08832323.2014.973828

Journal of Education for Business

ISSN: 0883-2323 (Print) 1940-3356 (Online) Journal homepage: http://www.tandfonline.com/loi/vjeb20

Assessing the Disconnect Between Grade
Expectation and Achievement in a Business
Statistics Course
Mark L. Berenson, Renu Ramnarayanan & Alan Oppenheim
To cite this article: Mark L. Berenson, Renu Ramnarayanan & Alan Oppenheim (2015)
Assessing the Disconnect Between Grade Expectation and Achievement in a Business Statistics
Course, Journal of Education for Business, 90:2, 72-80, DOI: 10.1080/08832323.2014.973828
To link to this article: http://dx.doi.org/10.1080/08832323.2014.973828

Published online: 21 Nov 2014.

Submit your article to this journal

Article views: 46

View related articles


View Crossmark data

Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=vjeb20
Download by: [Universitas Maritim Raja Ali Haji]

Date: 11 January 2016, At: 19:08

JOURNAL OF EDUCATION FOR BUSINESS, 90: 72–80, 2015
Copyright Ó Taylor & Francis Group, LLC
ISSN: 0883-2323 print / 1940-3356 online
DOI: 10.1080/08832323.2014.973828

Assessing the Disconnect Between Grade
Expectation and Achievement in a Business
Statistics Course
Mark L. Berenson
Downloaded by [Universitas Maritim Raja Ali Haji] at 19:08 11 January 2016

Montclair State University, Montclair, New Jersey, USA


Renu Ramnarayanan
Seton Hall University, South Orange, New Jersey, USA

Alan Oppenheim
Montclair State University, Montclair, New Jersey, USA

In an institutional review board–approved study aimed at evaluating differences in learning
between a large-sized introductory business statistics course section using courseware
assisted examinations compared with small-sized sections using traditional paper-and-pencil
examinations, there appeared to be a severe disconnect between the final grades students
expected to receive and the grades they actually earned. Given that the core-required
business statistics course has, for many years, had a reputation among students as being
anxiety producing, this disconnect was quite surprising. Using student responses to a survey
administered early in the semester, demographic information was available that enabled
assessment of this disconnect.
Keywords: academic performance, grades, statistics, student expectations, student
satisfaction

More than a century has elapsed since the science fiction

writer, journalist, and political commentator H. G. Wells
(1903) was attributed to have remarked that understanding
statistics would one day be as necessary for efficient citizenship as the ability to read and write (Wilks, 1951).
Unfortunately, however, the study of statistics in business
schools over the past six decades has failed to generate
widespread enthusiasm for the discipline and this has precluded the prediction’s realization. Far too frequently,
undergraduates interested in other business disciplines
approach their core-required statistics course with anxiety
and trepidation (Zeidner, 1991), never sensing its value,
never appreciating how it can make them better consumers
and users of data and information. And it is not uncommon

Correspondence should be addressed to Renu Ramnarayanan, Seton
Hall University, 400 South Orange Avenue, South Orange, NJ 07079,
USA. E-mail: renu.ramnarayanan@shu.edu

to hear undergraduates refer to the course as business
“sadistics.”
Considering this, should we anticipate a significant disparity between students’ expected versus actually achieved
course grades in a core-required introductory statistics

course? And if so, should we anticipate that such disparity
would be the result of students’ overestimation of the
grades they would earn in this course rather than
underestimation?
Impetus for Present Study
In an institutional research board (IRB)–approved study
aimed at assessing differences in learning when offering a
large-sized introductory business statistics course section
using courseware-assisted examinations compared with
small-sized sections using traditional paper-and-pencil
examinations, there appeared to be a severe disconnect
between the final grades students expected to receive and

DISCONNECT BETWEEN GRADE EXPECTATION AND ACHIEVEMENT

73

Student Questionnaire:
Name:______________________________________ Student ID: _ _ _-_ _-_ _ _ _ Class Section:___________
Gender: M___


F___

Major:______________________________________
Transfer Student: Yes, from community college___
Living on Campus:

Yes___

Yes, from other university___

No___

No___

Downloaded by [Universitas Maritim Raja Ali Haji] at 19:08 11 January 2016

Country of Birth:___________________________
Current Employment Status: Full-time___
‘Comfort Level’ with Math:


