Ch-4:Testing of Statistical

Hypothesis
-ER. MANIRAM BHATTARAI
Inferential Statistics
◦ Inferential statistics are often used to compare the differences between the
treatment groups.
◦ Inferential statistics use measurements from the sample of subjects in the
experiment to compare the treatment groups and make generalizations about
the larger population of subjects.
Types of Inferential Statistics
Testing of hypothesis
Hypothesis:
• Hypothesis can be defined as a formal statement, which gives the explanation
about the relationship between the two or more variables of the specified
population
• It helps researcher to translate the given problem to a clear explanation for the
outcome.
• Hypothesis is a quantitative statement about the population parameter.
• The population parameter will be correct if there is no significant difference
between population parameter and sample statistics and wont be correct if
there is significant difference.
Types of hypothesis
There are normally two types:
1. Null hypothesis:
◦ Null hypothesis means no difference or zero difference between population
parameter and statistics computed from selected random sample data.
◦ So, null hypothesis means presumption of no difference.
◦ Generally, the null hypothesis is set up in such a way that it should be
rejected. It is denoted by
Types of hypothesis
2. Alternative hypothesis:
◦ Another hypothesis complementary to null hypothesis is alternative
hypothesis.
◦ It is an alternation of the null hypothesis.
◦ If the decision maker rejects the null hypothesis on the basis of sample
information, then the alternative hypothesis is accepted.
◦ It is denoted by .
Error in hypothesis testing
State of nature

is true is false

Accept Correct Decision Error Decision

Decision (Type II Error)
Alternatives
Reject Error Decision Correct Decision
(Type I error)
Error in hypothesis testing
◦ Any decision made based on a hypothesis test may be incorrect. In the
framework of hypothesis tests there are two types of errors: Type I error and
Type II error.
◦ A type I error occurs if a true null hypothesis is rejected . (a “false positive”)
◦ A type II error occurs if a false null hypothesis is not rejected .(a “false
negative”)
◦ In other words, a type I error is detecting an effect that is not present, while a
type II error is failing to detect an effect that is present.
Level of Significance (α)
◦ The level of significance, usually denoted by α, is a threshold set by the
researcher that determines how much risk of a Type I error (incorrectly
rejecting a true null hypothesis) is acceptable in a hypothesis test.
◦ The level of significance is the probability of rejecting the null hypothesis
when it is actually true. For example, if α = 0.05, there is a 5% chance of
making a Type I error.
◦ Setting α: Commonly used values are 0.05, 0.01, and 0.10.
◦ A lower α (like 0.01) means you require stronger evidence to reject the null
hypothesis, reducing the chance of a false positive but possibly increasing the
chance of a false negative (Type II error).
Level of Significance
Role in Hypothesis Testing :
◦ Before collecting data, you choose a significance level to determine your
criteria for rejecting the null hypothesis.
◦ After computing the test statistic and the corresponding p-value, you
compare the p-value to α. Or you can find out the critical values from table
with help of α and find out respective test statistics values then compare.
◦ If the p-value is less than or equal to α, you reject the null hypothesis,
concluding that the observed effect is statistically significant. If the p-value is
greater than α, you fail to reject the null hypothesis.
Level of Significance
◦ Setting α is about balancing risk. A stricter significance level (e.g., 0.01)
minimizes the chance of falsely finding an effect but may also make it harder
to detect a true effect. Researchers must choose α based on the context and
potential consequences of errors.
◦ Understanding the level of significance is crucial because it underpins the
decision-making process in hypothesis testing and ensures that your
conclusions are backed by an acceptable level of statistical confidence.
Steps of hypothesis testing
◦ Hypothesis testing is a systematic procedure for deciding whether the results
of a research study support a particular theory which applies to a population.
◦ Hypothesis testing uses sample data to evaluate a hypothesis about a
population.
◦ There are generally two approaches but here we will discuss about critical
region approach.
Critical value approach for hypothesis
testing
Step 1: Set up null Hypothesis.
Step 2: Set up the alternative hypothesis.
Step 3: Test statistics: Define test statistics and compute it under is true.
Step 4: Selecting the level of significant:
The hypothesis are tested on pre-determined level of significance.
Generally we use 1%, 5% and 10% level of significance.
Critical value approach for hypothesis
testing
Step 5: Obtain Critical Value:
Obtain the critical value and critical region of the test statistics from the
appropriate table.
Step 6: Comparison and Decision:
If calculated value of test statistics is ≤ tabulated value then is accepted.
If calculated value of test statistics is > tabulated value then is rejected.
One tailed test and two tailed test
One tailed test and two tailed test
Basis of Comparision One-tailed test Two-tailed test
Meaning A significance test which has A significance test which has
only one end two ends
Hypothesis Directional Non- directional
Region of rejection Either left only or right only Both left and right
Determines If there exists a relation If there exists a relation
between variables in specified between variables in either
direction direction
Result Greater or less than certain Greater or less than certain
value ranges of value
Sign in alternative < or > ≠
hypothesis
Critical Region
Some symbols:
Parametric and Non-parametric Test
◦ Parametric and non-parametric tests are two broad categories of statistical tests
used to analyze data, depending on the nature of the data and the assumptions
that can be made about it.
◦ Here's a clear comparison between them:
Feature Parametric Tests Non-Parametric Tests
Requires assumptions about the
No strict assumptions about
Assumptions population distribution (usually
population distribution.
normal distribution).
Used for interval or ratio scale data Suitable for ordinal or nominal data
Data Type (continuous data). (categorical or ranked).
t-test, Z-test, ANOVA, Pearson Mann-Whitney U test, Kruskal-Wallis
Examples correlation test, Spearman correlation
Less robust to outliers and non- More robust to outliers and skewed
Robustness normality distributions
Generally more powerful if Less powerful when parametric
Power assumptions are met assumptions are true
Requires data with known scale and Works with data without equal
Measurement Scale equal intervals intervals or unknown scale
When data is normally distributed and When data is not normally distributed
Use Case variances are equal or sample size is small
Usually simpler and more
Computational Complexity straightforward
Can be more complex or less intuitive
Z-test
◦ Z test is one of the important parametric test which is based on the assumption
of normality and large the sample size which is more than 30. Z-test is used for
:
1. Test of significance of single mean (Testing the significance of difference
between sample mean and population mean)
2. Test of significance of difference between two means (Testing the
significance of difference between two independent sample means)
3. Test of significance of a sample (single) proportion (Testing the significance
of difference between sample proportion and population proportion)
4. Test of significance of two sample proportion (Testing the significance of
difference between two sample proportions)
Critical value table for Z-test
One-Tailed Test (Critical Two-Tailed Test (Critical
Significance Level (α)
z) z)
0.10 ±1.28 ±1.645

