28940012 Uttam Kumar Dey Assignment3

Uttam Kumar Dey

Marketing Analytics
Assignment 3
Part 1
Do box office rankings affect a movie’s long-term box office performance?
To understand the effect of a movie’s box office rankings has on its box office performance, Nonlinear
Estimation models can be used to interpret the relationships and make recommendations.
First, let us familiarize ourselves with the relationship, in the movies database, we are trying to analyze.
This relationship is as below:
Subsequent Earnings = b0*exponential(b1*ReleaseRank)
The above equation tells us that the relationship between Substantial Earnings and Release Rank is very different when the movie is ranked higher than when the movie is ranked lower. Hence, we use the logarithmic function to make it easier to use regression and depict the equation as below:
Log (Subsequent Earnings) = b0 + b1*Release Rank, or
Log (Subsequent Earnings) = 16.823 -0.699*Release Rank
This tells us that keeping all other variables that affect Subsequent Earnings, it will decrease by around
0.7th of its previous value every time the movie’s rank slips down by one position.
However, 16.823 indicates that the movie will still earn some unis of Subsequent Earnings even if it does not secure a rank, which for the given data may not hold true. It also indicates that there will be an increase of Subsequent Earnings with an increase in rank (here getting a …3, 2, 1 rank means increase in rank) (p<2.2e-16). A brief summary of the regression is listed below:
Casual Regression Model log(Subsequent Earnings) ~ Release Rank
Residuals
Min
-10.2718
Coefficients

1Q
-1.1163

Median
0.2456

3Q
1.2677

Max
8.6832

Estimate
Std. Error t value
Pr(>|t|)
Intercept
16.82333
0.03823
440.1
<2e-16 ***
Release Rank
-0.69905
0.00654
-106.9
<2e-16 ***
Residual standard error:
2.006 on 7843 degrees of freedom
Multiple R-squared:
0.5929
Adjusted R-squared:
0.5929
F-statistic:
1.142e+04 on 1
DF:

7843

P-Value:

< 2.2e-16

1|Page

Uttam Kumar Dey

Part 1C
Confounding factors that