Survey says: BCA released far fewer tickets on Tuesday than originally advertised

SL_SWTickets_brownbsa

In the aftermath of Monday morning’s ticketing fiasco, the Brown Concert Agency scrambled to deploy a new ticketing release schedule. In an announcement the BCA put up on its blog Monday afternoon, the new schedule was that 1250 tickets would be released later on Monday and then another 1250 — again to each concert — would be released yesterday, Tuesday. We quote directly from that announcement:

1250 tickets per concert will go on sale tonight (4/15) @ 8pm.

1250 tickets per concert will go on sale tomorrow morning (4/16) @ 7am.

Based on a Google poll deployed yesterday afternoon by BlogDailyHerald, there is excellent reason to believe that the latter statement turned out not to be true, and that anywhere between zero and around 485 tickets to the Saturday concert were actually released at 7 a.m. on Tuesday, a far cry from the 1250 advertised. Read our full analysis after the jump.

At 7:00 a.m. on Tuesday, upon trying to secure tickets to the Saturday concert, many encountered an error message indicating that Saturday tickets had sold out. At 7:02 by this blogger’s count, and 7:00 by others’, all of the 3200 tickets to the Saturday concert seemed to be gone. Not one person we spoke to was successful in securing Saturday concert tickets yesterday, in spite of the BCA’s announcement that 1250 Saturday tickets would be reserved for 7 am release. In an effort to understand why, we created and launched an online survey at 4 pm on Tuesday to test the validity of the BCA’s original announcement.

By Tuesday evening, we had collected enough responses to conclude with a high level of confidence that nowhere close to the advertised 1250 Saturday tickets were released at 7 a.m. on Tuesday.  In fact, with 95% confidence, we can say that the actual chances of getting a ticket Tuesday morning were between 0 and 8%. Even if we assume that every undergraduate at Brown had been vying for one of those tickets, which is impossible given that many had already secured a ticket, we find that the actual number of Saturday concert tickets available at 7 a.m. was anywhere between 0 and 485.

BCA did not respond to our request for comment.

________

Methodology:

The survey was launched as a Google Form at 4 p.m. yesterday. It was promoted through the BlogDailyHerald Facebook page, and respondents needed to log in with their Brown University email addresses in order to register their answers. 64 respondents were logged.

The survey asked three questions:

  1. Did you try to buy a ticket to the Saturday concert on Tuesday, April 16, at 7 a.m.?
  2. If you did try to buy a Saturday ticket, were you able to get one?
  3. If you did try to buy a Saturday ticket, at what time (please try to be specific)?

We pared down our data set to only include responses to which the answer to the first question was “Yes.” After this paring was complete, we examined the 40 remaining answers to the second question. Every single one of these responses was “No,” except for three. One of these exceptions listed its time of purchase (in response to the 3rd question) as 8:00, and the other two listed their respective times of purchase as 20:00 — these were ruled out, as we deduced that these attempts were not, in fact, made at 7 a.m. on April 16, but rather at 8 p.m. on April 15. It is likely that these respondents did not read the first question completely. This left us with 37 respondents, all of whom tried to purchase tickets at 7 a.m. on April 16th, and none of whom succeeded.

We then applied the rule of three for a binomial proportion confidence interval to find the 95% confidence interval for the rate of successful ticket acquisition throughout the population that made attempts to acquire Saturday tickets on Tuesday, April 16. This test indicated that the true proportion was, with 95% certainty, between 0/37 and 3/37, or about 0% to 8%.

Find the full results of the survey at this link.

Analysis contributed by Rakesh Patel ’13.

4 Comments

  1. Alexa Van Hattum

    Just pointing out- this survey has doesn’t have much statistical validity. Confidence intervals can only be used if based off of a simple random sample… and a survey posted on Facebook doesn’t meet this requirement. If there were people who got Saturday tickets at 7 am Tuesday, most of them probably went back to sleep and didn’t think about it again. The people who didn’t get tickets are the ones who are much more likely to go on Facebook and put the effort into taking this survey. Yes, these numbers still suck, and the tickets released could very well have been less than expected, but you can’t really rely on a confidence interval to show this.

  2. Charlie

    Two words: response bias.

Leave a Reply