Tables, Charts, Graphs, and Diagrams The Power of Graphs and Tables

1
Tables, Charts, Graphs, and Diagrams The Power
of Graphs and Tables

• Session 10

2
TablesBeginning to Establish Causality

• Tables
• Editorial material
• Table number
• Title
• Headnote
• Footnote
• The Box
• Column Captions
• Row Captions
• Totals

3
Presenting Table Data Effectively

• Round to two significant digits
• Use row and column averages
• Figures are easier to compare in columns than in
  rows
• Arrange to show the structure of the data
• Space it to make it easy to read
• Use graphs only when appropriate

4
Tables Designed to Examine Relationships

• Bivariate Tables
• Multivariate Tables
• nested tables
• Title
• Independent variable
• Labels
• Order of values
• Percentages
• Totals
• Footnotes

5
GraphsUses and Abuses (the Mark Twain effect)

• Basics of Graphs
• Two coordinates
• Horizontal axis X independent variable
• Vertical axis Y dependent variable
• and data points
• Positive Linear
• Negative Linear
• Curvilinear
• S-shaped
• Exponential

6
Graphs

• Positive Linear

• Negative Linear
• Job losses have been especially severe in the
  state's manufacturing sector. Over 40 of the
  total recession-based job losses came from this
  one employment sector. While total seasonally
  adjusted nonfarm employment in the state is up
  from its fourth-quarter 1992 low, manufacturing
  employment has continued to slide. Manufacturing
  job losses are expected to continue until the
  year 2000. It is significant to note that in 1950
  over 50 of the state's labor force worked in
  manufacturing. Today the manufacturing sector
  employs about 19 of the workforce.

7

• Curvilinear
• S-shaped
• Exponential
• Few processes continue to follow this form of
  growth for ever. The streets of London are, in
  the hackneyed example, not in fact shoulder-deep
  in horse manure, as they would have been if an
  exponential increase in horse-drawn taxi traffic
  had continued indefinitely... Many processes of
  change instead follow an S-shaped or "sigmoid"
  curve ("sigma" is Greek for "S")
• Note that the early parts of the curves - the
  first 40 years in these arbitrary examples - are
  very similar. It is thus extremely difficult to
  tell when a process is makingthe transition. It
  is conversely very easy to "cry wolf" in either
  sense "population is increasing exponentially!"
  or "food production is trailing off!" Dr Evans
  calls this latter "sigmoid fraud".
• Further breeding can produce even more extreme
  dwarf varieties. A number of ways in which plants
  might use sunlight more efficiently are worth
  exploring, though the fact that we start from the
  results of more than 5000 years of selective
  breeding makes further staggering increases
  unlikely. Possibly the biggest opportunity for
  increasing grain yields is in producing varieties
  which are more precisely adapted to local
  conditions.

8

• Pie Chart shows the proportion of the whole for
  each category
• Bar Chart shows the number within each category
• Frequency Lines or Distribution Curves depicts
  the general shape of the distribution

9
Choosing Among Graphics

• Choose the representation that most accurately
  represents the data
• Avoid deliberate distortions

10
Checklist for Graphics

• Simplicity
• Choose according to the nature of the variable
• Accuracy
• Titles and source material

11
Presenting Qualitative Data

• Descriptive versus explanatory intent
• Single site versus multiple site data
• Ordered versus non-ordered
• Time Ordered versus not time ordered
• Categories of variables
• Kinds of information