Transcript Statistics 101 (An Introduction)

Introduction to Statistical Methods
Session 3
Univariate Data
Daniel D Reidpath
28 June 2011
Kinds of statistics

Location


Spread


Mean, 25th percentile, Mode
Range, variance, standard deviation
Shape

Uniform, “Normal”

Interrelationship/Difference

Fit
Kinds of Data

Ratio data


Interval data


2Y is twice as big as Y
The difference between Y+2 and Y+5 is 3
Ordinal data

Sometimes treated as truly ordered data

Sometimes treated as a “Factor”

Sometimes treated as interval or
even ratio (*sigh*)
Kinds of Data...



Nominal

Factors

Groups
Binary/Dichtomous

Logical (True/False)

Nominal
Missing
Nominal Data
• In general the numerical descriptions for
nominal data are
• Frequencies – counts
• Percentages or Proportions
• The common visual plot for nominal data is
the barchart (or in R .. The barplot() )
• Load Session3.R and have a look at the
Titanic Data
Ordinal Data
• Numerical Descriptions for ordinal data
include medians, range and quartile range
• The typical plot is the boxplot.
• Using Session3.R, have a look at the
famous Iris data from RA Fisher.
Photograph by Eric Guinther
Continuous Data
• Common summary descriptions for data
involve some kind of central point measure,
and some measure of variation.
• What would be good for continuous data?