View
215
Download
0
Category
Preview:
Citation preview
PSY 1950Graph Design
December 8, 2008
QuickTime™ and a decompressor
are needed to see this picture.
Why graph?• Exploratory data analysis
– Usually raw data– Tukey: a good graph “forces us to notice
what we never expected to see”– e.g., scatter plot, funnel plot– Design less important
• Communication– Usually summary statistics– “a picture is worth a thousand words”– e.g., line graphs, bar graphs– Design extremely important
Bar Graphs vs. Line Graphs• Bar graphs
– Convey position (i.e., bar height)• Effective in conveying relative point values• Ineffective for conveying interaction effects
– X axis has nominal or ordinal scale
• Line graphs– Convey change in position (i.e., line
slope)• Ineffective in conveying relative point values• Effective for conveying interaction effects
– X axis has interval or ratio scale• Okay for 2-level nominal scale
Values vs. Trends
Relevance• Communication is most effective
when neither to much nor too little information is presented
• Tufte (1983): “Maximize the data-ink ratio, within reason”, where data-ink ratio is the “proportion of a graphics ink devoted to the non-redundant display of data.”
Distracting Backgrounds
4
4.5
5
5.5
6
Control Experimental
Count
Men Women
4
4.5
5
5.5
6
Control Experimental
Count
Men Women
Appropriate Knowledge• Communication requires prior
knowledge of relevant concepts, jargon, and symbols
4
4.5
5
5.5
6
Control Experimental
Score on Dependent Variable
Men Women
**
*
Jargon, Symbols
Concepts
4
4.5
5
5.5
6
Novel Suppress Concentrate
Stimulus Condition
Score on Dependent Variable
Meditators
Controls
-1
-0.5
0
0.5
1
Suppression Concentration
Priming Condition
Score on Dependent Variable
Meditators
Controls
Salience• Attention is drawn to large
perceptible differences
Discriminability• Two properties must differ by a large
enough proportion or they will not be distinguished
Size
0
1
2
3
A B
Group
Dependent Variable
0
1
2
3
A B
Group
Dependent Variable
Proportion
0
1
2
3
A B
Group
Dependent Variable
0
1
2
A B
Group
Dependent Variable
Scale
Perceptual Organization• People automatically group elements
into units, with they then attend to and remember
Compatibility• A message is easiest to understand if
its form is compatible with its meaning
4 4.5 5 5.5 6
4-5 Year Olds
7-8 Year Olds
9-10 Year Olds
Average Word Length
4
4.5
5
5.5
6
4-5 Year Olds 7-8 Year Olds 9-10 Year Olds
4
4.5
5
5.5
6
Control Experimental
Score on Dependent Variable
Women
Men
4
4.5
5
5.5
6
Experimental Control
Score on Dependent Variable
Women
Men
Color, Category Placement
Between vs. Within Subjects
4
4.5
5
5.5
6
Control Experimental
Score on Dependent Variable
Men Women
4
4.5
5
5.5
6
Control Experimental
Score on Dependent Variable
Informative Changes• People expect changes in properties
to carry information
3.5
4
4.5
5
5.5
6
6.5
Novel Suppress Concentrate
Stimulus Condition
Score on Dependent Variable3.5
4
4.5
5
5.5
6
6.5
Novel Suppress Concentrate
Stimulus Condition
Score on Dependent Variable
Capacity Limitations• People have a limited capacity to
retain and to process information and will not understand a message if too much information must be retained or processed
Placement is Key
4
4.5
5
5.5
6
Novel Suppress Concentrate
Stimulus Condition
Score on Dependent Variable
Controls Meditators
4
4.5
5
5.5
6
Novel Suppress Concentrate
Stimulus Condition
Score on Dependent Variable
Meditators
Controls
Horizontal vs. Vertical Bars: Labels
0 5 10 15 20 25 30 35
Amphetimines
Cocaine
Hashish or Marijuana
Heroin, opium, or morphine
LSD
Prescription drugs
Solvents
Percentage Use
0
5
10
15
20
25
30
35
Amphetimines
Cocaine
Hashish or Marijuana
Heroin, opium, or morphine
LSD
Prescription drugs
Solvents
Percentage Use
Font Typehttp://www.surl.org/usabilitynews/32/font.asp
QuickTime™ and a decompressor
are needed to see this picture.
Font Typehttp://www.surl.org/usabilitynews/32/font.asp
QuickTime™ and a decompressor
are needed to see this picture.
Font Style• ALL UPPERCASE TEXT IS HARDER
TO READ than all lowercase text or Mixed Case Text
• Italicized text is harder to read than plain text
• Underlined text (e.g., pajamas) is harder to read than non-underlined text (e.g., pajamas)
Scatterplots
ExampleZachs, J., & Tversky, B. (1999). Bars and lines: A study of
graphic communication. Memory & Cognition, 27, 1073-1079.
Recommended