1 Hypothesis Testing Rev

HYPOTHESIS
TESTING
What is Hypothesis
A Hypothesis is the statement or an
assumption about relationships
between variables which can be
proven false.
or
A Hypothesis is a tentative
explanation for certain behaviors,
phenomenon or events that have
occurred or will occur.
HYPOTHESIS TESTS
Null hypothesis HO:
There is no difference
Alternative hypothesis (HA):

There is a difference
If the null hypothesis is false, it is likely
that our alternative hypothesis is true.
TYPES OF HYPOTHESIS
Each of the following statements is
an example of a null hypothesis
and alternative hypothesis.
A difference is considered
significant if the probability of
getting that difference by random
chance is very small.
P value:
The probability of making an error by
chance
Historically we use p < 0.05
Criteria for Hypothesis Construction

It should be empirically testable, whether it
is right or wrong.
It should be specific and precise.
The statements in the hypothesis should
not be contradictory.
It should specify variables between which
the relationship is to be established.
It should describe one issue only.
Steps in Hypothesis Testing
Establish Critical or Rejection region
FORMULATE A DECISION RULE

TO ACCEPT NULL HYPOTHESIS
Accept H0 if the test statistic
value falls within the area of
acceptance.
Reject otherwise.
Select the Suitable Test of

Significance or Test Statistic
Whether the test involves one
sample, two samples, or many
samples?
Whether two or more samples
used are independent or related?
Is the measurement scale nominal,
ordinal, interval, or ratio?
ERRORS IN HYPOTHESIS
TESTING
The choice of a probability distribution of

a sample statistics is guided but the
sample size n and the value of
population standard deviation .
Student
t - Test
Students t-test
Student was W. S.
Gossett. He published his
test anonymously as
Student because he was
working for the brewers
Guinness and had to keep
the fact they were using
statistics a secret.
The test is used to compare
samples from two different
batches.
It is usually used with small
(<30) samples that are
normally distributed.
TWO TYPES OF T-TEST

Matched Pairs and Independent Pairs
If there is some link between the
data then use the matched pairs
test (eg a before and after).
If there is no link between the data
then use the independent pairs
test.
T-TEST INDEPENDENT SAMPLES

If there is no before and after relationship
between the samples then the independent
samples test is used.
T - TEST FOR TWO

INDEPENDENT
SAMPLES

The reason for hypothesis
testing is to gain
knowledge about an
unknown population.
Independent samples t-test
is applied when we have
two independent samples
and want to make a
comparison between two
groups of individuals. The
parameters are unknown.
How is this different than a
Z-test and One Sample ttest?

We are interested in the difference
between two independent groups. As
such, we are comparing two populations
by evaluating the mean difference.
In order to evaluate the mean difference
between two populations, we sample
from each population and compare the
sample means on a given variable.
Must have two independent groups
(i.e.samples) and one dependent
variable that is continuous to compare
them on.
Examples:
Do males and females significantly
differ on their level of math anxiety?
IV: Gender (2 groups: males and females)
DV: Level of math anxiety
Do older people exercise significantly
less frequently than younger people?
IV: Age (2 groups: older people and
younger people)
DV: Frequency of getting exercise
Examples:
Do 6th graders have significantly more
unexcused absences than 4th graders
in LNC?
IV: Grade (2 groups: 6th grade and 4th
grade)
DV: Unexcused absences
Note that Independent t-test can be
applied to answer each research
question when the independent
variable is dichotomous with only two
groups and the dependent variable is
continuous.
Assumptions
The two groups are independent of
one another.
The dependent variable is normally
distributed.
The two groups have approximately
equal variance on the dependent
variable.
Step 1: State the hypotheses

Ho: The null hypothesis states that the two samples come
from the same population. In other words, There is no
statistically significant difference between the two
groups on the dependent variable.
Symbols:
Non-directional:
Directional:
Ho: 1 = 2
H 0:
1
or
H 0:
1
If the null hypothesis is tenable, the two group means

differ only by sampling fluctuation how much the
statistics value varies from sample to sample or
chance.
Step 1: State the hypotheses

H1: The alternative hypothesis states that the two samples
come from different populations. In other words, There
is a statistically significant difference between the two
groups on the dependent variable.
Symbols:
Non-directional:
Directional:
H1 : 1 2
H 1:
1
or
H 1:
1
If the null hypothesis is tenable, the two group means

differ only by sampling fluctuation how much the
statistics value varies from sample to sample or
chance.
Step 2: Set a Criterion for Rejecting Ho
Compute degrees of freedom

Set alpha level
Identify critical value(s)
T Distribution Table
Computing Degrees of Freedom

Calculate degrees of freedom (df) to
determine rejection region.
df = n1 + n2 - 2
sample size for sample1+ sample size for
sample2 2
df describe the number of scores in a
sample that are free to vary.
We subtract 2 because in this case we
have 2 samples.
Identify Critical Value(s)

Directional or non-directional?
Look at the T Distribution Table
To determine your Critical Value
you need to know:
df if df are not in the table, use the

next lowest number to be
conservative
directionality of the test
alpha level
Step 3: Collect data and Calculate t

statistic
x x
1
1
s n
s n 1
2
n n
n n
variance
Whereby:
n : Sample size
x : Sample mean
s2 = variance
df
subscript1 = sample 1 or group 1
subscript2 = sample 2 or group 2
Step 4: Compare test statistic to criterion
df = 18 = .05 , two-tailed test in this example

critical values are 2.101 in this example
30
Step 5: Make Decision
Fail to reject the null hypothesis and conclude that there is no statistically
significant difference between the two groups on the dependent variable,
t = , p > .
OR
Reject the null hypothesis and conclude that there is a statistically
significant difference between the two groups on the dependent variable,
t = , p < .
If directional, indicate which group is higher or lower (greater,
or less
31
than, etc.).

Example:
Some brown dog hairs were found on the clothing
of a victim at a crime scene involving a dog.
The five of the hairs were measured: 46, 57, 54,
51, 38 m.
A suspect is the owner of a dog with similar
brown hairs. A sample of the hairs has been taken
and their widths measured: 31, 35, 50, 35, 36 m.
Is it possible that the hairs found on the victim
were left by the suspects dog? Test at the %5
level.
[From D. Lucy Introduction to Statistics for Forensic Scientists Chichester: Wiley,
2005 p. 44.]

1. Calculate the mean and standard deviation for the
data sets.

data sets.
Total
Mean
Standard deviation
Dog A
Dog B
46
31
57
35
54
50
51
35
38
36
246
187

data sets.
Dog A
Dog B
46
31
57
35
54
50
51
35
38
36
Total
246
187
Mean
49.2
37.4
Standard deviation

data sets.
Dog A
Dog B
46
31
57
35
54
50
51
35
38
36
Total
246
187
Mean
49.2
37.4
7.463
7.301
Standard deviation

data sets.
.
2. Calculate the magnitude of the difference between
the two means
49.2 37.4 = 11.8
Total
Mean
Standard
deviation
Dog A
Dog B
46
31
57
35
54
50
51
35
38
36
246
49.2
7.463
187
37.4
7.301

data sets.
.
the two means
49.2 37.4 = 11.8
3. Calculate the standard error in the difference
= 4.669 4.67
Total
Mean
Standard
deviation
Dog A
Dog B
46
31
57
35
54
50
51
35
38
36
246
49.2
7.463
187
37.4
7.301

data sets.
.
the two means
49.2 37.4 = 11.8
= 4.669 4.67
4. Calculate the value of t = 2.527
11.8
= 2.527
t=
4.67
Total
Mean
Standard
deviation
Dog A
Dog B
46
31
57
35
54
50
51
35
38
36
246
49.2
7.463
187
37.4
7.301

data sets.
.
the two means
49.2 37.4 = 11.8
= 4.669 4.67
5. Calculate the degrees of freedom
n
=5+5 -2=8
Dog A
Dog B
46
31
57
35
54
50
51
35
38
36
Total
246
187
Mean
49.2
37.4
Standard
7.463
7.301
df = n1 + n2 - 2
deviation

data sets.
.
the two means
49.2 37.4 = 11.8
= 4.669 4.67
=5+5 -2=8
6. Find the critical value for the particular significance

you are working to from the T Distribution table.
At the 0.05 level
tcrit = 2.306

data sets.
.
the two means
49.2 37.4 = 11.8
= 4.669 4.67
=5+5 -2=8
6. Find the critical value for the particular significance

you are working to from the T Distribution table.
If t < critical value then there is no significant difference
between the two sets of data.
If t > critical value then there is a significant difference
between the two sets of data
At the 0.05 level
tcrit = 2.306
So, at 0.05 level there is a significant

difference between the two data sets.
So it could not come from the same dog.

Group A
Group B
(leading
(control)
questions)
.
2
3
4
4
2
4
3
MeansA = ?
SDA = ?
3
5
0
3
4
0
MeansB = ?
SDB = ?
Diff of means= ?
SE of diff = ?
DoF = ?
t=?

_
XA - xA
Group A
Group B
.
xB - xB
3
_
xA=
Diff of means =
df = n A + n B
xB=
SDA=
xA - xB
2=
SDB=
SE of diff = SDA - SD B =
t=?

Group A
Group B
(leading
(control)
questions)
.
2
3
4
4
2
4
3
Mean = 3.143
SD= 0.8997
3
5
0
3
4
0
Mean = 2.50
SD = 2.07
Diff of means= 0.64

SE of diff = 0.861
DoF = 11
t = 0.7433

.
t < tcritical at 5% level, 2-tail
df = 11
t = 0.7433
There is no significant difference between the data test
Single-Sample
Z Test
A one-sample Z-test for proportions is a test

that helps us compare a population proportion
with a sample proportion.

30%
Sample Mean
()
30%
Population Mean

30%
Sample Mean
()
This is the
symbol
for a
sample
mean
30%
Population Mean
A one-sample Z-test for proportions is a test that

helps us compare a population proportion with a
sample proportion.
30%
Sample Mean
()
30%
Population Mean
And this is the

symbol for a
population mean
(called a mew)

30%
Sample Mean
()
Here is our question:
30%
Population Mean

30%
Sample Mean
()
30%
Population Mean
Here is our question: Are the population and

the sample proportions (which supposedly
have the same general characteristics as the
population) statistically significantly the same
or different?
Consider the following example:

A survey claims that 9 out of 10 doctors
recommend aspirin for their patients with
headaches. To test this claim, a random sample
of 100 doctors is obtained. Of these 100
doctors, 82 indicate that they recommend
aspirin. Is this claim accurate? Use alpha = 0.05

Which is the sample?

What is the population?

The sample proportion is .82 doctors
recommending aspirin.

The sample proportion is .82 of doctors
recommending aspirin.
The population proportion .90 (the claim that 9
out of 10 doctors recommend aspirin).
We begin by stating the null hypothesis:

The proportion of a sample of 100 medical doctors
who recommend aspirin for their patients with
headaches IS NOT statistically significantly different
from the claim that 9 out of 10 doctors recommend
aspirin for their patients with headaches.

The alternative hypothesis would be:

The alternative hypothesis would be:

headaches IS statistically significantly different from
the claim that 9 out of 10 doctors recommend aspirin
for their patients with headaches.
State the decision rule:
State the decision rule: We will calculate what

is called the z statistic which will make it
possible to determine the likelihood that the
sample proportion (.82) is a rare or common
occurrence with reference to the population
proportion (.90).
State the decision rule: We will calculate what

is called the z statistic which will make it
possible to determine the likelihood that the
sample proportion (.82) is a rare or common
occurrence with reference to the population
proportion (.90).
If the z-statistic falls outside of the 95%
common occurrences and into the 5% rare
occurrences then we will conclude that it is a
rare event and that the sample is different from
the population and therefore reject the null
hypothesis.
Before we calculate this z-statistic, we must

locate the z critical values.
Before we calculate this z-statistic, we must

locate the z critical values.
What are the z critical values? These are the
values that demarcate what is the rare and the
common occurrence.
Lets look at the normal distribution:
It has some important properties that make it

possible for us to locate the z statistic and
compare it to the z critical.
Here is the mean and the median of a normal

distribution.
50% of the values are above and below the

orange line.
50% -
50% +
68% of the values fall between +1 and -1

standard deviations from the mean.
34% -
34% +

34% -
-1
34% +
mean
68%
+1

34% -
-2 -1
34% +
mean
95%
+1 +2
Since our decision rule is .05 alpha, this

means that if the z value falls outside of the
95% common occurrences we will consider it a
rare occurrence.
34% -
-2 -1
34% +
mean
95%
+1 +2
Since are decision rule is .05 alpha we will see

if the z statistic is rare using this visual
rare
rare
-2 -1
mean
+1 +2
2.5%
95%
2.5%
Or common
Common
-2 -1
mean
95%
+1 +2
Before we can calculate the z statistic to see

if it is rare or common we first must determine
the z critical values that are associated with
-2 and +2.
Common
-2 -1
mean
95%
+1 +2
We look these up in the Z table and find that

they are -1.96 and +1.96
Common
Z
values
-2 -1
mean
-1.96
+1 +2
+1.96
95%
So if the z statistic we calculate is less than

-1.96 (e.g., -1.99) or greater than +1.96 (e.g.,
+2.30) then we will consider this to be a rare
event and reject the null hypothesis and state
that there is a statistically significant
difference between .9 (population) and .82 (the
sample).

sample).
Lets calculate the z statistic and see where if
falls!

sample).
Lets calculate the z statistic and see where if
falls!
We do this by using the following equation:
So if the z statistic we calculate is less than -1.96

(e.g., -1.99) or greater than +1.96 (e.g., +2.30) then we
will consider this to be a rare event and reject the null
hypothesis and state that there is a statistically
significant difference between .9 (population) and .82
(the sample).
Lets calculate the z statistic and see where if falls!
So if the z statistic we calculate is less than -1.96 (e.g.,

-1.99) or greater than +1.96 (e.g., +2.30) then we will
consider this to be a rare event and reject the null
hypothesis and state that there is a statistically significant
difference between .9 (population) and .82 (the sample).
Lets calculate the z statistic and see where if falls!
Zstatistic is what we are trying to find to see if it is outside or

inside the z critical values (-1.96 and +1.96).
Heres the problem again:
A survey claims that 9 out of 10 doctors recommend

aspirin for their patients with headaches. To test this
claim, a random sample of 100 doctors is obtained. Of
these 100 doctors, 82 indicate that they recommend

Note this
little hat ()
over the p
means that
this proportion
is an estimate
of a population
(.90)
(100)

(100)

(100)
Lets plug in the numbers
Sample Proportion

Sample Proportion

Sample Proportion
Population Proportion


The difference
The difference
Now for the denominator which is the estimated

standard error. This value will help us know how
many
standard error units .82 and .90 are apart from one
another (we already know they are .08 raw units
apart)
Now for the denominator which is the estimated

standard error. This value will help us know how many
standard error units .82 and .90 are apart from one
another (we already know they are .08 raw units apart)
Note - If the standard error is small then the z

statistic will be larger. The larger the z
statistics the more likely that it will exceed the
-1.96 or +1.96 boundaries, compelling us to
reject the null hypothesis. If it is smaller than
we will not.
Lets continue our calculations and find out:

aspirin for their patients with headaches. To test this claim,
a random sample of 100 doctors is obtained. Of these 100
doctors, 82 indicate that they recommend aspirin. Is this
claim accurate? Use alpha = 0.05
Sample Size:

aspirin for their patients with headaches. To test this claim,
a random sample of 100 doctors is obtained. Of these 100
doctors, 82 indicate that they recommend aspirin. Is this
claim accurate? Use alpha = 0.05
Sample Size:
Lets continue our calculations:
Now we have our z statistic.
Lets go back to our distribution:
Common
rare
-2 -1
mean
-1.96
rare
+1 +2
+1.96
95%
Lets go back to our distribution: So, is this result

rare or common?
Common
rare
-2.67
-2 -1
-1.96
mean
95%
rare
+1 +2
+1.96

rare or common?
Common
rare
-2.67
-2 -1
-1.96
This is the
Z-Statistic
we calculated
mean
rare
+1 +2
+1.96
95%

rare or common?
Common
rare
-2.67
-2 -1
-1.96
This is the
Z Critical
mean
rare
+1 +2
+1.96
95%
Looks like it is a rare event therefore we will

reject the null hypothesis in favor of the
alternative hypothesis:
Looks like it is a rare event therefore we will

reject the null hypothesis in favor of the
alternative hypothesis:
The proportion of a sample of 100 medical
doctors who recommend aspirin for their
patients with headaches IS statistically
significantly different from the claim that 9 out
of 10 doctors recommend aspirin for their
patients with headaches.
Summary of Certain Values for Sample Statistics Z

1 Hypothesis Testing Rev

Uploaded by

Copyright:

Available Formats

1 Hypothesis Testing Rev

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 Hypothesis Testing Rev

Uploaded by

Copyright:

Available Formats

HYPOTHESIS

Alternative hypothesis (HA):

Criteria for Hypothesis Construction

Steps in Hypothesis Testing

Establish Critical or Rejection region

FORMULATE A DECISION RULE

Select the Suitable Test of

The choice of a probability distribution of

TWO TYPES OF T-TEST

T-TEST INDEPENDENT SAMPLES

T - TEST FOR TWO

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

Step 1: State the hypotheses

If the null hypothesis is tenable, the two group means

Step 1: State the hypotheses

If the null hypothesis is tenable, the two group means

Step 2: Set a Criterion for Rejecting Ho

Compute degrees of freedom

Computing Degrees of Freedom

Identify Critical Value(s)

df if df are not in the table, use the

Step 3: Collect data and Calculate t

subscript1 = sample 1 or group 1

subscript2 = sample 2 or group 2

Step 4: Compare test statistic to criterion

df = 18 = .05 , two-tailed test in this example

Step 5: Make Decision

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

6. Find the critical value for the particular significance

At the 0.05 level

T-TEST INDEPENDENT SAMPLES

6. Find the critical value for the particular significance

At the 0.05 level

So, at 0.05 level there is a significant

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

T-TEST INDEPENDENT SAMPLES

Diff of means= 0.64

T-TEST INDEPENDENT SAMPLES

t < tcritical at 5% level, 2-tail

There is no significant difference between the data test

A one-sample Z-test for proportions is a test

A one-sample Z-test for proportions is a test

A one-sample Z-test for proportions is a test

A one-sample Z-test for proportions is a test that

And this is the

A one-sample Z-test for proportions is a test

Here is our question:

A one-sample Z-test for proportions is a test