(a)

Correlation Matrix Using Data Analysis tool:

Overall Comfort Amenities In-House Dining
Overall 1
Comfort 0.109459661 1
Amenities 0.639553781 0.011993901 1
In-House Dining 0.713669596 0.045137454 -0.02748379 1

Based on the output of the correlation matrix ‘In-House Dining’ has the highest degree of
correlation with overall satisfaction. The Pearson’s Correlation Coefficient value is in this case, r
= 0.713669 which indicates a strong positive correlation.

(b)

Overall Vs Comfort
100

95

90
f(x) = 0.143638432151104 x + 75.8147639793544
Overall

R² = 0.0119814174189037

85

80

75
86 88 90 92 94 96 98 100

Comfort
 Overall Vs Amenities
100

95

f(x) = 0.276942821660145 x + 65.8897553261017

R² = 0.409029039414936
90
Overall

85

80

75
60 65 70 75 80 85 90 95 100 105

Amenities

Overall Vs In-House Dining

100

95

f(x) = 0.462422706467535 x + 47.5464767524393

R² = 0.509324291651518
90
Overall

85

80

75
70 75 80 85 90 95 100 105

In-House Dining
SUMMARY OUTPUT

Regression Statistics
Multiple R 0.974094571
R Square 0.948860233
Adjusted R Square 0.939271526
Standard Error 0.844679296
Observations 20

ANOVA
df SS MS F Significance F
Regression 3 211.8103502 70.60345 98.95602 1.52586E-10
Residual 16 11.4157298 0.713483
Total 19 223.22608

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%
Intercept 14.30710329 7.799085397 1.834459 0.08525 -2.22621917 30.84042575 -2.226219174 30.84042575
Comfort 0.090109644 0.074270603 1.213261 0.242638 -0.067337 0.247556288 -0.067336999 0.247556288
Amenities 0.285258271 0.024492587 11.64672 3.17E-09 0.233336307 0.337180235 0.233336307 0.337180235
In-House Dining 0.472145619 0.036683874 12.87066 7.41E-10 0.39437928 0.549911958 0.39437928 0.549911958

The Multiple Regression Model is:

y=0.09011 x1 +0.28525 x 2 +0.47215 x 3+14.3071

Here,
Y = Overall Satisfaction
X1 = Comfort
X2 = Amenities
X3 = In-House Dining
Interpretation:
Based on the ‘Significance’ value for the overall model from the ANOVA table output since the
Alpha value is less than 5%, hence the model is valid. Again, based on the final table, the p-value
for ‘Comfort’ is more than 5% making the variable insignificant. The other two variables are
significant, and both are affecting the ‘Overall Satisfaction’ positively as the coefficients are
positive for both of the variables.
(c)
Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%
Intercept 14.30710329 7.799085397 1.834459 0.08525 -2.22621917 30.84042575 -2.226219174 30.84042575
Comfort 0.090109644 0.074270603 1.213261 0.242638 -0.067337 0.247556288 -0.067336999 0.247556288
Amenities 0.285258271 0.024492587 11.64672 3.17E-09 0.233336307 0.337180235 0.233336307 0.337180235
In-House Dining 0.472145619 0.036683874 12.87066 7.41E-10 0.39437928 0.549911958 0.39437928 0.549911958

The slope of Comfort = 0.09011

The Slope of Amenities = 0.28525
The Slope of In-House Dining = 0.47215
Intercept = 14.3071
All three variables have positive slopes which indicate that any increase in any of the variables
will increase the ‘Overall’ value. For a 1 unit increase in the comfort value, the overall will
increase by 0.09011units. For a 1 unit increase in the Amenities value, the overall will increase
by 0.28525 units. For a 1 unit increase in the In-House Dining value, the overall will increase by
0.47215 units. The intercept is positive on the y-axis means, even if all the three independent
variable values are ‘zero’ there will be an overall score of 14.3071.
(d)
Regression Statistics
Multiple R 0.974094571
R Square 0.948860233
Adjusted R Square 0.939271526
Standard Error 0.844679296
Observations 20

The value of Multiple Regression R squared is = 0.974095 and the adjusted R squared value is =
0.93927
The Adjusted R-squared value is a better measure for multiple linear regressions.
The value implies that the model can account for 93.927% of the variation in the dependent
variable (Overall Satisfaction). This can be regarded as a high Adjusted R-squared value and
indicates that the model almost accurately describes the data.
(e)
Regression analyses may not be able to fully represent the intricacy of the connection between
the independent and dependent variables if only one independent variable is taken into account.
This may result in the following restrictions:
Firstly, Lack of explanatory power: A model without explanatory power may not have enough
independent variables to adequately explain the variance in the dependent variable. Secondly, it
does not give the full picture of the model. Thirdly, the model’s R square value with one variable
can be much lower than the one considered with an actual number of independent variables.
Referring to the answers in (b) and (c), we can see that the R-square output in individual variable
graphs are all lower than the overall adjusted R-square value of the final model.