You are on page 1of 13

TYPE 1 AND TYPE 2 ERRORS

When you perform a hypothesis test,


there are four possible outcomes
depending on the actual truth (or
falseness) of the null hypothesis HO and
the decision to reject or not. This
outcomes are summarized in the table
shown.
REJECT

FAIL TO REJECT
Null Hypothesis is TRUE Correct Decision TYPE 1 ERROR
Null Hypothesis is FALSE TYPE 2 ERROR Correct Decision

Four Possible Outcome:


● The null hypothesis (H0) is true and we do not reject it. (correct decision)
● The null hypothesis (H0) is true and we reject it. (incorrect decision known as a Type I
error)
● The null hypothesis is false but we reject it. (correct decision)

● The null hypothesis is false but we do not reject it. (incorrect decision known as a Type II
error)

The Greek letters α and β represent the probabilities.

α = probability of a Type I error = P(Type I error) = probability of rejecting the null


hypothesis when the null hypothesis is true.
β = probability of a Type II error = P(Type II error) = probability of not rejecting the null
hypothesis when the null hypothesis is false.
α and β should be as small as possible because they are probabilities of errors. They are rarely
zero.

TYPE 1 ERROR
● Also known as a false positive
● It occurs when a researcher incorrectly rejects a true null hypothesis. This means that your report
that your findings are significant when in fact they have occurred by chance.
● It is the incorrect rejection of a true null hypothesis.
Symbol: α (alpha)
Example 1:
It is known that the number of hours of work allowed is 8 hours.
H0: The number of hours of work allowed is equal to 8 hours. (TRUE)
H1: The number of hours of work allowed is not equal to 8 hours.
Decision: tell your workers that the number of hours of work allowed is not equal to 8 hours.
(Reject Ho = Type 1 Error)The null hypothesis is true but you reject it which is not a correct
decision. Therefore you are committing a Type 1 Error.

Example 2:
A patient undergoes a test to see if they have cancer or not.
H0: Cancer Free (TRUE)
H1: Cancer
Decision: tell the patient they have cancer but they don't (Reject Ho = Type 1 Error)
The null hypothesis is true but you reject it which is not a correct decision. Therefore you are
committing a Type 1 Error.

TYLE 2 ERROR
● Also known as a false negative.
● It occurs when a researcher fails to reject a null hypothesis which is really false. Here a
researcher concludes there is not a significant effect, when actually there really is.
● It is the failure to reject a false null hypothesis.
Symbol: β (Beta)

Example 1:
According to the national geographic, the green anaconda is the largest snake in the world.
H0: The green anaconda is not the largest snake in the world. (False)
H1: The green anaconda is the largest snake in the world.
Decision: tell the students that the anaconda is not the largest snake in the world when in fact it
is. (Accept Ho = Type 2 Error)
The null hypothesis is true but you reject it which is not a correct decision. Therefore you are
committing a Type 1 Error.

Example 2:
A patient undergoes a test to see if they have cancer or not.
H0: Cancer Free (FALSE)
H1: Cancer
Decision: tell someone they don't have cancer but they do (Reject Ho = Type 1 Error)
The null hypothesis is true but you reject it which is not a correct decision. Therefore you are
committing a Type 1 Error.

Word Problem Example:


A large nationwide poll recently showed an unemployment rate of 10% percent in the Philippines. The
mayor of a local town wonders if this national result holds true for her town, so she plans on taking a
sample of her residents to see if the unemployment rate is significantly different than 10%.
Solution:
Let p represent the unemployment rate in her town.
Here are the hypotheses she'll use:
H0: μ = 0.10
H1: μ ≠ 0.10

Determine the TYPE 1 and TYPE 2 Error:

Assuming that from the sample taken, the unemployment rate is actually 10%.
TYPE 1 ERROR: The mayor concludes that the towns unemployment rate is not 10% when it is
in fact 10%.
This is a Type I error — H0: is true, but she rejects it.
Assuming that from the sample taken, the unemployment rate is greater/less than 10%.

TYPE 2 ERROR: The mayor concludes that the towns unemployment rate is 10%, when it is in
fact not.
This is a Type I I error — H0: is false, but she accepts it.

One-Tail and Two-Tail Test


Again, hypothesis testing is run to determine whether a claim is true or not, given a population
parameter.
One-Tail Test

● Is a statistical hypothesis test in which the values for which we can reject the null hypothesis,
Ho, are located entirely in one tail of the probability distribution.
● Are also known as directional and one-sided tests. Because you can test for effects in only one
direction.
● It determine if there is a relationship between variables in single direction.
● A test of any statistical hypothesis is considered a one-tailed if the alternative hypothesis is:

● Keywords: > (greater than) or < (less than)


⮚ Example :
A government official in Cebu City claims that the dropout rate for local
schools is less than 30% due to poverty. Last year 200 out of 830 students
dropped out. Is there enough evidence to reject the government official's
claim?
Drop out rate < 30%

● Example :
A government official in Cebu City claims that the dropout rate for local
schools is greater than 30% due to poverty. Last year 200 out of 830
students dropped out. Is there enough evidence to reject the government
official's claim?
Drop out rate >30%
● Sign in alternative hypothesis: > (more than) or < (less than)

Example1:
The average IQ of the average population is 100. A researcher believes the average IQ of adults
is lower. A random sample of 5 adults are tested and scored:
69 79 89 99 109 (sample standard deviation or s =15.81)
Is there enough evidence to suggest the average IQ is lower?

There are five steps in order to perform for a hypothesis test. This steps can be used for any type
of hypothesis testing example.
Hypothesis Testing Steps:
a. State the Null Hypothesis (H0 ) and the Alternative Hypothesis (H1 )
b. Choose level of significance (α)
c. Find critical values
d. Find test statistics
e. Show your conclusion.

Solution:
a. H0 : μ = 100
The null hypothesis is what is currently believed what is true or what is currently accepted. In
this example it says the average IQ of the adult population is 100. Therefore our null hypothesis
is mu equals to 100.
H1 : μ < 100
The alternative hypothesis is always what is being claimed . In this example it says a researcher
believes, he is claiming that the average IQ of adults is lower. Take note of the word lower. So
our alternative hypothesis is written with this greek letter miu less than 100. Remember that
anytime the alternative hypothesis is written with a less than or greater than symbol, this
means were going to perform a one tailed test.
b. α = 0.05
The level of significance is the area in the tails. Since there is no mentioned or given level
of significance then we need to choose. In here 0.05 is chosen since it is the most common
used level of significance. Notice how the only area or the only area that is shaded is at
the bottom tail? This is because we're forming a one tailed test because of our alternative
hypothesis that is mu less than the average of the population. So the only area of interest
is below the average of the population which is at the middle. And the shaded part have
an area equal to 0.05.
c. In order to find the critical values in a one tailed test, first we must take note that the critical
values are just the z values or the t values that separates the tail from the rest of the curve. In
this particular example, a T Test is used because of this conditions:
When to use T Value:
● σ is unknown.
● The sample size is less than 30 (n < 30)

Since we meet this two conditions then we are going to use the T Table.

Using the T Table:

Now when using the T Table we must keep in mind the area in one of the tails. In this example,
we are using the level of significance which is 0.05 which is the area of our lower tail. So if you
go to the T Table you'll see in the top row, we have the area of one tail which 0.05. This means
that the column of 0.05 will be used to locate the critical value.
Α = 0.05
Now as you can see, at the left side first column, we can locate the degrees of freedom. Now take
note that the degrees of freedom is always one less the sample size. Therefore:
df= n-1 = 5-1 = 4
now that we have a degrees of freedom which is 4 we can now locate the critical value of the one
tail test which is where the intersection of the 0.05 and 4 and that is t value 2.132.

Going back to the curve, the critical value lies below the average or below the middle of the curve.
This means that it is going to be a negative number. So instead of using +2.132, -2.132 is used as the
critical value. Take note that this critical value is important because it separates this area in the tail with
the rest of the curve and this area of the tail which is in red is called the rejection region. And the
reason why this rejection region is important because it will used to determine if we are going to reject
or accept the null hypothesis if our T Value happens to fall in the rejection region.
Critical Value = - 2.132

d.In the test statistics, it is going to be either a z value or t value and since T test is used so this is going
to be a t value. In order to find the t value we will be using this formula:
T is equal to the average of the sample minus the average of the population all divided by the standard
deviation of the sample divided by the square root of the sample size n.

= 89-100/15.81/√5 = -1.56 ---> TEST STATISTIC

We know that in order to get the average of the sample is we are just going to add the test scores and
then divide the sum by 5.
x̄ = 69 + 79 + 89 + 99 + 109 = 445 / 5 = 89
Now this value of – 1.56 is our test statistics. To draw the conclusion, we are going to estimate
and plot it on the curve. Now its somewhere in between the critical value and the average. Notice
that it is not located in the area of the tail. Where the rejection region lies. Which means we
cannot reject the null hypothesis. We have to accept the null hypothesis that the average IQ is
equal to 100.

