Transcript T test
Hypothesis Testing
Using T test Statistic
Hypothesis Statement
H1: College students sleep less than most folks
H0: College students sleep the same or more than
most folks
The statistic you need to decide whether to reject or
fail to reject the Null Hypothesis is the one-sample
t test. This statistic compares the sample mean to the
population mean and gets an estimate of the
probability that the sample mean is different by
chance.
Example
The average sleep time for the population is found to
be 8 hours a day
We think college students sleep a different amount,
maybe more - maybe less.
We survey 10 students to see how much they sleep.
Let us assume that the average sleep time is less.
The question is:
Did you happen to pick a group of light sleepers by
chance?
Or do college kids really sleep less?
Findings
Population mean = 8 hours
Data collected from 10 students
Hours sleep per day : 3,4,5,5,5,5,6,6,6,7
Mean = 5.2 hours
Std Dev = 1.1353
Std Error of the mean = 0.3590
Df = 9
Computer Generated Statistics
Here's the output from a computer package:
SE = SD / √N
Reject / Fail to Reject?
Does Your Group Come From a Different Population
Than the One Specified?
Population mean = 8
Sample mean = 5.2
2 Possibilities:
Your sample comes from a different population - reject
the Null Hypothesis.
The group comes from the same population and the mean
varies by chance. You just happened to pick a sample
group with a mean that misrepresents the population it
came from. The group isn't really different. When this
happens, you fail to reject the Null Hypothesis.
Do we reject the Null Hypothesis?
1) Value of the sample mean = 5.2
2) “Test value” = 8; Population mean
3) “Mean difference” = –2.8;
H0: College students sleep the same or more than most
folks
Decision based on t score and p value
T score is large = 7.799
p value (probability of this being by chance) is very low
< 0.001
Answer: YES / NO?
Fail to reject the Null Hypothesis?
1) Value of the sample mean = 7.5
2) “Test value” = 8; Population mean
3) “Mean difference” = –0.500
H0: College students sleep the same or more than most
folks
Decision based on t score and p value
t score is small = 2.236
p value (probability of this being by chance) is very
high 0.052
Answer: YES / NO?
We Conclude That…
Big t value and small p value
Reject Null Hypothesis
Significant difference
Sample comes from a different population
Group is different from Null Hypothesis mean
Small t value and big p value
Fail to reject Null Hypothesis
Data set is not significant
Sample comes from the same population & mean varies by
chance
Group is not different from Null Hypothesis
Errors in Hypothesis Testing