Wednesday, November 18, 2015

Semantic Analysis of the Third Democratic Debate



Semantic Analysis of the Third Democratic Debate
November 18, 2015

Quantified Perception (QPL) has analyzed the Democratic Debate held November 14 and offers the results of that analysis for your consideration.

Using the QPL Text Analysis Profiler (TAP), we find that all three candidates use a similar tone in the debate, the Myers-Briggs ESTP “Inventor” type, which portrays an enthusiastic and innovative language, demonstrating independence while maintaining a “team player” attitude. Sanders uses more spontaneous and less structured language, while Clinton emphasizes leadership with a more theoretical vocabulary. Clinton also uses a more personal vocabulary than her opponents. Sanders offers a slightly more emotional language. O’Malley falls between Clinton and Sanders in each of these areas.

This debate offers substantive discourse on issues over personalities, with the flagship themes of each candidate rising from their repetition and reuses of certain words and phrases. QPL weights the frequency of usage based on the total number of words in the sample and the length of a phrase.

Clinton emphasizes the international climate, with secondary domestic focus on healthcare and the economy. The phrase “ISIS is the leading threat” is used three times, which is quite frequent for such a long string of words (five), with a relative weight of just over nine. This is roughly equivalent to using that phrase as nearly one percent of her overall five-word phrases. “Affordable Care Act” is used five times or a relative weight of six.
O’Malley focuses on domestic economic issues, referring to the “big banks of Wall Street” twice and “Wall Street” eight times (weight = 6.86). “Debt free college” is used six times for the same weighted value.
Sanders is even more focused on economic issues, referring to “Wall Street” 17 times and the “corrupt campaign finance system” four times (weight=9.49). Sanders also emphasizes healthcare in general and family leave specifically as a focal area.

Clinton is tonally inclusive and inclusive and prescriptive, using “We are,” “we have to,” “we should,” and “we need to” as dominate constructions. O’Malley is less reliant on such phrasing, although he repeats “we need to” eight times. Sanders uses construction similar to Clinton’s with “we have to,” “we need to,” and “we should” as frequent phrasing. He also uses the common “I would” about as often as Clinton and O’Malley.
Overall, the candidates used functional language, with economics and international issues raised in their vocabulary. A baker’s dozen of substantive words is given in the chart below, all with a relative weighted index value above 0.5.

Next: A parallel analysis of the last Republican debate.

PhraseFrequency of usage Relative Weight Index
world 32 2.52
Wall [Street] 32 2.52
college 23 1.81
ISIS 17 1.34
work 14 1.10
income 13 1.02
working 13 1.02
money 11 0.87
young 9 0.71
gun 8 0.63
economy 7 0.55
international 7 0.55
Iraq 7 0.55

Copyright 2015 Quantified Perception, LLC

Tuesday, October 27, 2015

The Presidential Debates, unplugged



The Presidential Debates
Analysis by Quantified Perception
Have you watched the presidential debates so far? You may wonder how the democrats differ from the republicans, in terms of the language used and the concepts highlighted. Quantified Perception has analyzed the transcripts of all three debates to find the crucial areas where the two parties differ, and where they coincide. To our surprise, they really don’t differ much. The words, concepts, and phrases are so close to the same that the only meaningful conclusion is that everyone is talking the same talk.
We used the transcripts from the debates to evaluate the respective language, adjusting each for the number of words reported for each debate.
Of course the language of the campaign dominates, with the words “president” (or “commander in chief”) and “campaign” dominating the remarks. But other (slightly) less obvious words and concepts dominate the language of the candidates for both parties. In fact, the most common words in all three debates are the same, and rank the same:
“people”, “country”, “issues,” “things” and “going”.

The table below shows that the top ten words used in the second republican debate include five of the top ten words from the other two.
Word
Rank in second Republican Debate
Rank in first Republican Debate
Rank in Democratic Debate
people
1
3
1
about
2
2
2
but
3
5
3
do
4
1
4
will
5


they
6


he
7


think
8

9
I’m
9


would
10


if
11
7

want
12

10






 










The most common concepts show a similar pattern. They tend to be the same concepts in each debate, and in roughly the same order:
Business - battle - family - security

The most common concept in all three debates is "business," including the words economy, finance, industry, and wealth. Interestingly, the democrats said “wealth” often, while republicans said “deal”.

“War” is the next common concept, dominated by the word “war” and including “combat,” “fight,” and “battle.”

“Family” was the third most referenced concept, with “children,” “babies,” and “daughters” said frequently. “Daughters” was used much more often than “sons,” indicating some sort of gender bias in both camps.

“Security” was fourth most used concept, including the words “care” and “safety.”

Truth was another concept, indicated by “fact”, “actually,” “truth,” “reality,” and “real.”
“Way” is another frequent word, but that is partially due to Trumps penchant for asides, as in “by the way,” and the dominance he had in speaking in both republican debates.

The phrasing in all three debates was also surprisingly consistent, with the top 15 or so phrases used often in all three. The tables below show that the top 15 phrases from the democratic debate match five of the top phrases from the first republican debate and ten from the second.
Phrases
Rank in Democratic Debate
Rank in first Republican Debate
Rank in second Republican Debate
we have
1


thank you
2


that we
3
4
3
we need
4
1
1
this country
5
5
5
we’re going
6


it was
7

11
we are
8

19
you have
9

5
wall street
10


climate change
11


you know
12
2
2
united states
13

4
if you
14
9
17
but I
15

12


Quantified Perception used its Text Analysis Profiler for this analysis. Among other things, the Profiler determines a “Myers-Briggs”–like profile for a document. All three transcripts profiled as “ENTP”, with remarkable similar ratios of the profile vectors:
1st Republican Debate   2nd Republican Debate       Democratic Debate


The Wall Street Journal years ago referred to “Remocrats” and “Depublicans.” I wonder if they will revive that language, if this continues…