching introductory
statistics, we surveyed students in STAT212, an
introductory algebra-based statistical methods
course taught at Washington State University.
 

 
THE DATA COLLECTION
 

 
Three sections of STAT212 are offered each semester
at Washington State University. This course satisfies
a general education requirement for mathematics
proficiency at the university. In addition, it satisfies
the requirements of many departments.
In the class, students use statistical software and
they use the Internet for the purpose of printing
out class handouts as well as running JAVA
applets. Thus, computer and Internet access are
important. Students enrolled at Washington State
University have access to computer labs on cam-
pus, where they can use university software. They
also have access to the Internet as the dormitories,
libraries and computer labs are connected to the
Internet through Ethernet.
 22
 
•
 
Teaching Statistics. Volume 31, Number 1, Spring 2009
 
© 2009 The Authors
Journal compilation © 2009 Teaching Statistics Trust
 
STAT212 consists of 3 hours of lecture and 2
hours of lab per week. In the labs, students partic-
ipate in hands-on activities, in which they collect
their own data and analyse the data using Minitab
statistical software (Minitab Inc. 2004). Professors
conduct the lectures and graduate students con-
duct the labs. In the spring of 2006, three different
professors taught the three sections of STAT212.
Two of the sections each enrolled approximately
125 students and the third section enrolled approx-
imately 75 students.
The three sections used the same text and had a
similar structure. Each professor required three
exams, homework, completion of labs and a project.
For the class project, students were asked to
complete the survey found in the Appendix and
analyse the data resulting from the survey. The
instructors purposely made the survey short to
make the data analysis more manageable for the
introductory statistics students. At the end of the
semester, the students were each required to
turn in a two- to three-page report in which they
discussed the results of their analysis.
The authors have observed that having students
complete a project, in which they collect and
analyse data that are of interest to them, enhances
the learning of students in introductory statistics
classes. Consequently, the authors will continue to
incorporate projects. One difficulty encountered,
however, involves the amount of time required to
grade the projects. If the reader is considering
incorporating such a project in his or her class,
the authors would suggest requiring group
projects as opposed to individual projects to cut
down on the grading time, especially if the class
size is large.
In our study, we combined the survey responses
from the three sections of STAT212, resulting in a
sample size of 267. A numerical description of the
sample is provided in table 1.
 

 
ANALYSIS
 

 
In our analysis, we examined the percentage of all
study participants in favour of an Internet approach
and then decomposed the preferred teaching
approach by each of the following variables:
class (freshman, sophomore, etc.), gender and
student type. The majority, 75% (95% confidence
interval: 69%, 80%), of the students were in favour
of the traditional approach. Also, regardless of
class, gender and student type (honours college
or not), the traditional approach was the preferred
choice of students, as shown in figures 1–3.
To identify factors that relate to preferred teaching
method, we conducted a logistic regression analysis
using the variables in table 2. Logistic regression was
used here instead of ‘ordinary’ regression because
the response variable (preferred teaching method)
is a binary response variable (0 = No, 1 = Yes) as
opposed to a quantitative variable as one would
encounter in an ‘ordinary’ regression analysis. If a
survey similar to the one described in this study
was conducted with question 5 being changed so as
to ask students to identify on a continuous scale, say
from 0 (strongly disagree) to 10 (strongly agree),
how strongly they agree or disagree that they
would prefer an Internet class, the results could be
analysed using ‘ordinary’ regression techniques.
In a logistic regression analysis, the probability of
the occurrence of one of the two outcomes of the
Table 1. Description of students in STAT212 (Spring 2006)
Quantitative variable Minimum 25th percentile Median 75th percentile Maximum
Average number of hours on Internet per week 1 5 10 18 80
Average number of hours on Internet per week for class 0 0.5 1 1 15
Math skill 1 6 7 8 10
Qualitative variable Count Percentage
Class
Freshman 40 15
Sophomore 96 36
Junior 81 30
Senior 50 19
Gender
Female 187 70
Male 80 30
Student type
Honours college student 28 10
Regular university student 239 90
 © 2009 The Authors
 
Teaching Statistics. Volume 31, Number 1, Spring 2009
 
•
 
23
 
Journal compilation © 2009 Teaching Statistics Trust
 
binary response variable is modelled, via the logit
transformation, as a function of one or more
explanatory variables. One of the purposes of logistic
regression, like ordinary regression, is to identify
variables that are good predictors of the response
variable. For more information on logistic regres-
sion, please refer to Hosmer and Lemeshow (2000).
One approach to identifying good predictors is to
use a variable selection procedure. We used a step-
wise procedure, with the stopping rule being that
the 
 
p
 
-value of all variables not in the model is
> 0.10 (for both entering and leaving the model).
We chose this 0.10 cut-off as opposed to the tradi-
tional 0.05 cut-off mark because of the exploratory
nature of our study. In using stepwise logistic
regression procedure, GENDER and AVGINT-
CLASS were identified as being statistically signi-
ficant predictors of PREFERRED TEACHING
APPROACH. The results of this stepwise logistic
regression analysis are presented in table 3. In
our discussion of the results, we interpret them in
terms of odds ratios (Hosmer and Lemeshow 2000),
holding the other variable constant.
In regard to GENDER, we estimated that females
are 0.574 times as likely to prefer an Internet class
as males, holding AVGINTCLASS constant. This
finding of males preferring computer usage more
Fig. 1. Preferred teaching approach by gender
Fig. 2. Preferred teaching approach by class
Table 2.  Description of explanatory and response variables used in logistic regression analysis
Variable name Description
Response variable
PREFERRED TEACHING APPROACH Lecture = 0, Internet = 1
Explanatory Variables:
CLASS Freshman = 1, Sophomore = 2, Junior = 3, Senior = 4
GENDER Male = 0, Female = 1
STUDENT TYPE Whether or not student belongs to honours college: No = 0, Yes = 1
AVGINT Average number of hours on Internet per week
AVGINTCLASS Average number of hours on Internet per week for class
MATH SKILL Student’s rating of his or her math skill: 1 (low) to 10 (high)
Fig. 3. Preferred teaching approach by student type
 24
 
•
 
Teaching Statistics. Volume 31, Number 1, Spring 2009
 
© 2009 The Authors
Journal compilation © 2009 Teaching Statistics Trust
 
than females is consistent with findings from other
studies (Temple and Lips 1989; Whitley 1997).
In regard to AVGINTCLASS, for every 1-hour
increase in the average number of hours a student
spends on the Internet for the class, we estimated
that he or she is 0.673 times as likely to prefer an
Internet class, holding GENDER constant. In
STAT212, students are required to retrieve class
notes and assignments from the Internet. This was
not meant to take much time for the students. The
professors did observe, however, that for some
students retrieving the information was time-
consuming and they became frustrated as a result.
The frustration experienced by some students can
be used as one possible explanation for an increase
in Internet usage for the class being associated with
being less likely to prefer an Internet class.
The overall concordance for our final model was
54.9%. The Hosmer–Lemeshow Goodness of Fit
indicates the logistic regression model provides a
good fit for the data (
 
p
 
 = 0.456). Two outliers were
detected in the logistic regression analysis. For
each of these two observations the amount of time
spent on the Internet for class purposes was
unusually large. Specifically, one student indicated
that an average of 15 hours a week was spent on
the Internet for class purposes and the other
student indicated an average of 10 hours. As
mentioned earlier, students in STAT212 are only
required to use the Internet for retrieving class notes
and assignments. Given this, these two responses
were deemed questionable and the decision was
made to leave them out of the final analysis. How-
ever, it should be noted that when the two outliers
were included in the analysis, only GENDER was
detected as a significant predictor of preference.
 

 
DISCUSSION
 

 
As mentioned earlier, the students in the three
sections of STAT212 were required to write a
report for their class project. In the write-ups,
many students expressed their opinions regarding
why they favoured lecture-based classes over Inter-
net classes. Below are some of the comments made
by students, which help to shed some light on the
preference students have for a lecture-based course
over an Internet-based approach. These comments
are grouped according to the different themes that
emerged from the opinions.
 
Face-to-face interaction
 
Many of the students in this study value the
face-to-face learning experience, and thus prefer
a lecture-based approach over an Internet-based
approach, as illustrated by the comments below.
 
I personally feel that the interaction between students and
their teacher/professor is crucial for successful learning. So, the
option of an Internet class should be overridden.
The social interaction between fellow students, the professor, and
the TAs [teaching assistants] is a vital part of the learning process
and one would not succeed nearly as easily if they were taking an
online-based course.
It is my belief that students are better able to learn material
when they have the opportunity for face-to-face contact with the
instructor and fellow students. In a face-to-face setting students
can more easily ask questions and learn from the questions and
input of other students.
 
Subject matter
 
It is the belief of some of the students that the
most effective teaching approach depends on the
subject matter, as discussed below.
 
For a history class that has strictly lecture notes, exams and
quizzes could be taken online because you don’t have any
examples and the information is either read to the class or
presented in a way that notes are needed. A math class that has
different equations and examples should not be taught online
because if there is a misunderstanding, usually the teacher can
help explain it another way if you are in class.
Science-based classes should not be used for Internet purposes,
but classes such as history would be beneficial to students
because there are no interactive ‘hands-on’ activities.
Some classes simply must be taught in the classroom such as those
with ‘hands-on’ labs or very detailed subject matter such as calculus.
 
Procrastination
 
According to some of the students, Internet classes
have the potential of fostering procrastination, as
detailed in the comments below.
Table 3. Results of logistic regression analysis
Parameter Odds ratio p-value
90% lower confidence 
limit for odds ratio
90% upper confidence 
limit for odds ratio
INTERCEPT – 0.2836 – –
GENDER 0.574 0.0625 0.351 0.937
AVGINTCLASS 0.673 0.0364 0.493 0.919
 © 2009 The Authors
 
Teaching Statistics. Volume 31, Number 1, Spring 2009
 
•
 
25
 
Journal compilation © 2009 Teaching Statistics Trust
When a student completes a course online, it is done at his or
her own pace and within the comfort of their home, which
allows the student to fall behind and fail to learn the material
thoroughly.
A disadvantage of online courses is the complete lack of super-
vision. If a person is not motivated to finish what they start,
then it is probably not a good idea to take an online course.
While there may be deadlines and expectations from the online
course instructor, there is no one there to remind him/her to do
the assignments or to take the tests.
It is difficult for some people to muster the self-discipline to do
the homework every day or every week, without a teacher, it
becomes easier to procrastinate.
 
Convenience factor
 
Some students expressed inconveniences associated
with the use of an Internet-approach and these are
mentioned below.
 
Most students I have talked to prefer to have a professor to be
able to go talk with when they are having problems, not try to
surf the Internet for answers.
It is difficult to understand the material on your own and not be
able to ask questions directly to your professor when instead
you have to correspond through e-mail and wait for a response.
There are disadvantages to Internet classes. Say for instance,
the Internet network on campus was not functioning for a week,
then the Internet class could not be held due to technicalities.
 

