ADVANCE STATISTICS REVIEWER

I. Hypothesis Testing

Hypothesis Testing
> a decision-making process for evaluating claims about population
> testing an assumption that we can make about a population

Hypothesis
> an assumption or conjecture about a population parameter which may or may not be true
> an assertion or tentative theory which aims to explain facts about the real world

Steps involved in hypothesis testing:

1. State the null and alternative hypotheses
2. Assign the level of significance
3. Determine the test statistics to be used
4. Define the critical region or rejection region
5. Make the appropriate computation
6. Make a statistical decision
7. Interpret the result and draw valid conclusion

Null Hypothesis (Ho)

> the initial claim
> shows no significant difference, no changes, nothing happened, no relationship between
two parameters
> The independent variable has no effect on the dependent variable
> Ho: μ1 = μ2

Alternative Hypothesis (Ha or H1)

> the contrary to the null hypothesis
> shows that there is significant difference, an effect, change, relationship between a parameter
and specific value
> The independent variable has an effect on the dependent variable.

= equal to, the same as, not changed from, is

≠ not equal, different from, changed from, not the same as

> Greater than, above, higher than, longer than, bigger than, increased, at least

< Less than, below, lower than, smaller than, shorter than, decreased or reduced
from, at most
II. Types of Error

Parameter
> any numerical quantity that characterizes a given population or some of its aspects
> denotes the true value

Statistic
> the numerical measure that is calculated from the sample
> known number and a variable that depends on the portion of the population

Two parameters that can characterized distribution:

1. Mean
2. Standard Deviation

Different symbols used to denote parameters:

Measure Statistic Parameter

mean x̄ μ

variance s² σ²

Standard deviation s σ

proportion p̂ p

Types of Error:
Decision Actual condition

Ho is true Ho is false

Reject ho Type I (a) error correct decision

Do not reject Ho correct decision Type II (b) error

Type I error
> rejecting the null hypothesis when it is true
> probability denoted by alpha (α)
> known as false positive
> tends to assert something that is not really present, a false hit

alpha region
> the normal curve that shows the critical region
Type II error
> accepting the null hypothesis when it is false
> probability denoted by beta (β)
> known as false negative
> fails in identifying something that is present, a miss

beta region
> the normal curve that shows the acceptance region

Note!
• the larger the value of alpha, the smaller the value of beta

Correct decision
> when the null is false and you reject it
> when the null is correct and you accept it

Examples:

1. Maria insists that she is 30 years old when, in fact she is 32 years old. What errors is Maria
committing?
– Type I error

2. A man plans to go hunting the Philippine monkey - eating eagle believing that it is a proof of
his mettle. What type of error is this?
– Type II error

3. Stephen says that he is not bald. His hairline is just receding. Is he committing an error? If
yes, what type of error?
– Type I error

4. How to state type of errors in a sentence,

The mean number of years a teacher work before retiring is 30.

Type I error:
We conclude that the mean number of years a teacher work before retiring is not 30 years,
when it really is 30 years.

Type II error:
We conclude that the mean number of years a teacher work before retiring is 30 years, when in
fact it really is not 30 years.
III. Level of Significance and Type of Test

Significance
> defined as the quality of being statistically significant

Level of Significance (α)

> refers to the degree of significance in which we accept or reject the null hypothesis
> 100% accuracy is not possible in accepting or rejecting a hypothesis
> the probability of making the wrong decision when the null hypothesis is true
> In public health research alpha is usually 0.01 or 1%
> in social science, alpha is usually 0.05 or 5% and 0.10 or 10% in other studies
> there is 1%, 5%, or 10% probability of rejecting a true null hypothesis

Examples:

1) Maria uses 5% level of significance in proving that there is no significant change in the
average number of enrollees in the 10 sections for the last two years. It means that the chance
that the null hypothesis (Ho) would be rejected when it is true is 5%.
– α = 0.05 is the area

2) In 2015, it was recorded that around 34% of the population in 2015 were not married. A
researcher surveyed a random sample of 500 couples. He found out that 18% of them were
living together but unmarried. Test at 5% significance level if the current percentage of
unmarried couples is different from 34%.
– α/2 = 0.025

3) An average construction worker hourly rate pay in the Philippines is Php 62.50 with a
standard deviation of Php 6.01. A random sample of 20 manufacturing workers were asked on
their hourly rate and found out that they had an average of Php 50 hourly rate pay with a
standard deviation of Php 5.00. Construct a 90% confidence interval that the average hourly
rate for construction workers is higher than the average hourly rate for manufacturing workers
– α = 0.10

4) A chemist invented an additive to increase the lifespan of rechargeable battery. The said
additive will extend on average the battery’s lifespan to 48 months. Test the hypothesis at 0.01
level of significance that the average lifespan is higher than 48 months.
– α = 0.01

