Hypothesis Testing

Delivered by​
Dr. Tushar Shaw
Assistant Professor
Dept. of Allied Health Sciences
Faculty of Life and Allied Health Sciences ​
2
Set up a Hypothesis
• Conventional approach – Set two different hypothesis
• So constructed if one is accepted other is rejected
1. Null Hypothesis (H0)
2. Alternative Hypothesis (Ha)
Example: A psychologist wishes to test difference in mean IQ of two
groups then he will establish a null hypothesis is that two groups have
equal means and alternative hypothesis is that their means are not
equal.

3
Types of error
Null hypothesis is
Type I error: Rejecting H0 when H0 is true True False
[false positive]. Decision Fail to reject Correct Type II error
about null inference (false
Type II error: Accepting H0 when H0 is hypothesis (true negative)
negative) (probability =
false II error [false negative] Probability= β)
1-α
Reject Type I error Correct
(false positive) inference
• Clearly, the consequences of type II error (probability =
α
(true positive)
(probability =
are more serious than Type I error. ) 1-
β)
• So, in any testing, we must decide the
maximum amount of Type I error which
we are prepared to risk, which is called
as the level of significance
4
Set up a suitable significance level

5
Suggested Statistical Test based on
Data Type
Type of Research Question Numerical Data Categorical Data
Parametric Non Parametric
Is there a difference between Unpaired t Mann Whitney Chi-Square
2 unpaired groups? test U test test
Wilcoxon ranksum Fisher exact
test Test

Is there a difference between Annova Kruskal Wallis Chi-Square

> 2 unpaired groups?
Is there a difference Paired t test Wilcoxon pairs Mc Nemars test
between 2 paired groups? signed-rank test
Is there an association Pearson’s Spearman’s rank Relative risk
between variables? correlation correlation Odds ratio
coefficient Chi square test for
test (r) Logistic regression
Survival Analysis Kaplan Mayer curves and Logrank
test
6
7
Power of a Hypothesis Test
• The power of a hypothesis test is defined is the probability that the
test will reject the null hypothesis when the treatment does have an
effect.
• The power of a test depends on a variety of factors including the size
of the treatment effect and the size of the sample.

8
The p-value
• Is the probability of obtaining a specific observed result by chance,
assuming that the null hypothesis is true.
• If the calculated p-value of your data is 0.47 it means the probability
of the null hypothesis being true is 47%, which is very high, so you
cannot reject the null hypothesis.
• If the calculated p-value of your data is 0.03 (as in Table 1), it can be
stated that the probability of the null hypothesis being true is only
3%, and since we have fixed the level of significance at 5%, we can
safely reject the null hypothesis and hence accept the alternative

9
The Limitations of p-Values
• Whether intentional or not, there is a tendency for p-values to devolve into a conclusion of "significant"
or "not significant" based on whether the p-value is less than or equal to 0.05. This can be very
misleading.
• P-values depend upon both the magnitude of association and the precision of the estimate (the sample
size). If the magnitude of effect is small and clinically unimportant, the p-value can be "significant" if the
sample size is large. Conversely, an effect can be large, but fail to meet the p<0.05 criterion if the sample
size is small.
• There is a temptation to embark on "fishing expeditions" in which investigators test many possible
associations. When many possible associations are examined using a criterion of p< 0.05, the probability
of finding at least one that meets the critical point increases in proportion to the number of associations
that are tested.
• Many investigators inappropriately believe that the p-value represents the probability that the null
hypothesis is true. However, p-values are computed based on the assumption that the null hypothesis is
true. The p-value is the probability that the data could deviate from the null hypothesis as much as they
did or more. Consequently, the p-value measures the compatibility of the data with the null hypothesis,
not the probability that the null hypothesis is correct.
• Statistical significance does not take into account the evaluation of bias and confounding.

10
Disclaimer
• All data and content provided in this presentation are
taken from the reference books, internet – websites
and links, for informational purposes only.

11