 
CONCLUSION
 

 
The Internet enjoys great popularity with college
and university students. This was illustrated in this
study, with the median number of hours spent on
the Internet by the introductory statistics students
in this study being 10 hours. The authors anticipated
that this popularity would translate into the
students preferring an Internet approach to teach-
ing introductory statistics over the lecture/lab
approach that is currently used at Washington State
University. This was not the case. The majority of
the students preferred the lecture/lab approach
over an Internet approach and this was true
regardless of gender, class (freshman, sophomore,
etc.), and student type (honours college or not).
We note that our study was on the teaching
approach (Internet or lecture) preferred by students
and not on the effectiveness of the two approaches.
Just because students advocate a lecture-based
approach to an Internet-based approach, this does
not mean that students will necessarily learn more
when such an approach is used. Several studies
have been conducted examining the effect of
Internet instruction on performance in statistics
classes. In these studies, different conclusions were
reached regarding the effect of Internet instruction
on course performance. Dutton and Dutton (2005)
concluded that students in a Business Statistics
class performed better when an online teaching
approach is used compared to when a lecture-based
approach is used. But Utts et al. (2003) and Ward
(2004) concluded that there is no significant
improvement in performance when a hybrid-
Internet approach is used compared to when a
traditional lecture-based approach is used.
Our study was conducted to see if students would
advocate switching from the traditional lecture-based
approach to teaching an introductory course in
statistical methods currently used at Washington
State University, to an Internet-based approach.
We conclude that, at Washington State University,
the majority of students favour the lecture-based
approach to instruction, and this is true regardless
of gender, class and student type. Based on the
results of other studies, showing lack of compelling
evidence that using an Internet-based approach
improves performance, and our results, the authors
will continue to use a lecture-based approach in
teaching the introductory statistical methods classes
at Washington State University.
 
References
 
Dutton, J. and Dutton, M. (2005). Character-
istics and performance of students in an
online section of business statistics. 
 
Journal
of Statistics Education
 
, 
 
13
 
(3), http://www.
amstat.org/publications/jse/v13n3/Dutton.html.
Hosmer, D.W. and Lemeshow, S. (2000).
 
Applied Logistic Regression
 
 (2nd edn).
New York: John Wiley & Sons.
Minitab Inc. (2004). 
 
Minitab, Version 14
 
(Computer software). State College, PA:
Minitab Inc.
Temple, L. and Lips, H.M. (1989). Gender
differences and similarities in attitudes toward
computer. 
 
Computers in Human Behavior
 
,
 
5
 
(4), 215–226.
Utts, J., Sommer, B., Acredolo, C., Maher,
M.W. and Matthews, H.R. (2003). A study
comparing traditional and hybrid internet-
based instruction in introductory statistics
classes. 
 
Journal of Statistics Education
 
, 
 
11
 
(3),
http://www.amstat.org/publications/jse/v11n3/
utts.html.
 26
 
•
 
Teaching Statistics. Volume 31, Number 1, Spring 2009
 
© 2009 The Authors
Journal compilation © 2009 Teaching Statistics Trust
 
Ward, B. (2004). The best of both worlds:
a hybrid statistics course. 
 
Journal of Statistics
Education
 
, 
 
12
 
(3), http://www.amstat.org/
publications/jse/v12n3/ward.html.
Whitley, B.E. Jr. (1997). Gender-difference in
computer-related attitudes and behavior:
a meta-analysis. 
 
Computers in Human
Behavior
 
, 
 
13
 
(1), 1–22.
 
Appendix
Survey used for STAT212 Project
 
1. What is your gender?
0 = Male 1 = Female? _________
2. What year are you in Washington State
University? ____________
1 = Freshman 2 = Sophomore 3 = Junior
4 = Senior 5 = Graduate student
3. How often do you access the Internet? (number
of hours on an average week) _______________
4. How often do you access the Internet for
STAT212 class purposes? (number of hours on
an average week): __________
5. Would you prefer an Internet class as opposed
to a face-to-face lecture with lab-type class for
STAT212? 
1 = Yes 0 = No ________
6. On a scale of 1 (low) to 10 (high), how would
you rate your Math skills? ________
7. Are you a part of the honours program?
1 = Yes 0 = No ________
 
XXX
 
 
 
Original Articles
 
SHORT TITLE RUNNING HEAD:AUTHORS RUNNI G HEAD:
LOOK AHEAD
 
Forthcoming articles include
Analyzing the world population: Using population pyramids and 
 
If the World Were a Village
 
The binomial distribution in shooting
Correlation between coin random variables
Experimental probability in elementary school
Explaining a positive test result to a patient
Hypothesis testing with the Three Stooges
An inside look at the two envelopes paradox
Madlibs
Playing with residuals
Statistics online computational resource for education
Teaching decision theory with ‘Deal or No Deal’
Understanding of patterns, streaks, and independence by grade school children
A website that provides resources for assessing students’ statistical literacy, reasoning, and thinking