Accept H0 : μ = 100
e. Is there enough evidence to suggest the average IQ is lower?
Conclusion:
No, there is not enough evidence because we accept that the average IQ of adults is equal to 100
Example 2:
The PBA believes the chance of the Ginebra winning the coin toss is greater than 50%. In a
random sample of 200 coin tosses, the Ginebra won 118 times. Is there enough evidence to
suggest the Ginebra are cheaters?

Solution:
a. H0 : p = .50
The null hypothesis is what is currently believed what is true. In this case we have to assume that
the Ginebra Basketball Team are innocent. Innocent until proven guilty. So if they are innocent,
we know that the chance of winning a fair coin toss is 50%. So the null hypothesis is that the
proportion that the Ginebra Baketball Team win a coin toss is equal to 50%.
H1 : p > .050
The alternative hypothesis is always what is being claimed. In this example it says the PBA
believes that the chance of the Ginebra winning the coin toss is greater than 50%. Take note of
the word greater than. The alternative hypothesis is assuming that the Ginebra Basketball team
are guilty. Where they win more than 50% of the time and anytime you see again a greater than
or less than symbol in the alternative hypothesis this means, your going to use a one tailed test.
b. α = 0.01
The level of significance is the area in the tails. Since there is no mentioned or given level of
significance then we will have to choose. In here, lets choose 0.01 but it could be any value it's
usually between .01 and .05. Since were using the level of significance .01 this means the area in
this tail is equal to .01. and why is it again at the upper tail that 0.01 is located? It is because of
the alternative hypothesis which is p greater than 50%.

c. Now again, the critical values are the z values or t values that separates the tail from the
rest of the curve. Now how do we know if its z value or t value? With proportions, we use a z
value for a large enough sample size. How do we determine if its a large enough sample size? We
use the formula Px N. Where P is the average proportion of the population and n is the sample
size that must be greater than 5. and also 1-p times the sample size n must be greater than 5.
When to use Z Value
Formula:
Pn > 5
(1-P) n > 5
now by subtituting this... we will have....
50% (200) = 100 > 5
( 1 - 50%) 200 = 100 > 5
So this means we have a large enough sample size so that's why we can approximate this using a
normal curve and use a z test and critical value is going to be a z value. And to find the z value
we are going to use the z table.

Using the Z Table:

So in order to find this critical value of Z in the Z Table first remember that We have an area of
the tail which is 0.01.
α = 0.01

Going to the Z table, we can locate the 0.01 at


the 3 column and with its corresponding z value at the 4th column beside the 0.01 which is 2.326.
rd

Take note that this critical value is important because it separates this area in the tail with the rest of the
curve and this area of the tail which is in red or shaded is called the rejection region. And the reason
why this rejection region is important because it will used to determine if we will reject or accept the
null hypothesis. And in this example, if this T value happens to fall in the rejection region then that
means that we can reject the null hypothesis that the the proportion is equal to 50% and accept the
alternative hypothesis that the proportion is greater than 50%.
Critical Value = 2.326

d. In the test statistics, it is going to be either a z value or t value and since Z test is used so we are
going to find the z value. In order to find the z value we will be using this formula:
Z= proportion of the sample-proportion of the population all divided by the square root of the
proportion of the population times 1 – the proportion of the population all divided by the
square root of the sample size n.

= 0.59-0.50/√0.50(0.50)/200 = 2.55 ---> TEST


STATISTIC

to get our proportion of the sample we will simply divide the number of observed events over the
sample size n.

= 118 / 200 = 59% or 0.59

this value of 2.55 is our test statistics To draw a conclusion, we are going to estimate and plot it
on the curve. Now when plotting it in the normal curve , we know that our value of 2.55, our
test statistics is certainly greater than 2.326 which is the critical value. Since we know its greater
so it will be somewhere above the critical value or to the right of the critical value. notice that it
lies inside the rejection region. This means that we can reject the null hypothesis which states
that the proportion is equal to 50% and we can accept the alternative hypothesis which states
that the proportion is greater than 50%.
Accept H1 : p > 50%
e. Is there enough evidence to suggest the Ginebra are cheaters?
Conclusion:
Yes, there is enough evidence to suggest that the Ginebra are cheaters. Because we accept the
alternative hypothesis that the proportion is greater than 50%.
TWO-TAILED TEST
● Is a statistical hypothesis test in which the critical area of a distribution is two-sided and tests
whether a sample is greater or less than a range of values.
● Are also known as non-directional and two-sided tests. Because you can test for effects in both
directions
● It determine if there is a relationship between variables in either direction.
● A test of any statistical hypothesis is considered a two-tailed if the alternative hypothesis is:

● Keyword: = (equal)
● Sign in alternative hypothesis: ≠ (not equal to)
In this test, α is bifurcated or divided into two equal parts, placing half on
each side, i.e. it considers the possibility of both positive and negative
effects. It is performed to see, whether the estimated parameter is either
above or below the assumed parameter, so the extreme values, work as
evidence against the null hypothesis.

