1

1. Introduction
The purpose of this report is to analyse sample data related to the compliance with
corporate governance. The data of Environmental Accounting Compliance (EAC) scores are
corrected from 50 accounting firms across the world over the three years, this year, last year,
and the year before. The two main objectives are to find the difference in EAC scores
throughout the period, and the relation between the university degree and EAC scores. This
report will analyse the data using various statistical testing methods such as ANOVA,
statistical inferences about two populations (when the population variances are unknown)
difference between two population proportions and a chi-squared test of a contingency table to
address the main objectives mentioned above.

2. Relation between degree and country type

2.1. Objective
The purpose of this analysis is to find whether the percentage of accountants in each firm who has
at least one university degree in accounting related to country type.
The data is qualitative. This problem combination indicates that the statistical procedure to be
employed is the chi-squared test of contingency table.
The Hypotheses are:

The two variables are independent (i.e. the percentage of accountants

in each firm with accounting degree is independent from country type).

The two variables are dependent.

2.2. Statistical Analysis
The data table is rearranged and shown below to compare the two variables: the percentage of
accountants in each firm who has at least one university degree in accounting, and country type.
Country classification
Acc-Dgr% at time t 1 2 3 TTL
Less than 50% 0 6 17 23
At least 50% 13 9 5 27
TTL 13 15 22 50

Then frequency for a contingency table is decided, according to
size sample
total) (row total) (column i j
e
ij

=
Country classification
2

Acc-Dgr% at time t 1 2 3 TTL
Less than 50% 0(5.98) 6(6.9) 17(10.12) 23
At least 50% 13(7.02) 9(8.1) 5(11.88) 27
TTL 13 15 22 50

The value of the test statistic is

= =

=
c
j
r
i ij
ij ij
e
e o
1 1
2
2
) (
_
88 . 11
) 88 . 11 5 (
12 . 10
) 12 . 10 17 (
1 . 8
) 1 . 8 9 (
9 . 6
) 9 . 6 6 (
02 . 7
) 02 . 7 13 (
98 . 5
) 98 . 5 0 (
2 2 2 2 2 2

=
95315 . 19 =
2
_ is chi-squared distributed with 2 ) 1 3 )( 1 2 ( ) 1 )( 1 ( . . = = = c r f d
With the 10% significance level, rejection region is 61 . 4
2
) 1 3 )( 1 2 ( , 10 . 0
2
) 1 )( 1 ( ,
2
= = >

_ _ _
o c r

With the 5% significance level, rejection region is 99 . 5
2
) 1 3 )( 1 2 ( , 50 . 0
2
) 1 )( 1 ( ,
2
= = >

_ _ _
o c r

With the 1% significance level, rejection region is 21 . 9
2
) 1 3 )( 1 2 ( , 010 . 0
2
) 1 )( 1 ( ,
2
= = >

_ _ _
o c r

Since , 61 . 4 99 . 5 21 . 9 95315 . 19
2
> > > = _

is rejected with all 1%, 5%, and 10% significance levels.

2.3. Conclusion:
The two variables are dependent. That is, the percentage of accountants in each firm who has at least
one university degree in accounting is related to country type.
2.4. Comment:
The reason why we have the
2
_ value as high as 19.95 is mainly because there are significant
differences between country type 1 and 3 as to the percentage of the accountants who have one
university degree in accounting. The tendency shows that the companies from the first world countries
are more likely to have the accountants with high qualification than the companies from the third
world countries.

3. Difference of EAC scores over three years

3.1. Objective
3

The objective is to identify whether there is evidence of different environmental compliance scores
(EAC) over the three-year period. To achieve this objective, two-way ANOVA with randomised blocks
is employed, since:
The problem objective is to compare three populations.
The data are quantitative.
The EAC scores of each firm can be considered a treatment.
Each three data point (per firm) can be blocked because they belong to the same firm.
This procedure eliminates the variability of EAC scores within firms and helps detect
differences in the mean EAC scores between three years.
3.2. Checking of assumptions
Before proceeding with the test, the following assumptions need to be made:
1. The populations are normally distributed
2. The variances of the populations are equal
The following histograms show that it is reasonable to assume that each of the
populations is approximately normally distributed.