1
Low

2

Part-time___
3

4
Neutral

Unemployed___
5

6

7
High


Intention to Go to Graduate School to Pursue a Master’s or Doctorate Following Graduation:
Yes___

No___

Not Sure___

Anticipated Course Grade:
A___

A-___

B+___

B___

B-___

C+___


C___

C-___

D+___

D___

D-___

F___

FIGURE 1. Questions extracted from survey used for developing a student profile.

the grades they actually earned. This disconnect was quite
surprising, given that the core-required business statistics
course has, for many years, had a reputation among students as being anxiety producing.
Questions to be addressed here are whether the observed
disconnect between expected versus actual course grades is
significant and, if so, what might be the determinants of

this disconnect? Student responses to a survey administered
early in the semester as part of the aforementioned study
provided demographic information that enabled such investigation. Figure 1 contains the extracted questions from the
survey that pertain to this present study.

Understanding the Students’ Survey Responses
It was observed from responses to the final survey question
in Figure 1 that all students reported that they expected
their performance to be at a high level (i.e., that they would
earn at least a grade of B in their core-required business statistics course).
Unfortunately, it cannot be determined whether the
students’ responses to such a question on course grade
expectation are truly honest. Perhaps they exemplify what
Bandudra (1997) defined as self-efficacy: a belief in an

individual’s own capabilities to produce the desired
outcomes.
Several other reasons, however, are also possible. Perhaps some students, part of the present “me generation,” lie
to themselves by providing exaggerated grade expectations
while other students, aware that their instructor will be

reading their nonanonymous responses, perhaps lie to
please the teacher (i.e., the halo effect) or perhaps intend to
work more diligently (i.e., the Hawthorne effect), motivated
by responding to the questionnaire (i.e., expectancy theory).
And perhaps still other students may consciously exaggerate their expectations because they perceive that in the business world one must show confidence to be successful.
Such supposition on why students may respond to the
grade question as they do must be the subject of future
research by colleague practitioners in various fields of psychology. What is more germane here, however, is assessing
the magnitude of the observed disconnect and, if significant, exploring the potential ramifications that this may
have.
Potential Impact of a Grade Disconnect
If initial grade expectations (Vroom, 1964) are not realized
through actual course performance, cognitive dissonance

Downloaded by [Universitas Maritim Raja Ali Haji] at 19:08 11 January 2016

74

M. L. BERENSON ET AL.


(Festinger, 1957; Festinger & Carlsmith, 1959) would
likely set in, and students may try to search for reasons to
rationalize this disconnect in order to reduce their dissonance. Toward that end, with respect to their business statistics course in particular, they may attribute their poor
performance to inadequate instruction, or they may characterize the course as useless or a necessary evil. By word of
mouth, course negativity could be passed down to the next
class of enrollees, creating an overall anxiety and fear of
taking this core-required course and perpetuating its poor
image and bad reputation.
A question that naturally arises is whether there are
greater levels of disconnect and potentially more dissonance to overcome among some unpopular core-required
courses such as business statistics than others. An assessment of whether the level of disconnect between expected
and actual grade differs significantly among different types
of core-required courses must become the subject of needed
future research because if the disconnect correlates with
course enjoyment and satisfaction it can impact on student
evaluations of faculty and the latter are widely used in reappointment, tenure, promotion, salary adjustment, and sabbatical decisions.

LITERATURE REVIEW
A review of the literature on grade expectation surprisingly
indicates that the results observed in this study are not surprising but rather a confirmation of findings obtained in several other research endeavors across various disciplines—
there is an apparent disconnect between students’ expected
versus actual grades. A study by Landrum (1999) speculated
that grading practices and grade inflation in secondary
schools might be cause for an expectation among college
students that attendance and effort should be sufficient to
earn a good grade and several studies have shown that students overpredict their grades at the beginning of the semester or quarter, and that they also overestimate their abilities
and the amount of time and effort they can devote to a class
(Andrews, Swanson, & Kugler, 2007; Cann, 2005; Gaultney
& Cann, 2001; Khachikian & Guillaume, 2002, 2011; Khachikian, Guillaume, & Pham, 2011; Perlman & McCann,
2007; Prohaska, 1994; Wendorf, 2002). Moreover, McCann,
Immel, Kadah-Ammeter, and Priniski (2013) recently found
that university students expected to receive high grades in
lower level (i.e., core-required) courses. And a study conducted nearly 40 years ago indicated that the majority of
students overestimated their initial course grade by an average of one letter grade and that men predicted higher grades
for themselves than did women (Cole & King, 1977). This
latter finding perhaps goes somewhat counter to the argument that today’s generation of students displays more selfinterest and narcissism, has higher expectations, and feels
more entitled (Twenge, 2009).

DEVELOPING THE PRESENT STUDY
The aforementioned studies on grade disconnect were primarily exploratory and descriptive; the present study takes
a confirmatory approach by employing appropriate inferential methodology to investigate both the magnitude of this
disconnect and its potential determinants.
Objectives of Present Study
The objectives of the present study then were threefold:
 To determine whether there is evidence of an overall
real difference between expected and actual grades of
students taking the core-required introductory business statistics course,
 To determine whether there is evidence of a real difference between expected and actual grades between
the two major groupings from the initial IRBapproved study—students in large class sections with
courseware assisted examinations versus students in
small class sections with traditional paper-and-pencil
examinations, and
 To explore possible differences in grade expectancy
versus reality based on student responses to a survey
(see Figure 1) containing a variety of demographic
variables
METHOD
Study Sample
The initial group of subjects in the original IRB-approved
study consisted of students in four specific sections of the
required undergraduate business statistics course. Only students who provided written consent to participate in that
study, finished the course with an earned grade of A–F, and
completed an in-class survey early in the semester were
part of the set of analyses for this study. This resulted in
128 participants.
Participants
The most typical student characteristics indicated a
U.S.-born (68.5%) woman (50.8%), majoring in a nonquantitative subject (52.0%), who started college at this university (53.1%) but lived off campus (84.1%), and was
working part-time (45.3%). The typical student was comfortable with mathematics (66.4%), was not sure or not
planning on attending graduate school (57.8%), and
expected an A-level grade (71.1%) in the course. Table 1 is
a supertable (Tufte, 1983) displaying this profile between
the two major groupings from the initial study—students in
large class sections with courseware assisted examinations
versus students in small class sections with traditional

DISCONNECT BETWEEN GRADE EXPECTATION AND ACHIEVEMENT

75

TABLE 1
Characteristics of 128 Students in Two Types of Course Sections Along With the Distribution of Actual Course Grades

Downloaded by [Universitas Maritim Raja Ali Haji] at 19:08 11 January 2016

Responses

All students
Gender
Female
Male
Major
Quantitative
Nonquantitative
Transfer status
No
Yes
Live on campus
No
Yes
Country of birth
USA
Other
Employment status
Full-time
Part-time
Unemployed
Graduate school intent
Yes
No or not sure
Comfort level with mathematics
Higher (5–7)
Lower (1–4)
Anticipated grade
A
A–
BC
B
Actual grade
A
A–
BC
B
BCC
C
C–
DC
D
F

Section

n

%

Large class, courseware

Small class, traditional

128

100.0%

47.7%

52.3%

65
63

50.8%
49.2%

56.9%
38.1%

43.1%
61.9%

61
66

48.0%
52.0%

52.5%
43.9%

47.5%
56.1%

68
60

53.1%
46.9%

54.4%
40.0%

45.6%
60.0%

106
20

84.1%
15.9%

43.4%
65.0%

56.6%
35.0%

87
40

68.5%
31.5%

47.1%
50.0%

52.9%
50.0%

46
58
24

35.9%
45.3%
18.8%

23.9%
63.8%
54.2%

76.1%
36.2%
45.8%

54
74

42.2%
57.8%

46.3%
48.6%

53.7%
51.4%

85
43

66.4%
33.6%

47.1%
48.8%

52.9%
51.2%

63
28
24
13

49.2%
21.9%
18.8%
10.2%

46.0%
53.6%
45.8%
46.2%

54.0%
46.4%
54.2%
53.8%

20
21
11
16
16
7
13
12
1
4
7

15.6%
16.4%
8.6%
12.5%
12.5%
5.5%
10.2%
9.4%
0.8%
3.1%
5.5%

55.0%
42.9%
45.5%
56.2%
50.0%
57.1%
53.8%
33.3%
0.0%
0.0%
57.1%

45.0%
57.1%
54.5%
43.8%
50.0%
42.9%
46.2%
66.7%
100.0%
100.0%
42.9%

Note: One student was undecided with respect to major, two students did not respond about living on campus, and one student did not respond about country of origin.

