Methods Week 11 - Visualizations (2/2)

how do u understand an error bar?

it is significant if it doesn’t pass 0. (same idea as CI)

what is data visualization important?

bypass assumptions that data sets look similar just based on similar data

distortion of scale

changing y axis to make the differences look very significant

• harder to identify in line graphs than bar graph bc we don’t know the range of the scale being used

distortion of meaning

changing the scale not to emphasize but to change the meaning to shown in the graph

ex. body temp vs. %above 98.6 body temp

assumed linearity

showing more data points show different variability rather than just a few data points that would show that something is insignificant

• helps with developing new studies to fix problems

multiple interpretations

• forgot 30% of information in BOTH groups

• ratio: 30%/80% they forgot less information than what they knew initially

• who remembered more at the end

what is wrong with this figure?

• 2 data points

• is the point is years or just 2 data points in general

• 2 groups don’t start off at similar base line so there could be confounds that statically are different and cannot be compared.