Reader Poll Roundup: Series Ten Edition

Every year when I examine the readers' reactions to the series's episodes, I have more data from previous years (S7, S8, S9) to compare and contrast, thus giving context to the current year's ratings. Because I eliminated the zero-star option in all of the Reader Polls this year, I won't be able to do an "apples to apples" analysis, but will do what I can with what I've got.

I'll begin, as usual, with the average (mean) ratings. For each episode, I multiplied any given star rating (e.g., 5 stars) by the number of votes it got, added the results, and divided that sum by the total number of votes. The relatively even-keeled results are below:

Although the separation between best and worst scores are not as extreme, this year there's a slightly wider spread in the middle half of the data than last year. The six middle-ranking episodes span a bit more than half a star (0.58), as opposed to S9's three-tenths of a star. S7's middle seven episodes were separated by 0.46 stars; S8's middle half by 0.44. For ease of visualization, here are the S10 ratings from best to worst (note that different colors have been assigned to the episodes between the above and below charts; I couldn't help that):

Unlike S9, in which the relative dud was a full 1.2 stars behind the next lowest-ranking episode, the bottom ranked episode of S10 (The Lie of the Land, or TLotL) lagged the penultimate one (The Pyramid at the End of the World or TPatEotW) by less than a half-star (0.42).

When we look at the entire series, the difference between highest-rated World Enough and Time (4.33 stars) and lowest-rated TLotL (2.87 stars), yields a total series range of only 1.46 stars—nearly a full star tighter than S9's 2.42-star spread. Given everything I've heard (and felt) about how good S10 has been overall—with some even claiming it's one of (if not the) strongest of the modern era, that felt appropriate. Then I compared to S7 and S8.

The overall span of scores for S8 episodes was 1.36 stars. S7 was even tighter, at a mere 1.08 stars. Perhaps that is a result of trying to do the statistics of small numbers (but more on number of votes below).

Looking at the entire series, we get an average score of 3.71 stars with standard deviation (σ) of 0.43 stars. S9 scored 3.58 ± 0.65, S8 was 3.53 ± 0.46, an S7 had 3.76 ± 0.34. Again, Series Seven seems to have been the big winner here (though only last year's Heaven Sent scored higher than World Enough and Time). Remember, though, that these S7 statistics include The Snowmen (none of the other series took Christmas specials into account) and the most-ever-voted-upon The Name of the Doctor.

As far as individual episodes go, though, Series Ten is still holding its own. Four episodes—a third of the season—were rated at 4.00 stars or above (World Enough and Time, The Doctor Falls, The Pilot, and Thin Ice); Series Nine, Eight, and Seven had three, three, and four, respectively (though note that S7 had two more total episodes, as well). On the low end, S10 had only two episodes at 3.33 stars or less (TPatEotW and TLotL), while its predecessors had (respectively) three, four, and two such episodes.

Most S10 episodes garnered the most votes at either four or five stars, with the exception of TLotL, which peaked at three stars, and TPatEotW, which received something of an inverse bell curve of votes:

Because zero-star votes were so prominent last year, I eliminated the option. Now I'm not sure whether or not that was smart, because I have no way to know whether or not S10 episodes were any more satisfying to those zero-star voters. However, I can look at percentages of one-star votes as a means of comparison.

The maximum percentage of 1-star votes any S10 episode got was 29% (TPatEotW), which was also the episode that garnered the least votes overall. On average, 1-star votes were only 8% of an episode's total. In other series, one-star vote maxes and averages were 26% / 5% (S9), 19% / 5% (S8), and 15% / 4% (S7, without Snowmen). If I combine zero-star votes with those 1-star ones, we instead get 42% / 14% (S9), 30% / 10% (S8), and 15% / 5% (S7). Thus, S10 has the fewest low-star ratings since S7 when both 1- and zero-star votes are taken into account.

More so than any previous year, though, the results are suspect by virtue of there being so little data on which to base the ratings. Despite re-opening the three-week-long polls again in the last week-and-a-bit, only the two-part finale garnered any new votes: World Enough and Time got a single extra vote, while The Doctor Falls finished out its three week period during the run up to this post. Check out the response rates below.

Poll participation continues to fall. I don't know if that's because I've lost readership with my move away from weekly posts, everyone's focus is elsewhere because of world events, or some other factor(s), but it makes pulling reliable statistics out of the data nigh impossible. This year's most-voted-upon episode (The Doctor Falls) got only four more total votes than the least-voted-upon episode from S7. However, it was still only three votes behind both S8's and S9's top vote-getter.

At the other end of the spectrum, though, we're down to a mere seven votes for TPatEotW. That's not even half the lowest performer from S8 and S9 (15 and 16 votes, respectively), and less than a third of S7's (24 votes). Last year, only Under the Lake failed to garner more than fifteen votes; this year only a third of the series exceeded that mark.

I'll keep posting polls as new episodes come out for as long as I continue the blog, but I may need to find new ways to market the fact that my polls are out there. When the majority of entries have fewer than fifteen votes each, it's hardly a representative sample. Perhaps that's why I come away feeling that a really stellar season performed only adequately according to the numbers.



Real Time Analytics