Overview of seminar
- What makes a good graph?
- Main principles
- Different types of graphs
- Histograms, box plots, line charts, scatterplots
- How to plot them in SPSS
What makes a good graph
- Show the data
- Help the reader to focus and think about the data presented (avoid adding
distracting components to your graph) - Present any numbers with minimum ink
- Make large data set coherent
- Reveal the underlying message of the data
Tufte (2001)
How to know when you use the following graphs for data:
-histograms
-box plots
-scattergraphs
-line graphs
Histogram
Purpose: Show distribution
Data focus: Single quantitative variable
More for experiments
Box plot
Purpose: Summarize distribution and variability
Data focus: Single quantitative variable (groups optional)
More for experiments
Line Graph
Purpose: Show trends over time
Data focus: Quantitative variable over time
To examine a relationship
Scatterplot
Purpose: Examine relationships
Data focus: Two quantitative variables
To examine a relationship
How to know when to use a histogram over box plot
Historgram if you want to see distribution - this is clearly visible
Box plot if you want to summarise mean, median aetc and see standard deviation
How to know when to use line graph or scatterplot
Scatterplot if looking for a relationship. Easier to see anomalies.
Line graph more for looking at trends over time. Typically has a time or ordered variable.