paper-and-pencil examinations. Also included in Table 1 is
the distribution of actual course grades.
This overall profile, reflecting the modal response for
each of the eight characteristics, may not be the same as the
most typical joint response to all eight questions. Given
that the sample contained only 128 study participants and
the joint responses to all eight questions form a multidimensional table of 1,920 cells (i.e., seven questions categorized
into two possible responses, one question on employment
status with three possible responses, and the anticipated

grade question with five possible grade level responses
ignoring pluses and minuses), the sparseness of such data
preclude any reasonable descriptive drill down.

RESULTS: COMPARING ANTICIPATED VERSUS
ACTUAL COURSE GRADES
Table 2 is a cross-tabulation of grades that students
expected, or were striving for, as reported in their survey

76

M. L. BERENSON ET AL.
TABLE 2
Cross-Classification of Anticipated and Actual Student Grades

TABLE 4
Cross-Classification of A Versus Not A Grades, Anticipated
and Actual

Actual grade
Actual grade
Anticipated
grade

A

A
12*
A–
4
BC
3
B
1
Grand total 20

Grand
A¡ BC B B¡ CC C C¡ DC D F total
10
7*
2
2
21

6 10 7
1 4 3
4* 1 5
0 1* 1
11 16 16

5
1
0
1
7

4 6
3 2
4 2
2 2
13 12

1
0
0
0
1

0
1
2
1
4

2
2
1
2
7

63
28
24
13
128

Anticipated grade
A
Not A
Grand total
%

A

Not A

Grand total

%

12
8
20
15.6

51
57
108
84.4

63
65
128

49.2
50.8
100.0

Downloaded by [Universitas Maritim Raja Ali Haji] at 19:08 11 January 2016

Note: * indicates students that correctly predicted their grades.

questionnaire matched against the actual grades received. It
is interesting, if not ironic, to note that 71.1% of the students anticipated a grade level of A or A- but no students
expected a grade lower than B in this core-required business school course known to be anxiety producing. In actuality, 32.0% A-level grades were given, 33.6% B-level
grades were given, 25.1% C-level grades were given, 3.9%
D-level grades were given, and 5.5% F grades were given.
Only 24 students of 128 (18.8%) correctly predicted their
grades. Overwhelmingly, students thought they would get
better grades than they actually did—perhaps resulting in
them leaving the course with much disappointment. Only
12 students out of 128 (i.e., 9.4%) received a higher grade
than anticipated while 92 students out of 128 (i.e., 71.9%)
received a lower grade than anticipated. That is, for each
student who received a higher grade than expected,
7.67 times as many students received a lower grade than
expected.
For purposes of symmetry in the category groupings
needed for further comparative analysis, the columns of
Table 2 are collapsed to match the anticipated grade rows
and the 4 £ 4 square matrix based on the grades A, A¡,
BC, and B or below are presented in Table 3.
Using these four groupings, the Bowker test for symmetry of response (Bowker, 1948) showed a severe disconnect
between perception and reality (B ! x2a;c.c ¡ 1/=2; a D
0:05 and c D 4; B D 61:90 > x20:05;6 D 12:592; p D .000). A
description of the Bowker test is provided in the Appendix
(Koppel & Berenson, 2007).
TABLE 3
Cross-Classification Table Used for Bowker Test of Symmetry
Actual grade
Anticipated grade
A
A–
BC
B
Grand total

A

A–

BC

B

Grand total

12*
4
3
1
20

10
7*
2
2
21

6
1
4*
0
11

35
16
15
10*
76

63
28
24
13
128

Note: *indicates students that correctly predicted their grades.

TABLE 5
Cross-Classification of A or A- Versus BC or Below Grades,
Anticipated and Actual
Actual grade
Anticipated grade
A or A–
Not A or A–
Grand total
%

A or A–

Not A or A–

Grand total

%

33
8
41
32.0

58
29
87
68.0

91
37
128

71.1
28.9
100.0

Using a McNemar-type a posteriori comparison procedure
(Stuart, 1955), there were obvious significant disconnects
with respect to expected and actual grade at each grade
level displayed in the cross-classifications shown in
Tables 4–6. At each grade level, students significantly
overestimated the grades they actually earned. These results
are highlighted in Table 7. Stuart’s a posteriori comparison
procedure is described in the Appendix.
Based on Tables 2 and 3, Table 8 displays a summary of
cross-classifications of anticipated versus actual grades for
the overall study, for the two key groupings of students
(large class, courseware-assisted exams; small class, traditional paper-and-pencil exams), and for the breakout groupings of eight demographic characteristics shown in Table 1.
The Bowker and McNemar test statistics (see the Appendix
for a description of the McNemar test) are provided for
each of the 20 cross-classifications.

TABLE 6
Cross-Classification of BC or Better Versus B or Below Grades,
Anticipated and Actual
Actual grade
Anticipated grade
A, A–, BC
x21 ¡ a;c.c ¡ 1/=2 ; the upper tail
critical value from the chi-square distribution with
c.c ¡ 1Þ=2 degrees of freedom at a chosen level of significance a.
Stuart A Posteriori Comparisons and Decision Rule
For a family of c a posteriori comparisons, using an
experiment-wise error rate a, declare that a significant
difference exists in the column classification i and
the corresponding row classification i if the absolute
difference in the two sample proportionsqffiexceeds
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiits
ffi
critical range, that is, jp^:i ¡ p^i: j > x21 ¡ a;ðc ¡ 1Þ 
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
^p:i ^pi’:
^p:i ^pi’:
2.p
^ii ¡ ^p:i ^pi: /
where p^:i D nx:i0 and p^i: D nxi:0 ; where
n0 C n0 ¡
n0
i0 is the complement of i, the combined responses from all
c – 1 other classifications, and n0 is the actual sample size
in the study (not the n individuals whose responses were
evaluated in the Bowker test).

McNemar Test and Decision Rule
Given a 2-by-2 cross-classification table containing the tallies from a sample of n0 individuals. Let xij represent a tally
of initial responses i and follow-up responses j (such as
anticipated grade i and actual grade j). The null hypothesis
of symmetry (or equality of paired population
proportions)
P
to be tested is conditioned on those n D
8 i6¼j xij individuals whose responses change and the probability of a
switch from response i to response j is equal to the
probability of a switch from response j to response i,
and this probability is 0.5. Therefore, interest is in the
comparison of the tally x21 (and marginal total x1 .) below
the main diagonal of the 2-by-2 contingency table with the
corresponding tally x12 (and marginal total x1 .) above the
main diagonal. The n0 – n individuals with tallies along the
main diagonal whose responses do not change (i.e., correctly anticipating the actual grade) are discarded. The
null hypothesis is rejected if the McNemar test statistic
M D p.xffi.xffi12ffiffiffi¡ffiffiCffixffiffix21ffiffiffi/ffi/ffi > Z1 ¡ a=2 or if M < Za=2 ; where Z1 ¡ a=2 and
12
21
Za=2 are the respective upper tail and lower tail critical values from the standardized normal distribution at a chosen
level of significance, a (Marascuilo & McSweeney, 1977).
Test for Equality of Proportions: Two Independent
Populations and Decision Rule
This test procedure is found in all introductory business statistics textbooks. Reject the null hypothesis of equal proportions in two independent populations if the test statistic
ZSTAT > Z1 ¡ a=2 or if ZSTAT < Za=2 ; where Z1 ¡ a=2 and Za=2
are the respective upper tail and lower tail critical values
from the standardized normal distribution at a chosen level
of significance a.
Test for Equality of Proportions: c Independent
Populations and Decision Rule
This test procedure is found in almost all introductory business statistics textbooks. Reject the null hypothesis of equal
proportions in c independent populations if the test statistic
x2STAT > x21 ¡ a;.c ¡ 1/ ; the upper tail critical value from the
chi-square distribution with .c ¡ 1/ degrees of freedom at a
chosen level of significance a.

Dokumen yang terkait