0
5
10
15
F
r
e
q
u
e
n
c
y

Bin
EAC-scores this year
Frequency
0
5
10
15
F
r
e
q
u
e
n
c
y

Bin
EAC-scores last year
Frequency
4


The sample variance of EAC scores in each of the three years.
EAC-Scrs this
year (t)
EAC-Scrs
last year (t-1)
EAC-Scrs the
year before
last year (t-2)
77.09 76.68(min) 91.79(max)
Since 5 . 1 20 . 1
68 . 76
79 . 91
2
3
< = =
s
s
, therefore the variances of the populations are equal.

0
5
10
15
F
r
e
q
u
e
n
c
y

Bin
EAC-scores last year
Frequency
5

3.3. Statistical analysis

Testing Treatment means,
Hypotheses:
differ. means treatment least two at : H
: H
A
3 2 1
0
= =

Test statics:
MSE
MST
F =
Testing block means,
Hypotheses:
differ. means block least two at : H
equal all are means Block : H
A
0

Test statics:
MSE
MSB
F =

1
=
k
SST
MST
| |
2
k
1 j
j
) x T x ( b SST

=
=
From the data table,

=63.110,

=61.638,

=60.324,

61.6907
2 2 2
) 6907 . 61 324 . 61 ( 50 ) 6907 . 61 038 . 61 ( 50 ) 6907 . 61 110 . 63 ( 50 SST + + =
2529 . 194 =
Therefore, 12647 . 97
1 3
2529 . 194
1
=

=
k
SST
MST
1 +
=
b k n
SSE
MSE

=
=
b
i
i
X B X k SSB
1
2
) ] [ (
2 2 2
61.6907) - 53.967 ( 3 .... 61.6907) - 43.367 ( 3 61.6907) - 43.467 ( 3 + + + =
88 . 11774 =

= =
=
k
j
b
i
ij
X x Total SS
1 1
2
) ( ) (

2 2 2
) 6907 . 69 2 . 55 ( .... ) 6907 . 69 0 . 45 ( ) 6907 . 69 8 . 44 ( + + + =
6


2 2 2
) 6907 . 69 0 . 54 ( .... ) 6907 . 69 0 . 43 ( ) 6907 . 69 6 . 45 ( + + + +

2 2 2
) 6907 . 69 7 . 52 ( .... ) 6907 . 69 1 . 42 ( ) 6907 . 69 0 . 40 ( + + + +
61 . 12226 =
Since: SSE SSB SST Total SS + + = ) (
Therefore, SSB SST Total SS SSE = ) (
4737 . 257 88 . 11774 2529 . 194 61 . 12226 = =
Therefore,
1 50 3 150
4737 . 257
1 +
=
+
=
b k n
SSE
MSE
627283 . 2 = ,
and 3037 . 240
49
88 . 11774
1
= =

=
b
SSB
MSB
Test statistics for treatment is: 96841 . 36
627283 . 2
12647 . 97
= = =
MSE
MST
F
Test statistics for blocks is: 46471 . 91
627283 . 2
3037 . 240
= = =
MSE
MSB
F
Testing the mean responses for treatments
F > Fa,k-1,n-k-b+1
With 5% significance level, the rejection region is F > F0.05, 2, 98
With 1% significance level, the rejection region is F > F0.01, 2, 98
Since F=36.96841>4.82>3.09, null hypothesis is rejected with the significance level of both 5%
and 1%.
There is sufficient evidence to indicate that differences exist among the EAC scores between
the three years.

- Testing the mean response for blocks
F> Fa,b-1,n-k-b+1
With 5% significance level, the rejection region is F > F0.05, 49, 98
With 1% significance level, the rejection region is F > F0.01, 49, 98
Since F=91.46471>1.74>1.48, null hypothesis is rejected with the significance level of both 5%
and 1%.
7

There is enough evidence to conclude that the block means differ. This means the design of
the experiment, which is to remove the variability due to the block differences, is suitable.
3.4. Conclusion:
There is sufficient evidence of different environmental compliance scores (EAC) over the three-year
period.
3.5. Comment:
Major differences over the three years are found especially in Firm5, 10, 19, which has the
variance of 18.96, 29.81, and 12.60, respectively. Those firms are the major reason for the F value of
treatment to be big enough to reject the hypothesis that the treatment means are all equal.
Since the difference in the EAC scores over the three years is evidenced, and the mean scores have
been slightly increased, from 60.32 in the year before to 61.64 last year and 63.11 this year, it can be
said that the compliance level of the accountants through the world has been improved.




