In this section, we will analyze the economic impact of diversity in regards to the gender of the cast and crew as well as the ethnicities of the cast. What we would like to find is if there is a significant correlation between how well a film performs in the box office and how diverse the cast and crew is. To do so, we will test our null hypothesis that diversity in the cast and crew has no affect on a film’s profitabiliy by performing both independent and paired t-tests to assess significance and see if there is a positive or negative statistic.
In order to quantify diversity, we decided to create different diversity scores. In regards to gender, we calculate the percentage of cast and crew members who identified as female to the total number of cast and crew. In regards to ethnicity, we calculate the percentage of cast members who do not identify as Caucasain to the total number of cast members. We then created an overall diversity score that equally weighs the gender diversity score and ethnicity diversity score of each film to compute the average of the two. Each diversity score ranges between 0 to 1, 1 being the most diverse with all female and no Caucasain cast or crew.
After conducting a multitude of t-tests, our results is as follows:
- OVERALL OVERALL:
- Ttest_indResult(statistic=-5.217337297586662, pvalue=1.8308840767272648e-07)
- Paired t-test statistic = -5.267764949074156 , pvalue = 1.405914720843303e-07
- Mean difference = -0.014081434303650543
- OVERALL GENDER:
- Ttest_indResult(statistic=-7.402233055319847, pvalue=1.3851604032703372e-13)
- Paired t-test statistic = -7.408244717081399 , pvalue = 1.3694607125081795e-13
- Mean difference = -0.019609662795609637
- OVERALL ETHNICITY:
- Ttest_indResult(statistic=-1.759053764571897, pvalue=0.07858168498393178)
- Paired t-test statistic = -1.7817683338054782 , pvalue = 0.07481342400398133
- Mean difference = -0.008553205811691453
According to these results, we can see that there is a significant negative relationship between how well a film performs in the box office and how diverse the cast and crew are. In other words, we reject our null hypothesis as it is evident that more profitable films have lower diversity rates.
We can tell because both p-values for our independent and paired t-tests for overall diversity as well as gender divserity are less than 0.05, indicating a significant relationship. The t-test statistics for these tests are also highly negative, suggesting that it is highly reliable to conclude a negative relationship between a film’s profitability and its diversity rate.
Interestingly, the p-values for ethnic diversity were both above 0.05, indicating that ethnic diversity has no significant affect on a film’s profitability, however the t-test statistic is not as highly negative, suggesting that this conclusion is not as reliable.
This can be better visualized through graphs: