Tuesday, October 7, 2008

Debate Word Histograms

Here's the distribution of single words over the two debaters tonight, with a histogram of all speakers (including questioners) thrown in for comparison. (You can find the full transcript here.)

First, with only some basic function words filtered:

All:
we 359
i 267
you 254
sen 132
going 121
our 116
not 96
do 91
can 84
know 80
McCain:
i 14
know 8
my 7
you 6
country 4
we 4
times 4
Obama:
i 11
we 10
you 6
going 5
my 4
same 4
Now, with more robust filter of some other function words and support verbs:

All:
sen 132
know 80
mccain 77
now 59
obama 57
think 57
people 54
McCain:
know 8
country 4
times 4
like 4
believe 3
Obama:
same 4
challenges 3
things 3
know 3
Maybe I'll try a deeper linguistic analysis tomorrow night as I travel for work. But I'd be surprised if one didn't show up sooner on Language Log.

No comments: