Transcript Inferential statistics

Inferential Statistics
Inferential statistics
Use the sample to make inferences about
the population
 Use at least two variables at a time
 Can never be sure that results in sample
are 100% representative of population
 Instead, statistics tell us the probability
that results in sample are true in
population

Significance Levels



p<.10
 This means that the probability that the finding was by chance is
less than 1 in 10.
 This indicates that there was a 'trend' in the data. It is not
considered a significant finding, but it is encouraging enough to
continue with the line of investigation.
p<.05
 This means that the probability that the finding was by chance is
less than 5 in 100.
 This indicates that there was a significant finding in the data.
You should certainly continue with this line of investigation.
p<.01 or less
 This means that the probability that the finding was by chance is
less than 1 in 100.
 This means that there was a highly significant finding in the data.
You should continue with this research with considerable
confidence.
Inferential Statistical Tests: Which
Test?
Variable 1
Variable 2
Test to use
Statistic
Continuous
Continuous
Correlation*
r
Categorical
1. 2 categories
2. IV
Continuous
1. DV
T-test
t
Categorical
1. 3 or more
categories
2. IV
Continuous
1. DV
One-way
ANOVA
F
Categorical
Categorical
Chi-Square*
2
* For Correlation and Chi-Square analyses, variables are not distinguished as
IVs or DVs.

Does GPA tend to
increase in people
with high selfesteem?
Correlation coefficient
indicates the degree
of association
between the two
variables.
GPA

High
Does GPA Correlate With SelfEsteem?
Self-Esteem
Two Features of a Correlation
Perfect
Moderate
Direction
+
Weak
-
Strength
Chi-Square
Expected Results for Random Relationship between Gender and Ability to
Drive
Can drive
Can’t drive
Females
10
10
Males
10
10
What if this is what we really found:
Can drive
Can’t drive
Females
Males
16
4
4
16