I found this on my Facebook feed yesterday: Graph TV. It’s a website that reads the IMDB ratings of TV show episodes and graphs them, with the linear regression showing the ratings trend over each season. If you thought the last season of Dexter was disappointing, you should check out what it looks like on this site. Of course, the first thing I did was type in Doctor Who, and the graph is very interesting. I’ve included a screenshot of it here but I recommend going to the site and looking at the real display, because you can mouse over the dots to see which episodes they are and their individual ratings.
The biggest problem with graphs like this (and statistics) is that if you don’t really understand the mathematics behind them, it’s really easy to misinterpret data and use them to support whatever theory you already believe. Now, I’m not a statistics expert, so I can’t draw a lot of conclusions from the graph, but it’s fun to try.
First, it’s important to understand how this data was collected. Each dot represents an episode’s rating, which is an average of IMDB user ratings. You can view the list of episodes, number of votes, and ratings by clicking on the IMDB link below the graph (or just click here), and you can see that each episode has somewhere between 1,000 and 3,000 ratings. The graph does not take into account any episodes that don’t have an episode number, so the David Tennant specials (“The Next Doctor,” “Planet of the Dead,” “The Waters of Mars,” and “The End of the Time) and the final Matt Smith episodes (“The Day of the Doctor” and “The Time of the Doctor”) are not included in the graph.
It’s also important to note the averages of the ratings, because they are hard to read off the graphs. I’ve calculated them, and here they are, with a few tweaks as noted:
- All 7 series: 8.13
- Series 1: 8.09
- Series 2: 8.02
- Series 3: 8.12
- Series 4: 8.29
- David Tennant specials: 8.14
- Series 5: 8.18
- Series 6: 8.21
- Series 7: 7.97
- Series 7.1: 8.05
- Series 7.2: 7.92
- Series 7 with “The Day of the Doctor” and “The Time of the Doctor”: 8.07
The first thing you probably notice when you see the graph is the marked difference between the Russell T. Davies (RTD) era (series 1-4) and the Steven Moffat era: except for series 2, the RTD series’ regression lines have a strong positive slope, while the Moffat series’ regression lines are much flatter, and series 7 has a negative slope. What does this mean? RTD’s episodes get a lot better at the end of the season, while Moffat’s episodes keep a consistent amount of quality. What happened to series 2? Well, the bottom three episodes, “The Idiot’s Lantern,” “Love and Monsters,” and “Fear Her” are clear outliers: only six episodes in the entire run of the show are rated 7.0 or below, and three of them are these three episodes. While you can’t really just drop them out of the graph, if you ignore them for a moment, you see that the rest of series 2 has the same trend as series 1, 3, and 4 and its ratings average pops up to 8.42, which is better than any other season.
What do these trends tell us about RTD’s episodes? RTD’s style was to present us with a number of adventure episodes with hints about the theme tying the season together (you know, “Bad Wolf,” Torchwood, Harold Saxon), then use the last few episodes to tell the overarching story. So, we see that the quality of the episodes improve as they start to tell this story. The actual non-conforming series in this run is series 3, Martha’s season. Its finale, “Utopia”/”The Sound of Drums”/”The Last of the Time Lords”, is rated much worse (though not badly) than the series 1, 2, and 4 finales, but its overall high positive slope is maintained by the fantastic ratings of “Human Nature”/”The Family of Blood” and “Blink.” Without those stories, unrelated to the overall arc, series 3’s regression line’s slope would be closer to series 5’s slope.
What about the Moffat’s episodes? For his first two seasons Moffat set up story arches that were introduced in the first episode of each season (the crack in the wall and the impossible astronaut), teased us with the story in a couple of episodes during the season but otherwise presented adventure episodes, then tied up the story in the last couple of episodes. Looking at the graphs, we can see that the story arc episodes in general rate very high, but the rest of the episodes have poor ratings. This results in a flat regression line, as both series start high, dip low, and end high. We can see this happen again in series 7, in a bit more complicated way: the impossible girl is introduced in “Asylum of the Daleks,” restarted again in “The Snowmen,” and wrapped up in “The Name of the Doctor,” and apart from those three episodes and “The Angels Take Manhattan” (the departure of Amy and Rory, another story arc episode), the adventure episodes in series 7 are seem to be below average for the entire 7-series run of the show.
Another interesting thing to look at are the best episodes. Take a look at all the episodes that are rated 8.8 or above. They all have at least one of these three qualities: they are part of the season arc, they are written by a very highly-regarded writer (Moffat or Neil Gaiman), or they have something important to say (“The Impossible Planet”/”Satan Pit”, “Human Nature”/”Family of Blood”, “Midnight”, “Turn Left”, “Vincent and the Doctor”). Doctor Who is at its best when it tackles interesting themes.
A few more points:
- It looks like RTD was just hitting his stride when he left. Series averages were going up, more high-rated episodes, less low-rated episodes. I wonder what his series 5 would have been like?
- Series 7 was really kind of terrible. Out of fifteen episodes, only four of them rated higher than the average ratings of all of the other series.
- I’ve been referring mostly to the showrunner of the series, but I wonder how much the different Doctor affects these ratings? How many people who, say, don’t like the Tenth Doctor will automatically rate a Tenth Doctor episode low? Will a Tenth Doctor fan rate the episode higher? Does this cancel out?
One last thing I noticed on this graph was that it supports something I like to tell my friends when they start watching Doctor Who: some episodes are good and some episodes are bad, like any TV show, but Donna had the best season and never had a stinker episode. Take a look at her graph, series 4: not a single episode under 7.6, and four episodes at 9 or above! Her average is 8.29! No other series comes close to that record. Maybe her first episodes weren’t earth-shattering, but they were fun and enjoyable. I’d love to say it’s because she’s the best companion ever (because she is!), but honestly, I just simply think that series was just well done.