Академический Документы
Профессиональный Документы
Культура Документы
variables the number of ATM transactions, the number of other services used, whether the
individual has a debit card, and whether interest is paid on the particular account, write a
report indicating which of the variables seem related to the account balance and how well
they explain the variation in account balances. Should all of the independent variables
proposed be used in the analysis or can some be dropped?
By using regression analysis we can now estimate or predict the account balances if we
know the number of ATM transactions, the number of other services used, whether the individual
has a debit card, and whether interest is paid on the particular account.
Y=174.9+90.47(Trans)+59.99(Services)+161.01(Debit)+218.83(Interest)
Model Summary
Adjusted R
Model
R
.765a
R Square
Square
.585
.555
398.389
ANOVAb
Sum of
Model
1
Squares
Regression
Residual
df
Mean Square
1.229E7
3073046.498
8729246.940
55
158713.581
2.102E7
59
Total
Sig.
.000a
19.362
Coefficientsa
Model
1
Unstandardized
Standardized
Coefficients
Coefficients
Std. Error
(Constant)
174.920
167.053
Transaction
90.471
12.452
Services
59.992
Debit
Interest
Beta
Sig.
1.047
.300
-159.862
509.701
.651
7.265
.000
65.516
115.426
27.427
.199
2.187
.033
5.026
114.957
161.017
108.416
.135
1.485
.143
-56.254
378.287
218.834
119.929
.163
1.825
.073
-21.508
459.177
Thecoefficient is positive and shows a direct relationship between account balance and number
of ATM transactions. If we increase the number of ATM transaction by 1 and hold other
independent variables constant, we can estimate an increase of 90.47$ in the account balance.
The number of other services used variable also shows direct relationship between the
account balance and the number of other services used. It is logical that individuals with a higher
account balance use more services. So for each additional service used we estimate an increase
of 59.99$ in the account balance.
The third variable is number of other services used, our regression analysis shows that
there is positive relationship between individuals who have debit card and account balances. The
data shows if the individual has a debit card then it will increase the account balance by 161.01$.
The fourth variable is whether interest is paid on the particular account. Again our data
shows positive and direct relationship between account balance and whether interest is paid on
the particular account. So the data dhows if there interest is paid in a particular accountthen we
can estimate there will be an increase of 218.83$ in the account balance.
However, not all the independent variables should be used because not all of them are
effective predictors of the account balance. In order to determine which independent variable to
tale out we look at the p-value, if the p-value is greater than the .05 significant levels we
conclude it is not a good predator. Looking at the table we see the variables debit card and
interest their p-value are higher than the significant level. We remove the variable debit and we
run the regression analysis again.
Coefficientsa
Unstandardized Coefficients Standardized Coefficients
Model
1 (Constant)
Std. Error
Beta
Sig.
Lower Bound
Upper Bound
228.688
164.829
1.387 .171
-101.505
558.881
Transaction
89.785
12.577
64.590
114.980
Services
67.663
27.225
13.125
122.201
Interest
178.305
118.033
-58.144
414.754
We observe the p-value associated with interest is still greater than the significant level.
So next we remove the interest variable and run the regression analysis again.
Coefficientsa
Unstandardized Coefficients Standardized Coefficients
Model
1 (Constant)
Std. Error
Beta
Sig.
Lower Bound
Upper Bound
258.292
165.490
1.561 .124
-73.096
589.681
Transaction
91.433
12.670
66.062
116.804
Services
67.883
27.529
12.757
123.009
We observe that the Adjusted R Square values have declined. Using all four
independent variables the Adjusted R Square value was .555. With the two nonsignificant variables removed the Adjusted R Square value is .53. This means that
the two independent variables account for 53% of the variation in the account
balance.