Lecture Notes
Chapters 1-2
Part II
Turning Data Into Information
Armstrong
Statistics
1
EXPLORATORY
DATA
ANALYSIS
(EDA)
Before you start
doing a bunch of
calculations it is wise
to explore your data
first.
Categorical Quantitative
Variables Variables
One Variable Two Variables One Variable Two Variables
Bar Chart Contingency Boxplot Scatterplot
Table
Pie Chart Stemplot
Frequency Histogram
Table
EDA helps us describe the DISTRIBUTION of a variable:
A DISTRIBUTION TELLS US WHAT VALUES A VARIABLE CAN
TAKE AND HOW OFTEN IT TAKES THESE VALUES.
Armstrong
Statistics
2
DISPLAYING
THE
DISTRIBUTION
OF A
CATEGORICAL
VARIABLE
TABLES:
BAR CHART PIE CHART FREQUENCY
TABLES
(one variable)
CONTINGENCY
TABLES
(two variables)
FREQUENCY
CATEGORY NAMES
or
X AXIS: Y AXIS: CIRCLE: SECTIONS:
RELATIVE
FREQUENCY
Category Frequency or Represents Represent
names/labels relative the entire the category
frequency data set percentages
Prefer RF!
Ex:
Armstrong
Statistics
3
An article in the Winter 2003 issue of Chance magazine reported on the Houston Independent
School District’s magnet schools programs. Of the 1755 qualified applicants, 931 were accepted,
298 were waitlisted, and 526 were turned away for lack of space.
1. What is the variable of interest?
2. Find the relative frequency distribution of the decisions made
3. Make an appropriate display of these data
4. Interpret your graph.
Armstrong
Statistics
4
a CONTINGENCY TABLE shows a table of the JOINT DISTRIBUTION of two variables
The rows contain the groups of one variable.
The columns contain the categories of the other variable.
O is the observed frequency in each of the cells.
T is the Total
Category 1 Category 2 Category 3 Category 4
Group 1 O11 O12 O13 O14 TG1
Group 2 O21 O22 O23 O24 TG2
TC1 TC2 TC3 TC4 n
JOINT FREQUENCY
represent the number of units that are in a particular group AND category.
JOINT PROPORTION
represent the proportion of units that are in a particular group AND category.
divide the observed frequency by your sample size
MARGINAL DISTRIBUTION
Just look at the margins of the contingency table
the distribution of the group variable while ignoring the category variable
or
the distribution of the category variable while ignoring the group variable.
The marginal distribution for the category variable is:
Category Category 1 Category 2 Category 3 Category 4
Frequency TC1 TC2 TC3 TC4
The marginal distribution for the group variable is:
Group Group 1 Group 2
Frequency TG1 TG2.
Armstrong
Statistics
5
EX: Below is a contingency table of frequencies (counts)
Freshman Sophmore Junior Senior
Male 20 25 39 12 96
Female 20 42 23 14 99
40 67 62 26 195
What is the joint frequency of being a Sophmore and Female? _______________
What is the joint frequency of being Male and a Junior? _______________
What is the joint frequency of being a Freshman and Female? _______________
What is the joint frequency of being Female and a Freshman? _______________
What is the probability of being a Sophmore and Female? _______________
What is the probability of being a Sophmore and Male? _______________
What is the probability of being Male and a Junior? _______________
What is the probability of being Female and a Junior? _______________
What is the probability of being a Freshman and Female? _______________
What is the probability of being Female and a Freshman? _______________
The marginal distribution for class standing is:
Class Standing Freshman Sophmore Junior Senior
Rel Freq
What is the probability of being a Sophmore? ____________
What is the probability of being a Freshman? ____________
The marginal distribution for gender:
Gender Male Female
Rel Freq
What is the probability of being Male? ____________
Armstrong
Statistics
6