5.) The average number of years to finish basic education is 14. A sample of 30 senior high
school students were asked and found out that the mean number of years to finish their basic
education is 12 with a standard deviation of 2 years. Test the hypothesis at 93% confidence
interval that the average number of years to finish basic education is less than 14 years.
– α = 0.07
 Z-TEST (σ is given) T-TEST(σ is not given)

The average number of years to finish basic In 2015, it was recorded that around 34%
education is 14. A sample of 30 senior high school of the population in 2015 were not married.
students were asked and found out that the mean A researcher surveyed a random sample of
number of years to finish their basic education is 500 couples. He found out that 18% of
12 with a standard deviation of 2 years. Test the them were living together but unmarried.
hypothesis at 93% confidence interval that the Test at 5% significance level if the current
average number of years to finish basic education percentage of unmarried couples is
is less than 14 years. different from 34%.

Two-Tailed Test (nondirectional)

> When the alternative hypothesis is two-sided like Ha: μ ≠ μo
> When the alternative hypothesis contains the inequality (≠) symbol

One-Tailed Test (directional)

> When the given statistics hypothesis assumes a less than (<) or greater than (>) value
> either right-tailed or left-tailed
Two-Tailed Test Right-Tailed Test Left-Tailed Test

Null Hypothesis Ηο: μ = μο Ηο: μ = μο or Ηο: μ = μο or

Ηο: μ ≤ μο Ηο: μ ≥ μο

Alternative Ηα: μ≠ μο Ηα: μ > μο Ηα: μ < μο

Hypothesis

Examples:
1. The school registrar believes that the average number of enrollees this school year is not the
same as the previous school year.
Let μo be the average number of enrollees last year.
Ho: μ = μo
Ha: μ ≠ μo If Ha uses ≠, use a two-tailed test.

However, if the school registrar believes that the average number of enrollees this year is less
than the previous school year, then you will have:
Ho: μ = μo
Ha: μ < μo If Ha uses <, use the left-tailed test.

On the other hand, if the school registrar believes that the average number of enrollees this
school year is greater than the previous school year, then you will have:
Ho: μ = μo
Ha: μ > μo If Ha uses >, use the right-tailed test.
IV. Decisions (p-value)

alpha value
> significance level chosen by the researcher to determine the threshold for rejecting the Ho

p-value
> probability of obtaining results as extreme as the observed results, assuming the Ho is true
> how likely you are to get the sample statistics if the null hypothesis were actually true

Note!
• reject Ho if p-value ≤ αlpha
• fail to reject Ho if p-value > αlpha

Examples:
1) Make a decision using the p-value 0.0752

a) α = 10% b) α = 0.05
p-value = 0.0752 < α = 0.10 p-value = 0.0752 > α = 0.05
– Fail to Reject Ho – Reject Ho

2) A random sample of 25 observations from a normal population has a mean of 22. Assuming
the population standard deviation is known to be 4.9.
Can we conclude at α = 0.01 that the population mean note!
is significantly below 24? σ – population standard deviation
n = 25 x̄ = 22 σ = 4.9 α = 0.01 n – sample size
x̄ – sample mean
μ – population mean
Ho: μ ≥ 24
Ha: µ < 24 (Left-tailed Test)

z = x̄ - μ Test Statistic
σ = 22 −24 = –2 = -2.04
√n 4.9/√25 0.98 Since p-value > α = 0.01 there is not enough
– fail to reject Ho evidence that μ < 24

3) A random sample of 25 observations from a normal population has a mean of 22. Assuming
the population standard deviation is known to be 4.9. Can we conclude at α = 0.05 that the
population mean differs from 24?
n = 25 x̄ = 22 σ = 4.9 α = 0.05
Ho: μ = 24
Ha: μ ≠ 24 (Two-tailed Test)

z = x̄ - μ Test Statistic p-value = 2(0.0207) = 0.0414

σ = 22 −24 = –2 = -2.04 Since p-value < α = 0.05 there is enough
√n 4.9/√25 0.98 – Reject Ho evidence that μ ≠ 24
V. Decisions (Traditional Method)

Critical Values α Values

Confidence Two-Tailed One-Tailed Confidence Two-Tailed One-Tailed
Level Test Test Level Test Test

90% ± 1.64 ± 1.28 90% α/2 = 0.05 α = 0.1

95% ± 1.96 ± 1.645 95% α/2 = 0.025 α = 0.05

99% ± 2.58 ± 2.33 99% α/2 = 0.005 α = 0.01

Graphical Representation of Analysis Error

Significance Level
> probability of committing a type 1 error

Levels of Significance:
▪ If p ≤ 0.05 ; the difference is said to be significantly at the 0.05 or 5%, or less level
▪ If p ≤ 0.01 ; the difference is said to be significantly at the 0.01 or 1%, or less level
▪ If p ≤ 0.001 ; the difference is said to be highly significant

