Hypothesis Testing

01/23/2023 1
 Learning objectives

At the end of this chapter the student will be able to:

– Define hypothesis testing
– Identify with the concepts of null and alternative hypothesis
– Differentiate between type I and type II errors
– Explain the meaning and application of statistical
significance
– Explain the meaning and application of p-values

01/23/2023 2
 Hypothesis Testing
• We often gather sample data in order to assess how much
evidence there is against a specific hypothesis about the
population
• We use a process known as hypothesis testing to quantify
our belief against a particular hypothesis
• The purpose of hypothesis testing is to aid the researcher
in reaching a decision concerning a population by
examining a sample from that population
01/23/2023 3
Definition:
 A statement about one or more population
 A statistical hypothesis is an assumption or a statement which may
or may not be true concerning one or more populations.
 It is a process whereby the statistical importance of an
investigation finding is determined by calculating the value of the
chosen test statistic, which is taken from a population sample, and
seeing if the value is equal to or greater than the preset
significance level. If it is, the result is termed statistically
significant.
01/23/2023 4
 Con…
• Hypothesis is a testable statement that describes the nature
of the proposed relationship between two or more variables
of interest
• The purpose of the study is to collect data which will allow
the researcher to test the hypothesis
• Hypothesis testing is a statistical method usually used when
a comparison has to be made, such as between two drugs or
two procedures
01/23/2023 5
 Examples of Research Hypotheses

Population Mean
• The average length of stay of patients admitted to the
hospital is five days
• The mean birth weight of babies delivered by mothers
with low Socio Economic Status (SES) is lower than
those from higher SES.
Population Proportion
• The proportion of adult smokers in Debre Markos town is
p = 0.40
• The prevalence of HIV among non-married adults is
higher than that in married adults, etc
01/23/2023 6
There are five ingredients to any statistical test

 Null Hypothesis

 Alternate Hypothesis
 Test Statistic
 Rejection/Critical Region

 Conclusion

01/23/2023 7
Types of Hypothesis

1. Null Hypothesis

2. Alternate Hypothesis

01/23/2023 8
 1. The null hypotheses

• The main hypothesis which we wish to test is called the null

hypothesis, since acceptance of it commonly implies “no
effect” or “ no difference.”

• It is denoted by the symbol H0

• H0 is always a statement about a parameter (mean,

proportion, etc.) of a population.

– H0 is always about a population parameter, not about
a sample statistic
01/23/2023 9
 · It is not about a sample, nor are sample statistics used in
formulating the null hypothesis.

· H0 is an equality ( μ = 14) rather than an inequality ( μ ≥ 14

or μ < 14).
· Is a statement claiming that there is no difference between
the hypothesized value and the population value.
· The effect of interest is zero = no difference

· H0 is a statement of agreement (or no difference)

· States the assumption (hypothesis) to be tested

01/23/2023 10
 Con…

• Begin with the assumption that the H0 is true

– Similar to the notion of innocent until proven guilty

• Always contains equality sign (“=” , “ ≤” or “≥ ”)
• May or may not be rejected

01/23/2023 11
e.g. The mean height of College of Medicine and Health Sciences
students of Wollo University is 1.63m
H0 : μ = 1.63 m
• At present only 60% of patients with leukemia survive more than
6 years. A doctor develops a new drug. Of 40 patients, chosen at
random, on whom the new drug is tested, 26 are alive after 6
years. Is the new drug better than the former treatment?
• The null hypothesis of the above statement is written as :
H0 : p = 0.60
– Here, we are questioning whether the proportion of patients
who recover under the new treatment is still 0.60 ( and hope
that it will be improved; this will be shown in our choice of HA
in the next section).

01/23/2023 12
 2. Alternative Hypothesis (HA)

• The notation HA (or H1 ) is used for the hypothesis

that will be accepted if HO is rejected.
• The alternate hypothesis (HA) is the opposite
supposition of the null hypothesis.
• It is the actual research question of interest, stated in
a particular form that can be tested indirectly

01/23/2023 13
 Con…
 Is a statement of what we will believe is true if our sample
data causes us to reject H0.
 Is generally the hypothesis that is believed (or needs to be
supported) by the researcher
 Is a statement that disagrees (opposes) with Ho
 (The effect of interest is not zero)
 Never contains “=” , “ ≤” or “≥ ” sign
 May or may not be accepted
01/23/2023 14
 Con…
• In clinical research, the alternate hypothesis would state
that there is a difference in treatment outcomes between
the new drug and a placebo.
• Consider the previous example (patients with leukemia)

HO: P = 0.60

HA: P > 0.60

01/23/2023 15
Steps involved in testing about a hypothesis

1. State the research question in terms of statistical

hypothesis
– The null hypothesis, H0, is a statement claiming that
there is no difference between the hypothesized value
and the population value
• The effect of interest is zero
– The alternative hypothesis, H1, is a statement that
disagrees with the null hypothesis
01/23/2023 16
Example: Hypotheses

 H0: μ = μ0  H0: μ = μ0

 H1: μ ≠ μ0  H1: μ > μ0 H1: μ < μ0

 H0: μ1=μ2  H0: μ1=μ2

 H1: μ1 ≠ μ2  H1: μ1 < μ2 or μ1 > μ2

 Two‐tailed  One‐tailed
01/23/2023 17
 Research Hypotheses: The thing we are primarily interested in
“proving”

HA1: μ > μ0 HA1: μ > 1.6m

HA2: μ < μ0 HA2: μ < 1.6m Average height of patients

HA3: μ ≠ μ0 HA3: μ ≠ 1.6m

 Null Hypothesis: Things are what they say they are, status quo

Ho: μ = μ0 Ho: μ = 1.6m

01/23/2023 18
 Con…
2. Select a sample and collect data
• Categorical, continuous
3. Decide on the appropriate test statistic for the
hypothesis (Z, t, χ2, F, etc.).
E.g., One population

or

01/23/2023 19
 Con…

4. Select the level of significance for the statistical test

(α=0.05, 0.01, 0.001, etc.)
5. Determine the critical value.
– A value the test statistic must attain H0: True to be
declared significant

01/23/2023 20
01/23/2023 21
 Con…

 N.B. The rejection region (also called the critical

region) is the range of values of a sample statistic that
will lead to rejection of the null hypothesis

01/23/2023 22
• The most frequently used values of α and the corresponding
critical values of Z are:
 α (level of significance)
Two-tailed,
0.10: ± 1.64
0.05: ± 1.96
0.01: ± 2.58
One-tailed (less than)
0.1: - 1.28
0.05: - 1.645
0.01: - 2.33
One-tailed (Greater than)
0.1: 1.28
0.05: 1.645
0.01: 2.33
01/23/2023 23
6. Perform the calculation
7. Draw and state the conclusion.
– If the numerical value of the test statistic falls in the
rejection region, we reject the null hypothesis and conclude
that the alternative hypothesis‐testing process will lead to
this conclusion incorrectly only 100α% of the time when
H0 is true.
– If the test statistic does not fall in the rejection region, we
do not reject H0.

01/23/2023 24
Errors of hypothesis testing:
 The null hypothesis is either true or false.
 Correspondingly, Ho is either not rejected or rejected

Type of H0 true H0 false

decision

Reject H0 Type I error (α) Correct decision

(1-β)
Accept H0 Correct decision Type II error (β)
(1-α)
01/23/2023 25
 Type I error

• The error committed when a true Ho is rejected

• Considered a serious type of error
• The probability of a type I error is the probability of
rejecting the Ho when it is true
• The probability of type I error is α
– Called level of significance of the test

• Set by researcher in advance

• The probability of making a type I error is denoted by α
01/23/2023 26
Type II error: not rejecting the null hypothesis when it is
actually false.

The probability of making a type II error is denoted by β.

Power: The probability of rejecting the null

hypothesis when it is false. Power = 1‐ β.

The power of a test tells the likelihood that the

hypothesis of no treatment effect will be rejected
when the treatment has an effect.
01/23/2023 27
 Con…

• The smaller size of the treatment effect the harder it is

to detect.
• The larger the sample size, the greater the power of
the test.
• The exact procedure to compute the power of a test
depends on the test itself

01/23/2023 28
Hypothesis Test for One Sample
• Test for single mean
• Test for single proportion

Hypothesis Test for Two Samples

• Test for the difference between two population
means
• Test for the difference between two population
proportions

01/23/2023 29
Hypothesis testing for single mean known
variance

• Example :
– Researchers are interested in the mean level of some
enzyme in a certain population. They are asking: can
we conclude that the mean enzyme level in this
population is different from 25?
– Sample of size 10 from a normally distributed
population with a known variance, σ2 = 45. The
calculated sample mean is 22.
01/23/2023 30
Step 1: H0: μ = 25

H1: μ ≠ 25

Step 2: They collect a sample of size 10 from a normally

distributed population with a known variance, σ2 = 45. The
calculated sample mean is 22.

Step 3: Testing a hypothesis about population mean

– The population is normally distributed Population variance
is known
⇒ Z ‐ statistic is the appropriate one

01/23/2023 31
01/23/2023 32
01/23/2023 33
Step 7: Since ‐1.41 falls in the acceptance region we
accept the null hypothesis.
 Reject the alternative hypothesis

01/23/2023 34
01/23/2023 35
Example
• Serum amylase determination were made on a sample
of 15 apparently health subjects. The sample yielded a
mean of 96 units/100ml and a standard deviation of 35
units/100ml. The population was normally distributed
but the variance was unknown. We want to know
whether we can conclude that the mean of the
population is different from 120.

01/23/2023 36
01/23/2023 37
01/23/2023 38
 Hypothesis Tests for single Proportion

• Involves categorical values

• Two possible outcomes

– “Success” (possesses a certain characteristic)
– “Failure” (does not possesses that characteristic)
• Fraction or proportion of population in the “success”
category is denoted by p

01/23/2023 39
 Con…

01/23/2023 40
Hypothesis Testing about a Single Population Proportion
(Normal Approximation to Binomial Distribution)

01/23/2023 41
• In the general population of 0 to 4-year-olds, the annual
incidence of asthma is 1.4%.
• If 10 cases of asthma are observed over a single year in a
sample of 500 children whose mothers smoke, can we
conclude that this is different from the underlying
probability of P=0.014? CI = 95%

 H0 : P=0.014

 HA: P≠0.014
01/23/2023 42
 Con…
• The test statistic is given by:

01/23/2023 43
 Con…

• The critical value of Zα/2 at α=5% is ±1.96.

• Don’t reject Ho since Z (=1.14) in the non-rejection

region between ±1.96.
• We do not have sufficient evidence to conclude that
the probability of developing asthma for children
whose mothers smoke in the home is different from
the probability in the general population
01/23/2023 44
 Example 2

• A pharmaceutical company claims that a drug which it

manufactures relieves cold symptoms for a period of 10
hours in 90% of those who take it. In a random sample
of 400 people with colds who take the drug, 350 find
relief for 10 hours. At a 0.05 level of significance, is the
manufacturer’s claim correct?

01/23/2023 45
Step 1: HO : P = 0.90
HA : P < 0.90

Step 2: They collect a sample of size 400 from a normally

distributed population and out of these 350 find relief for 10
hours
• Step 3: Testing a hypothesis about population proportion

⇒ Z ‐ statistic is the appropriate one

⇒ Z = P1 – P2/ √P (1-p)/n
⇒ Z calc = (.875 - .90) / √(.90 x .10 /400)

⇒(.875-.90)/.015 = -1.67
01/23/2023 46
4. Select the level of significance for the statistical test
(α=0.05)
5 and 6: Determine the critical value and Perform the
calculation
• A value the test statistic must attain H0: True to be
declared significant
• Z tab (α = .05) = -1.64 and
• reject HO if Z calc < -1.64

01/23/2023 47
 Con…

• Draw and state the conclusion

• Hence, HO is rejected: the manufacturer's claim is not

upheld

01/23/2023 48
 Two sample mean and proportion

• It is possible to compute point and interval estimation for the

difference of two sample means.

