12
these items are not logically expected to have high correla-
tion, for example, the distance of nearest railway siding will
not logically have correlation with quality of coal found in
a coal block on offer.
Thus, having satisfied ourselves that these items on the
scale are not redundant, we go on to use Cronbach’s alpha
to construct a Value variable. As described in the methodol-
ogy section, we converted each of the items on the Likert
scale of 0 to 10 by dividing the observations by their maxi-
mum. The scaled items were used to construct the value
variable, implemented through “alpha” command of Stata
that provides a measure reflecting items validity in measur-
ing Value and in their consistency with each other as well
as the with the constructed Value variable. This measure of
Cronbach’s alpha, as stated in the in the methodology sec-
tion of this paper should be 0.7 or more for a validity. The
result of the test is presented in Table 3.
The Cronbach’s Alpha for the Value variable is observed
to be 0.69, close to the acceptable level of 0.7. This we, thus,
consider appropriate for construction of the Value variable
using all the 8-items on the scale. We also examined the
option of dropping items that have low alpha values but
dropping any of them led to the resulting scale’s alpha val-
ues to be much lower than the acceptable level. Hence, we
have included all the 8 items and built the Value variable to
use for hypotheses testing, results of which are presented in
the subsections below.
Result of Testing the Base Model
We ran the test for the Base Model with two scenarios of
dependent variable being the Bid amount and Excess Bid.
This was to test the hypotheses 1 and 2 that are to esti-
mate the relationship between the dependent variable with
competitive intensity and value. The results of the tests are
presented in Table 4.
The result of tests for Bid amount shows that for
every unit increase in number of bidders, the bid amount
Table 2. Correlation matrix of 8 value drivers
Items SA GR MC GC SR FL PAP D
Surface Area (SA) 1
Geological Reserves (GR) 0.31 1
Mine Capacity (MC) 0.24 0.71 1
GCV (GC) 0.11 0.28 0.34 1
Stripping Ratio (SR) 0.08 0.30 0.40 0.40 1
Proportion of Forest Land
(FL)
0.08 0.07 0.03 –0.02 –0.11 1
Project Affected People
(PAP)
0.21 0.42 0.66 0.20 0.32 0.10 1
Distance to Railway
Siding (D)
0.14 0.14 0.18 0.24 0.21 –0.16 0.1 1
Number of observations for each item is 127
Table 3. Items on value variable construction and Cronbach’s
Alpha
Item
Number of
Observation Sign
Cronbach’s
Alpha
Surface Area 127 +0.68
Geological
Reserves
127 +0.60
Mine Capacity 127 +0.58
GCV 127 -0.65
Stripping Ratio 127 -0.64
Proportion of
Forest Land
127 -0.74
Project Affected
People
127 +0.62
Distance to
Railway Siding
127 +0.70
Scale 0.69
Table 4. Results of tests for Base Model
Dependent Variable →
Bid Amount Excess Bid
Estimate p-Value Estimate p-Value
Competitive intensity (Ni) 142.67 0.010‡ 164.87 0.003‡
Value (V
i )859.29 0.0001‡ 9.46 0.971
Number of observation =127
*p 0.10, † p 0.05, ‡ p 0.01.
these items are not logically expected to have high correla-
tion, for example, the distance of nearest railway siding will
not logically have correlation with quality of coal found in
a coal block on offer.
Thus, having satisfied ourselves that these items on the
scale are not redundant, we go on to use Cronbach’s alpha
to construct a Value variable. As described in the methodol-
ogy section, we converted each of the items on the Likert
scale of 0 to 10 by dividing the observations by their maxi-
mum. The scaled items were used to construct the value
variable, implemented through “alpha” command of Stata
that provides a measure reflecting items validity in measur-
ing Value and in their consistency with each other as well
as the with the constructed Value variable. This measure of
Cronbach’s alpha, as stated in the in the methodology sec-
tion of this paper should be 0.7 or more for a validity. The
result of the test is presented in Table 3.
The Cronbach’s Alpha for the Value variable is observed
to be 0.69, close to the acceptable level of 0.7. This we, thus,
consider appropriate for construction of the Value variable
using all the 8-items on the scale. We also examined the
option of dropping items that have low alpha values but
dropping any of them led to the resulting scale’s alpha val-
ues to be much lower than the acceptable level. Hence, we
have included all the 8 items and built the Value variable to
use for hypotheses testing, results of which are presented in
the subsections below.
Result of Testing the Base Model
We ran the test for the Base Model with two scenarios of
dependent variable being the Bid amount and Excess Bid.
This was to test the hypotheses 1 and 2 that are to esti-
mate the relationship between the dependent variable with
competitive intensity and value. The results of the tests are
presented in Table 4.
The result of tests for Bid amount shows that for
every unit increase in number of bidders, the bid amount
Table 2. Correlation matrix of 8 value drivers
Items SA GR MC GC SR FL PAP D
Surface Area (SA) 1
Geological Reserves (GR) 0.31 1
Mine Capacity (MC) 0.24 0.71 1
GCV (GC) 0.11 0.28 0.34 1
Stripping Ratio (SR) 0.08 0.30 0.40 0.40 1
Proportion of Forest Land
(FL)
0.08 0.07 0.03 –0.02 –0.11 1
Project Affected People
(PAP)
0.21 0.42 0.66 0.20 0.32 0.10 1
Distance to Railway
Siding (D)
0.14 0.14 0.18 0.24 0.21 –0.16 0.1 1
Number of observations for each item is 127
Table 3. Items on value variable construction and Cronbach’s
Alpha
Item
Number of
Observation Sign
Cronbach’s
Alpha
Surface Area 127 +0.68
Geological
Reserves
127 +0.60
Mine Capacity 127 +0.58
GCV 127 -0.65
Stripping Ratio 127 -0.64
Proportion of
Forest Land
127 -0.74
Project Affected
People
127 +0.62
Distance to
Railway Siding
127 +0.70
Scale 0.69
Table 4. Results of tests for Base Model
Dependent Variable →
Bid Amount Excess Bid
Estimate p-Value Estimate p-Value
Competitive intensity (Ni) 142.67 0.010‡ 164.87 0.003‡
Value (V
i )859.29 0.0001‡ 9.46 0.971
Number of observation =127
*p 0.10, † p 0.05, ‡ p 0.01.