Monday, June 17, 2013

From the Vault: SYTYCD Season 8 Post-Mortem

Ed. note: Way, way back in the day -- 2011! -- I crunched some numbers for contestant ages after season 8. I thought I posted it, but I guess not! I have no idea if I had anything more to say and thus didn't post it, or if I posted it and accidentally reverted it to a draft, or what. So consider this a blast from the past.

As promised, if belated. (Ed.: lol. That's from the original post.)

Earlier this season, a reader (I think Sara -- can't find the specific comment now) speculated that the Australian show has a higher level of dance possibly because it casts older contestants. I compiled the numbers, and unfortunately, I can't draw any firm conclusions from them because three seasons is fairly useless, statistically speaking.

But then a similar discussion bubbled up here, and I wanted to get in on the action, and looking specifically at the US show in isolation is more useful. (I'll put the Oz data below the jump, and some lightweight analysis with it to explain why the data is of little use.)

Season Median Mean SD
1 22.00 22.88 3.30
2 20.00 21.05 2.86
3 21.00 21.35 2.66
4 22.00 21.65 3.18
5 21.00 22.10 2.86
6 20.00 21.35 3.56
7 21.00 21.55 2.25
8 20.00 21.40 3.68
Total 21.00 21.64 3.09

First, I need to mention that the stats compiled by the TWOP posters (and I have to get a dig in at the site again -- I usually loathe its forums for being the pettiness and sanctimony of its posters, but that thread and the philosophy thread have a uniquely positive signal-to-noise ratio) focus only on the top 10 of each season, which is fine for looking at elimination patterns, but I think that in order to suss out any age biases, we ought to look at all the contestants who reach the competition because the casting decisions act as bottlenecks for the top 10.

No surprise, the show skews very young:

(click for higher resolution)

The 23-and-under crowd represents fully three-quarters of the dancers on the show. Almost half can't legally drink.

Now, here are individual charts, season-by-season, with the medians highlighted:

Obviously, all the charts tilt to the left, so trend-wise, the youth of the contestants isn't a recent phenomenon. (The first and fourth seasons come closest to being somewhat balanced, though.) Season 8 is crazy -- we all knew how imbalanced it was across all sorts of categories, but seeing just how heavy it was in 19-year-olds is good for a laugh.

Perhaps more interesting would be to compile the ages of dancers in the bottom for each week, but that'd be a fairly large undertaking.

Season Median Mean SD
1 22 22.65 3.98
2 24 22.45 3.07
3 21 22.30 3.79
Total 22 22.47 3.57

Looking at the mean averages alone doesn't give a great picture; Oz averages at about 22 years old, while the US averages at 21. Even if we look at the standard deviation for each program (standard dev basically indicating how spread out the data are, and two-thirds of all the data is within one standard deviation), they differ by only half a year. The median is marginally more interesting (median being the midway point where half the data is on either side), but only if we pick out one particular season, i.e. Oz season 2. Twenty-four years old! Definitely older. But that's just one season, and the two other seasons look fairly similar to any given US season (although the SDs are notably higher). So: inconclusive, because Oz just doesn't supply enough data.


Anonymous said...

"Season 8 is crazy -- we all knew how imbalanced it was across all sorts of categories, but seeing just how heavy it was in 19-year-olds is good for a laugh."

I'm not sure I see how S8 stands out from these graphs you made. According to them, S8 had 9 teenagers, the same number of teenagers S6 had. In fact, S2, S3, and S4, also had a similar number of teenagers (8, 10, 8, respectively). Is there something funny about having so many 19 year olds compared to 18 year olds?

Anonymous said...

oof! my apologies, S3 actually has 6 teenagers, not 10! >.< My misread, sorry. :(

Leee said...

You know, you're probably right that the numbers don't necessarily signal anything weird, and that I was reaching for a conclusion -- which may explain why I decided not to post this originally. The main reason why it went up at all was to get offer up the graphs themselves, not so much the commentary that they went with.