8

4. The shift of EAC scores since last year
4.1. Objective
The objective is to identify if that environmental compliance scores (EAC) for this year
(time t) is greater than last year (time t-1) among the 50 selected accounting firms.
For this purpose, t-statistic with matching pair is employed, since:
The data are quantitative.
Matching pair will eliminate the variability within firms and help to detect the
difference over the two year period.
4.2. Statistical analysis
The hypotheses test for this problem is
H
0
:
D
= 0
H
A
:
D
> 0

D = 1 2
nD = n1 = n2
The t-statistic is
D D
D D
n s
x
t

=

50 575584 . 1
0 4720 . 1
=
=6.6062

The rejection region is:
t > ta,d.f. with d.f. =50-1=49.

If a = .10, t.10, 49 =1.299
Since 6.6062 > 1.729, there is sufficient evidence in the data to reject the null hypothesis in
favor of the alternative hypothesis with 10% of significance level.

If a = .05, t.05, 49 =1.676
9

Since 6.6062 > 1.676, there is sufficient evidence in the data to reject the null hypothesis in
favor of the alternative hypothesis with 10% of significance level.

If a = .01, t.01, 49 =2.403
Since 6.6062 > 2.403, there is sufficient evidence in the data to reject the null hypothesis in
favor of the alternative hypothesis with 10% of significance level.


With significance level of 1%, the environmental compliance scores (EAC) for this year (time t) is
greater than last year (time t-1) among the 50 selected accounting firms.

4.3. Checking the required condition:
The histogram shown below confirms that assuming the normality of the differences is
reasonable.


0
2
4
6
8
10
12
14
-2 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 More
F
r
e
q
u
e
n
c
y

Bins
Histogram
10

4.4. Conclusion:
It is evidenced that the environmental compliance scores (EAC) for this year (time t) is greater
than last year (time t-1) among the 50 selected accounting firms.


11

5. Relation between degree and EAC score this year
5.1. Objective
The problem objective is to identify whether this year environmental compliance scores (EAC) of
accounting firms which has less than 50% accountants with at least one university degree in
accounting is lower than this year environmental compliance scores (EAC) of accounting firms which
has at least 50% accountants with at least one university degree in accounting.
The data table is rearranged and sorted according to the percentage of accountants in each firm
who have at least one university degree in accounting.
Acc-Dgr% at time t
=Less than 50%
Acc-Dgr% at time t
=at least 50%
Accounting firm EAC-Scrs this year (t) Accounting firm EAC-Scrs this year (t)
Firm 1 44.8 Firm 6 52.8
Firm 2 45.0 Firm 10 61.3
Firm 3 50.9 Firm 14 57.3
Firm 4 44.0 Firm 17 58.2
Firm 5 58.7 Firm 21 59.6
Firm 7 51.1 Firm 24 64.8
Firm 8 54.7 Firm 25 65.7
Firm 9 57.0 Firm 27 62.6
Firm 11 56.0 Firm 28 68.2
Firm 12 58.8 Firm 29 71.8
Firm 13 58.2 Firm 30 69.9
Firm 15 54.5 Firm 32 71.7
Firm 16 59.4 Firm 34 71.3
Firm 18 59.8 Firm 35 81.6
Firm 19 64.9 Firm 37 62.5
Firm 20 62.5 Firm 38 63.7
Firm 22 61.7 Firm 39 65.3
Firm 23 61.9 Firm 40 66.1
Firm 26 64.6 Firm 41 68.5
Firm 31 69.6 Firm 42 69.8
Firm 33 68.8 Firm 43 70.1
Firm 36 56.5 Firm 44 72.6
Firm 50 55.2 Firm 45 73.9
Firm 46 74.2
Firm 47 75.3
Firm 48 77.9
Firm 49 80.2
AVERAGE 57.3304 AVERAGE 68.0333
STANDARD
DEVIATION 6.9514
STANDARD
DEVIATION 7.0476
To address this problem, t-statistics is employed, since:
The data are quantitative.
There are two independent samples.
12

The parameter to be estimated is the difference between two means (1 2):
the mean environmental compliance scores (EAC) of accounting firms which has less
than 50% accountants with at least one university degree in accounting (1)
the mean environmental compliance scores (EAC) of accounting firms which has at
least 50% accountants with at least one university degree in accounting (2)
The population variances are unknown, but it appears that the population variances is equal
since the sample variances:




5.2. Statistical analysis
The claim to be tested is whether a difference between the two exists
H
0
: (
1
-
2
) = 0
H
a
: (
1
-
2
) < 0
Assuming that the populations are normally distributed, the t-test for equal variances
is employed:
)
1 1
(
) ( ) (
2 1
2
2 1
n n
s
X X
t
p
+

=
2 1


With
2
) 1 ( ) 1 (
2 1
2
2 2
2
1 1 2
+
+
=
n n
s n s n
s
p
=
48
6692 . 49 ) 1 27 ( 3213 . 48 ) 1 23 ( +
=49.0514


)
1 1
(
) ( ) (
2 1
2
2 1
n n
s
X X
t
p
+

=
2 1



3856 . 5
)
27
1
23
1
( 0514 . 49
0 ) 0333 . 68 3304 . 57 (
=
+

=

48 2 . .
2 1
= + = n n f d

The rejection region is
6692 . 49 , 3213 . 48
5 . 1 01 . 1
9814 . 6
0476 . 7
2
2
2
1
1
2
= =
< = =
s s
s
s
13

With 90% confidence level,
1.676 t t t
0.05,48 2
~ = >
/ , d.f.

or t < -ta/2,d.f. = -t0.05,48 = -1.676
or t > ta/2,d.f. = t0.05,48 = 1.676
Since t = -5.3856 < -1.676, there is sufficient evidence to reject the null hypothesis with
90% confidence level.
With 95% confidence level,
2.009 t t t
0.025,48 2
~ = >
/ , d.f.

or t < -ta/2,d.f. = -t0.025,48 = -2.009
or t > ta/2,d.f. = t0.025,48 = 2.009
Since t = -5.3856 < -2.009, there is sufficient evidence to reject the null hypothesis with
95% confidence level.
With 99% confidence level,
2.678 t t t
0.005,48 2
~ = >
/ , d.f.

or t < -ta/2,d.f. = -t0.005,48 = -2.678
or t > ta/2,d.f. = t0.005,48 = 2.678
Since t = -5.3856 < -2.678, there is sufficient evidence to reject the null hypothesis with
99% confidence level.

14









5.3. Conclusion
There is a sufficient evidence to infer, with the 1%, 5%, and 10% significance levels, that this
year environmental compliance scores (EAC) of accounting firms which has less than 50%
accountants with at least one university degree in accounting is lower than this year
environmental compliance scores (EAC) of accounting firms which has at least 50% accountants
with at least one university degree in accounting.
5.4. Comment
This finding infers that the companies with more accountants who have university degree perform better in
EAC test than the companies with fewer accountants who have university degree.

5.5. Checking the required condition
The histogram below confirms that it is reasonable to assume the populations are
normally distributed.

5.6. Conclusion
There is a sufficient evidence to infer, with the 1%, 5%, and 10% significance levels, that this
year environmental compliance scores (EAC) of accounting firms which has less than 50%
accountants with at least one university degree in accounting is lower than this year
0
5
10
15
44 49.37 54.74 60.11 65.49 70.86 76.23 More
F
r
e
q
u
e
n
c
y

Bin
Histogram
Frequency
2.678 -5.3856
15

environmental compliance scores (EAC) of accounting firms which has at least 50% accountants
with at least one university degree in accounting.
5.7. Comment
This finding infers that the companies with more accountants who have university degree
perform better in EAC test than the companies with fewer accountants who have university
degree.
6. Conclusion

Overall, the sample data shows that there is a significant difference between the EAC scores over the
years and a relation between the country type and the university degree. In fact with each statistical
test, there is a 99% certainty that the result is correct, meaning that the Board of the International
Union of Accountants is collecting very accurate information, and that the sample errors is very low.
An important finding is that the EAC scores are significantly related to the university degree, that is,
the companies with more accountants who have at least one university degree is outperforming the
companies with less such accountant. Therefore, it should be encouraged to employ an accountant with
university degrees since they are expected to have higher compliance levels. Another implication is that
the company in the third world counties should be encouraged to employ more accountants with
university degrees since the data show that university degree is an important factor for compliance level
and the percentage of accountants in a firm who have university degree is less in the third world countries
compared to the companies from the first world countries.