– Let x1, x2, …, xn1 are samples from the first population and

– and y1, y2, …, yn2 be sample from the second population.

X
• Sample mean for the first population be
Y
• Sample mean for the second population

• Then the point

( X estimate
Y) for the difference of means (µ1-µ2) is
given by
01/23/2023 49
 Two sample mean estimation

Two independent samples

 Confidence interval estimation for known variance
 A (1-α)100% confidence interval for the difference of
means is given by:
2 2
 
(x  y)  z 1
 2

2 n1 n2
 The steps to test the hypothesis for difference of means
is the same with the single mean

01/23/2023 50
Example:
– In a large hospital for the treatment of mentally retarded, a
sample of 12 individuals with mongolism yielded a mean
serum uric acid value of 4.5mg/100ml
– In a general hospital a sample of 15 normal individuals of the
same age and sex were found to have a mean value of 3.4
– if it is reasonable to assume that the two populations of values
are normally distributed with variance equal to 1, do these
data provide sufficient evidence to indicate a difference in
mean serum uric acid levels between normal individuals and
individuals with mongolism?
01/23/2023 51
•  

01/23/2023 52
• -1.96 0 1.96
Rejection rejection region region

2.82

⇒ Z== 2.82
• Since 2.82>1.96 reject the null hypothesis, i.e., there is
an indication that the means are not equal.
01/23/2023 53
01/23/2023 54
Hypothesis testing about the difference between two
population proportions

• Ho: π1= π2
• HA: π1≠ π2
• We use a pooled sample estimate (p) for the common hypothesized
proportion since the null hypothesis states that π1= π2, which is a
weighted average of the sample proportions, with the sample size
as weights

P = n1p1 + n2p2

n 1 + n2
 P can be easily computed by summing all of the successes and
dividing by the total sample size
01/23/2023 55
01/23/2023 56
Example
• Two hundred patients suffering from a certain disease
were randomly divided into two equal groups. Of the
first group, 78 recovered within three days. Out of the
other 100, who were treated by a new method, 90
recovered within three days. The physician wished to
know whether the data provide sufficient evidence to
indicate that the new treatment is more effective than
the standard.
01/23/2023 57
 Con…
• H0: π1 = π2
• H1: π1 < π2
• P1 = 78/100 = 0.78, P2 = 90/100 = 0.90
• P =n1p1+ n2 p2
n1+ n2
• P = 100(0.78) +100(0.90) = 0.84
100+100

01/23/2023 58
01/23/2023 59
 Con…

• Since ‐2.32 < ‐1.645 it falls in the rejection region.

• We reject the null hypothesis. The data suggests that
the new treatment is more effective than the standard.

01/23/2023 60
 Tests of Significance

Data are often collected to answer specified questions, such as:

 Do children under five from Urban have a lower prevalence
of malnutrition compared with Rural children?
 Is a new treatment beneficial to those suffering from a certain
disease compared with the standard treatment?

01/23/2023 61
 Confidence interval or p – value

P‐Value
• Another vital concept related to significance is the P
value, commonly reported in scientific journals.
• The P value is related to a hypothesis test; it is the
probability of obtaining a result as extreme as (or more
extreme than) the one observed, if the null hypothesis is
true.
01/23/2023 62
• In other words, the P‐value is the probability of the
observed outcome, assuming that chance alone was
involved in creating the outcome, i.e., assuming the null
hypothesis is correct, what is the probability that we
would have seen the observed outcome.
• We can use our data to calculate the probability that our
finding is just due to chance, under the null hypothesis