One population Test

> conducted on one sample purportedly coming from a population with mean μ
> sometime called a significance test for a single mean
> this is used if the sample is large (n ≥ 30)

Central Limit Theorem

> states that the sampling distribution of the mean will follow a normal distribution under the met
conditions

Steps in the Traditional Method of Hypothesis Testing:

1. Describe the population parameter of interest (mean, proportion)
2. Formulate the hypothesis
3. Check the assumptions
4. Choose the significance level size for α
5. Select the appropriate test statistic
6. State the decision rule for rejecting or not rejecting the null hypothesis
7. Compare the test statistic and critical value
Example:
A researcher used a developed problem-solving test to randomly select 50 grade 7 students. In
this sample, X=80 and s = 10. the mean and standard deviation were 75 and 15 respectively.
Use 95% confidence level to answer the following questions:
1. Does the sample mean differ significantly from the population mean

Step 1: Describe the population parameter of interest

The parameter of interest is the mean of the population where the sample came from

Step 2: Formulate the hypothesis

Ho : μ = 75
Ha : μ ≠ 75

Step 3: Check the assumptions

Since n= 50, we can use CLT; distribution is normal

Step 4: Choose a significance level

α = 1 – 95 = 0.05

Critical values ±1.96.

Step 5: Select the appropriate test statistic

σ = 15

Step 6: State the decision rule

Left One-tailed Test

Decision rule: Reject Ho if the test statistic is less than the critical value

Right One-tailed Test

Decision rule: Reject Ho if the test statistic is greater than the critical value

Two-tailed Test
Decision rule: Reject Ho if the test statistic is greater than the upper critical value or less than
the lower critical value

Step 7: Compare the test statistic and critical value

2.36 > 1.96
There is enough evidence to reject the null hypothesis.
There is significant difference between sample and
population mean
VI. Decisions (T-Test)

Test statistic
> random variable that is calculated from sample data and used in hypothesis test

Z - Test
> the population standard deviation (σ) is known
> population is normally distributed
> z-score is calculated with population parameters
> used to validate a hypothesis that the sample drawn belongs to the same population
> the sample size is large (n ≥ 30)

T - Test
> used when the population variance or standard deviation (σ) are not known
> population is normal or nearly normally distributed

Level of Significance (α)

> refers to the degree of significance in which we reject or do not reject the Ho
> basis for the critical or the rejection region dictated by the Ha
Common values of statistical significance:
0.01 – highly significant
0.05 – statistically significant
0.10 – significant

Rejection Region (Critical region)

> range of values of the test value which indicates that there is a significant difference that the
Ho should be rejected

Non - Rejection Region (Acceptance region)

> range of values of the test value which indicates that the difference was statistically
insignificant and that we failed to reject the Ho

Critical Value
> a point (boundary) on the test distribution that is compared to the test statistic to determine if
the null hypothesis would be rejected

Sample standard deviation

> used to make estimates or inferences about the population standard deviation
> tend to be lower than the real standard deviation of the population
Type of Levels of Significance

Test α = 0.01 α = 0.025 α = 0.05 α = 0.10

left-tailed test c = -2.326 c = -1.960 c = -1.645 c = -1.28

right-tailed test c = 2.326 c = 1.960 c = 1.645 c = 1.28

two-tailed test c = ±2, 575 c = ±2.326 c = ±1.960 c = ±1.645

Examples:
Determine the critical value and illustrate the rejection region under the normal curve by using
the given information.

1) Ha: p ≠ 0.52
α = 0.05
Critical Value: ±1.960

2) Ha: p > 0.35

α = 0.01
Critical Value: 2.326

3) Ha: p < 0.70

α = 0.10
Critical Value: - 1.28
4) A sample of 120 students is randomly selected from Masayahin High School and 21 of them
said they are left-handed. Test the hypothesis that less than 22% of the students are left-handed
by using a = 0.05 as the level of significance.

α = 0.05
Ha: p < 0.22
Critical Value: -1.645

5) In Kalinga Special Education School, a sample of 144 students was chosen and among
them, 48 are diagnosed with Attention Deficit Hyperactivity Disorder (ADHD). At a = 0.01, test
the hypothesis that the proportion of ADHD students in the school is not 0.40.

α = 0.01
Ha: p ≠ 0.40
Critical Value: ± 2.575

Formulas:

Population standard deviation

σ = population standard deviation
N = the size of the population
Xi = each value from the population
µ = the population mean

Sample standard deviation

s = sample standard deviation
Σ = sum of...
X = each value
x̄ = sample mean
n = number of values in the sample

Test statistics
x̄ = sample mean
s = sample standard deviation
μ = assumed mean
n = sample size