0.05 ±1.645 ±1.96

0.025 ±1.96 ±2.24

0.01 ±2.33 ±2.576

0.001 ±3.09 ±3.291

Z-test (Single Mean)
Step 1: State the Hypotheses
◦ Null hypothesis (H₀): μ = μ₀
◦ Alternative hypothesis (H₁):
Two-tailed test: μ ≠ μ₀
Left-tailed test: μ < μ₀
Right-tailed test: μ > μ₀
◦ Where: μ = population mean (unknown) and μ₀ = hypothesized mean
Step 2: Set the Significance Level (α)
◦ Common choices: α = 0.05, 0.01, 0.10
◦ This determines the rejection region in the standard normal (Z) distribution.
Z-test (Single Mean)
Step 3: Calculate the Test Statistic
Use the formula:
◦ where = sample mean, μ₀ = hypothesized population mean, σ = population
standard deviation (known), n = sample size and = standard error of mean
Step 4: Determine the Critical Value(s)
Use the Z-table to find the critical Z-value(s) based on α:
◦ Two-tailed test: use ±
◦ One-tailed test (left): use −
◦ One-tailed test (right): use +
Z-test (Single Mean)
Step-5: Decision
◦ If calculated z is lesser than or equal to critical value (tabulated value) of z
then null hypothesis is accepted.
◦ If calculated z is greater than critical value (tabulated value) of z then null
hypothesis is rejected.
Z-test (Single Mean)
1. A sample of 400 male students is found to have a mean height of 171.38 cm.
Can it be reasonably regarded as a sample from a large population with mean
height 171.17 cm and standard deviation 3.30 cm? Use critical value
approach.
[Ans: Null hypothesis is accepted]
2. The mean income of the random sample of 100 employees of an industrial
concern was found to be Rs.3000. If the standard deviation of the population
was Rs 25, find the standard error of the mean and also test whether the
sample mean differs from population mean of Rs 2850.
[Ans: 2.5,Null hypothesis is rejected]
Z-test (Single Mean)
3. The ABC bulb company claims that the bulb has average life of 15000 hours with standard
deviation of 1200 hours. However, there has been complaint against this claim from the
regular customer that a sample of 60 bulbs has average life of only 13000 hours. Can you
justify this claim at 1% level of significance?
[Ans: Null hypothesis s rejected]
4. A random sample of size 60 from a large population gave the following distribution: Test the
hypothesis that this sample comes from a population with mean 20.
Value 5-10 10-15 15-20 20-25 25-30
Frequency 5 19 24 8 4

[Ans: Null hypothesis is rejected]

Z-test (Single Proportion)
Step 1: State the Hypotheses
◦ Null hypothesis (H₀): p=
◦ Alternative hypothesis (H₁):
Two-tailed test: p≠
Left-tailed test: p<​
Right-tailed test: p>​
◦ Where: p= actual population proportion (unknown) and = hypothesized
proportion
Step 2: Set the Significance Level (α)
◦ Common choices: α = 0.05, 0.01, 0.10 which determines the rejection region.
Z-test (Single Proportion)
Step 3: Calculate the Test Statistic
Use the formula:
◦ where = sample proportion, =hypothesized proportion, n=sample size and x=
number of successes
◦ Also ( =1) and standard error of proportion =
Z-test (Single Proportion)
Step 4: Determine the Critical Value(s)
Use the Z-table to find the critical Z-value(s) based on α:
◦ Two-tailed test: use ±
◦ One-tailed test (left): use −
◦ One-tailed test (right): use +
Step-5: Decision
◦ If calculated z is lesser than or equal to critical value (tabulated value) of z
then null hypothesis is accepted.
◦ If calculated z is greater than critical value (tabulated value) of z then null
hypothesis is rejected.
Z-test (Single Proportion)
1. A coin is tossed 800 times and heads appear 480 times. Can you infer that the
coin is unbiased at 1% level of significance? [Ans: H0 is rejected]
2. A sample of 600 persons selected randomly from a large city gives the result
that males are 53%. Is there reason to doubt the hypothesis that males and
females are in equal number in the city? [Ans: H0 is accepted]
3. The manufacturer claims that only 3% of the tube it manufactures is defective.
To test the claim of manufacturer, a random sample of 700 tubes is taken and
found to have 30 defective. Do you agree with the manufacturer? [Ans: H0 is
rejected]
Z-test (Single Proportion)
4. A distributer claims that only 4% of the items supplied by him is out of order. A random
sample of 300 items contained 18 items as out of order. Test the claim of the distributer.[Ans:
Ho is rejected]