01/23/2023 63
 Con…
• A P-value is the probability of getting the observed
difference, or more extreme, in the sample purely by
chance from a population where the true difference is
zero.
• If the p‐value is small, meaning the observed
outcome would have been unlikely, we will reject that
chance played the only role in the observed difference
between groups and
• Conclude, say in the example above that new
treatment does in fact have an effect on outcome
compared with the standard
01/23/2023
• How small is P-value? 64
 Con…
• The P value is calculated after the statistical test has been
performed; if the P value is less than α (0.05), the null
hypothesis is rejected (statistically significant)

• If the P-value is greater than 0.05 then, we conclude that

the observed difference could have occurred by chance
and there is no statistically significant evidence for a
difference between the groups in the population (at the 5%
level of significance)
01/23/2023 65
 Con..

• The practice of reporting values less than some

traditional value was established prior to the
availability of computers, when statistical tables were
the only source of probabilities
• Reporting the actual p-value communicates the
significance of the findings more precisely

01/23/2023 66
 95% confidence interval

• A 95% confidence interval gives a plausible range of values

that should contain the true population difference.
• If the 95% confidence interval includes the point of zero
difference then, by convention, any difference in the sample
cannot be generalized to the population.
• If 95%CI contains the null or hypothezed value, then it is not
statistically significant.

01/23/2023 67
 Con…
• Confidence interval (CI) is far more informative measure than P
value to evaluate the role of chance
1. Provide information that p-value gives
– If null value is included in a 95% CI, by definition the
corresponding P-value is > 0.05
2. Indicate the amount of variability (effect of sample size) by the
width of the CI
– This information can not be obtained from p-value
– Wide CI indicates greater variability and suggest inadequacy of

01/23/2023 the sample size 68

 Chi square
• A chi square (χ2 ) distribution is a probability distribution.
• The chi-square is useful in making statistical inferences
about categorical data in which the categories are two and
above .
• A Chi-Square (χ2) is a probability distribution used to make
statistical inferences about categorical data (proportions) in
which the numbers of categories are two or more

01/23/2023 69
 Con…
• Chi-Square test allows us to test for association between two
categorical variables.

• Ho: No association between the variables.

• HA: There is association

or

• H0: The two variables are independent

• HA : The two variables are dependent

• Consequently a significant p-value implies association

01/23/2023 70
 Con…

• Chi-Square test compares observed to expected

counts (frequencies) under the assumption of no
association (or Ho is true)
• With this method, data are arranged in the form of a
contingency table.

01/23/2023 71
• Definition: A statistic which measures the
discrepancy between K observed frequencies O1, O2
……Ok and the corresponding expected frequencies
e1, e2 …… ek.
Chi square = χ2 = Σ (Oi - ei)2 / ei

01/23/2023 72
01/23/2023 73
 Calculation of expected values

• The expected frequency in each cell is the product of

the row and column totals divided by the sum of all
the observed frequencies (i.e. the sample size)

01/23/2023 74
Characteristics
1. Every χ2 distribution extends indefinitely to the right from 0.
2. Every χ2 distribution has only one (right ) tail.

3. As degree of freedom increases, the χ2 curves get more bell

shaped and approach the normal curve in appearance (but
remember that a chi square curve starts at 0, not at - ∞ )
 The skewness diminishes as n increases

01/23/2023 75
01/23/2023 76
01/23/2023 77
 Con…

• The value of this statistic is small when the observed

and expected frequencies are close and
• It is large when they are not close, thus large values
of this test statistic indicate that null hypothesis may
be false

01/23/2023 78
Degree of freedom for χ2

• Chi-square test is based on the table of χ2 for

different degrees of freedom (df)

Chi-square degree of freedom is given by:

=(No. of rows-1) x (No. of columns-1)
= (R-1)(C-1)

01/23/2023 79
01/23/2023 80
• For 2x2 table, when the total no of observations is less
than 20 or when it is greater than 20 and the smallest of
the four expected frequencies is less than 5, use
Fisher’s Exact test.

01/23/2023 81
01/23/2023 82
01/23/2023 83
01/23/2023 84
01/23/2023 85
 Thank
01/23/2023
You! 86