Example 3:
The average IQ for the adult population is 100 everyday with a standard deviation of 15. A
researcher believes this value has changed. The researcher decides to test the IQ of 75 random
adults. The average IQ of the sample is 105.I s there enough evidence to suggest that the average
IQ has changed?
Solution:

a. H0 : μ = 100
As we know the null hypothesis is what is currently believed what is true or what is currently
accepted. In this example it says the average IQ of the adult population is 100 and this will be
written as mu equal sign 100.
H1 : μ ≠ 100
Again The alternative hypothesis is always what is being claimed and must be the opposite of the
null hypothesis. In this example it says a researcher believes, he is claiming the value has
changed. Take note of the word has changed. He doesn't know if the average have gone up or the
average has gone down. But he believes that the average has changed, that the average is not 100
anymore. So our alternative hypothesis will be written with this greek letter miu not equal to
100.
Take note that anytime you see the alternative hypothesis written with the not equal sign. This
means we are going to perform a two tailed test.
b. α = 0.05
As mentioned Sometimes the level of significance is given to you in a problem but if not then you
need to choose the level of significance. The level of significance is just the area in the tails. In this
case, the level of significance is the area in the two tails. Remember that this is a two tailed test. So we
have to use both of the tails. So the level of significance is just the area in both the tails. In this
example, the level of significance chosen is 0.05 because it is the most common level of significance
used unless again if the level of significance is given to us. Take note that this two tails are also
symmetrical, they have the same exact area. Also remember that in this test, the alpha α is
bifurcated or divided into two equal parts, basically one tail will have an area of 0.025
and since it is symmetrical, the other will also have a 0.025 area and if you combine them you will have
the level of significance 0.05. Also one thing to point out, we know that the total area under the curve is
equal to 1 or 100%. So if the area of the tails is equal to 0.05 that means the middle of the curve (the
area) has to be equal to 95% or 0.95

c. In order to find the critical values


in a two tailed test, first we must take note that the critical values are
just the z values or the t values that separates the tail from the rest of the curve. Now, we always
use a Z Value if the population standard deviation is known. So in this example it says the
average IQ of the adult population is 100 with a standard deviation of 15. and since this is given
to us we are we are going to use a Z test instead of a t test.
So in order to find these critical values we are going to use a Z Table. So
theres lots of information we can use to find the values in the Z Table. This is a
95% interval so we can use a 95% confidence interval to find the Z values. Or we
could just use the area in one of the tails which is 0.025. whichever of the two used
there will be the same corresponding z value in the table.
Using the Z Table:
Now going to the Z Table, if we have a confidence level of 95% that means the area in one of the
tails is going to be 0.025. And that will give us a Z value of 1.96. So the critical value is going to
be 1.96.

Critical Value = 1.96

So we know now that the critical value on the upper tail will be 1.96 and the critical value on the lower
tails is going to be -1.96 since its below the average. Remember that this area in red is what we call
the rejection region. Both the area on the upper and lower tails which is in red are the rejection region.
Both rejection region will be used to determine if we reject or accept our null hypothesis. And if our Z
Value for our sample in the test statistics falls in any of the rejection region or if it falls in any of these
areas in red then that means we can reject our null hypothesis. Which says that the average is equal to
100. It is important because it allows us to draw a conclusion after we perform our two tailed test.

d. In finding the Z Value for the given sample. Since we chose to use a Z Value because the
population standard deviation was known then we are going to use the Z Test Statistic formula. Which
is Z test statistic equal to the average of the zample which is written with x bar minus the average of
the population which is written with the greek letter mu all divided by the standard deviation of the
population divided by the square root of the sample size n.

= 105-100/15/√75 = 2.89 ---> TEST STATISTIC

Where:
x̄ = average of the sample size
μ = average of the population
n = sample size
Now that we have our z test statistic value which is 2.89, if we plot it on the upper tail, it will be put
above the critical value of 1.96 since 2.89 is larger than it. If we also plot it in the left tail, it be be
below -1.96 since it will be -2.89. Notice that in both tails, the Z values lies inside the rejection region.
Therefore we can reject the null hypothesis that the average IQ of Adult is equal to 100 and accept the
alternative hypothesis which is the average is not equal to 100.

Reject H0 : μ = 100
Accept H1 : μ ≠ 100

e. Is there enough evidence to suggest the average IQ is lower?


Conclusion:
Yes there is enough evidence because we accept that the average has changed. It not
equal to 100

You might also like