Out of the 115.6 million homes with televisions in 2014, approximately 7.8 million tuned in for The Bachelor season 19 premier- 7% of all TV viewership at 8pm EST that Monday night. Of these 7.8 million, it is likely that a sizable portion of loyal fans were gambling on the show’s outcome- as measured by a contestant’s longevity before being eliminated based on the contestant’s biographical data (which on the ABC homepage includes a photo, their age, occupation, and hometown). Whether the gambling was formal- there exist a number of online sports books that allow for betting on reality television despite information that can leak prior to the show airing- or informal- pools and drafts that may or may not include a financial reward- the show has potentially produced enough data over its 19 seasons for a savvy gambler to position him/herself statistically.
Given the data available on the contestants before the season premier, and accounting for attractiveness not being readily or objectively quantifiable, an analysis can be performed that looks at the week a contestant was eliminated versus their age, hometown, and occupation.
The data was compiled for seasons 11- 19 and included 239 contestants. A contestant’s outcome was measured to be the week they were eliminated, with a higher number week being a better outcome than a lower one. Elimination weeks ranged from 1 to 11 (for seasons 15-19) and 1 to 9 (for seasons 11-15) with the winner being coded ‘week 11’ (or ‘week 9’) and the runner-up being coded ‘week 10’ (or ‘week 8’).
Age required no coding and was entered directly from the biographical data. The average age was 26.7 years with a high of 36 and a low of 21.
For occupation, I decided to create a binary variable that would be equal to 1 if the contestant’s job required a college degree and 0 if it did not. If it was not obvious that the listed position required a college degree (i.e. ‘account manager’) then it was coded as a 0. This required some judgment on my part and it is possible contestants are incorrectly coded. Overall, 86 out of 238 (36%) contestants were coded as having a college degree.
For hometown, I decided to create a binary variable that would provide a proxy for ‘Southern’. I coded the hometown as 1 if the contestant was from a state that voted for Mitt Romney in the 2012 presidential campaign and as 0 if they were from a state that voted for Barack Obama- international contestants were coded as 0. Overall, 68 out of 238 (29%) contestants originated from a state that supported Romney in 2012.
Next, I wanted to use regression analysis to see how these variables (age, hometown, and occupation) were associated with the number of weeks a contestant remained on the show. Of course, there are many other omitted variables that impact longevity and the choice of ‘bachelor’ impacts the desired attributes of a contestant. This analysis is merely designed to identify past correlations.
First, I ran a regression of age on elimination week:
While the coefficient on age is negative (meaning that as age increases their longevity or week eliminated decreases), it is not statistically significant- even at the 10% level. This means that it is likely that if the true relationship between age and elimination week were 0 (i.e. they are not related) we would find an effect as large as the one we found 19.2% of the time. We therefore fail to reject this null hypothesis and can’t conclude that age impacts a contestant’s longevity on the show.
Second, I ran a regression of occupation on elimination week:
Again, while the coefficient on occupation is negative, it is not statistically significant at the 10% level- in fact this time it is not even close. This means that while there is a very loose negative correlation between holding a college degree and elimination week, it can’t be distinguished from a purely random result.
Third, I ran a regression of hometown on elimination week:
In this regression, there is a statistically significant positive relationship at the 1% level. What does this mean? It means that if being from a state that voted for Mitt Romney had no real impact on your elimination week, we would only find a result as large as what we found in 0.006% of samples. This means we can conclude the following: being from a state that voted for Mitt Romney in 2012 is associated on average with remaining on the Bachelor for an additional 1.085 weeks.
So while being older and educated may not have a true negative relationship with contestant longevity, it appears to have been a very good bet to choose contestants from politically conservative states over the past 8 seasons- and to continue choosing them in the future.
Please fill in the contact form below to have new articles emailed to you directly!