# How should I best visualize my survey data?

I've conducted a Likert-style survey with $$45$$ questions in which $$15$$ people participated. Each question was repeated six times, so I have six answers from every participant for every question. All answers are $$\in \{1, 2, 3, 4, 5\}$$.

What I'm trying to visualize are the differences in consistency with which the participants answered the six repetitions of each question. For example, participant $$a$$ gave six answers to question $$q_1: [3, 5, 1, 1, 1, 2]$$ while participant $$b$$ gave the following answers to the same question $$q_1:[4, 4, 1, 1, 2, 2]$$. The order does not matter to me, what I'm interested in is the count ($$a$$ chose $$1$$ three times, $$2$$ one time, $$3$$ one time and $$5$$ one time while $$b$$ chose $$1$$ two times, $$2$$ two times and $$4$$ two times).

My data looks like this:

# participant 1
question, 1_count, 2_count, 3_count, 4_count, 5_count
q_1,1,1,1,0,3
q_2,0,6,0,0,0
q_3,3,0,0,0,3
q_4,1,1,2,2,0
q_5,1,1,0,0,4


# participant 2
question, 1_count, 2_count, 3_count, 4_count, 5_count
q_1,1,5,0,0,0
q_2,1,1,1,1,1
q_3,0,0,3,0,3
q_4,1,1,0,0,4
q_5,2,2,0,2,0


I wasn't able to find a suitable visualization technique. Do you have any suggestions?

• Do you care about differences between questions or people? Or an overall summary of consistency in answering?
– mkt
Aug 11, 2022 at 15:33
• I'm more interested in the differences between people. But an overall summary would also be nice. Aug 11, 2022 at 15:41
• As the order of a person's answers to any given question does not matter ($q_{1,a} : [3,5,1,1,1,2] \equiv q_{1,a} : [5,1,2,3,1,1]$), you seem to be interested in the distribution of responses. For that, I would make individual histograms (for each question answered by each participant), and then plot them on a $15 \times 45$ grid (using, facet_grid in R, like this). (You could also make ridge histograms, though they do not render nicely for this few observations and this many distributions.) Aug 11, 2022 at 18:04 