INTERPRETATION OF

STATISTICAL TESTS
GROUP 5

DR SANGEETA PATANKAR
DR SANDIPKUMAR SURYAWANSHI
DR SHITAL RASKAR
DR PALLAVI PATLE
DR VANDANA KAMBLE
DR RAJENDRA JANGLE
1
2
STATISTICS is the systematic collection and analysis of numerical
data in order to investigate or discover relationships among
phenomena so as to explain, predict, and control their occurrence.
3
4
 WHAT IS STATISTICAL ANALYSIS?
• STATISTICAL ANALYSIS is the science of collecting, exploring,
and presenting large amounts of data to discover underlying patterns
and trends.

5
 STEPS OF STATISTICAL ANALYSIS
3. Create a model to
2. Explore the relation summarize an
1. Describe the nature
of data to the understanding of how
to be analysed
underlying problem data relates to the
underlying population

5. Employ predictive
4. Prove or disprove
analytics that will
the validity of the
help guide future
model
actions
6
TYPES OF STATISTICS

7
 INTERPRETATION OF RESULTS OF
STATISTICAL ANALYSIS
• It relies on consideration of null hypothesis, P values, concept of
statistical vs clinical significance, study power, etc.
• Deciding which statistical test to employ depends on the type of data
being analysed and whether or not the data is normally distributed.
• Statistical tests are used in hypothesis testing. They can be used to
determine whether a predictive variable has a statistically significant
relationship with an outcome variable.

8
• If the value of the test statistics is more extreme than the statistic
calculated from null hypothesis, there is a statistically significant
relationship between predictor and outcome variables.
• If the value of the test statistics is less extreme than the statistic
calculated from null hypothesis, there is NO statistically significant
relationship between predictor and outcome variables.
• For a statistical test to be valid, the sample size needs to be large
enough to approximate the true distribution of the population to be
studied.

9
10
 CONFIDENCE INTERVALS
11 Dr Sandipkumar Suryawanshi
Asso.Professor, GAAC, Ahmedabad
‘POINT ESTIMATION’
 During Ante-Natal check-up = the weight of pregnant women are being
recorded to see, whether they are having the anticipated weight.
 mean anticipated weight during 32-36 weeks of pregnancy is 52 kg,
according to population true value.
 This sort of estimation of Mean, Median, S.D. etc. from the study sample,
for assessing its closeness to a specific true value part of the population is
called ‘Point Estimation’

12
INTERVAL ESTIMATES

The anticipated weight of a pregnant woman of 32-36 weeks of gestation is,

50-55 kg instead of a specific value (say 52 kg.) Then such estimates are
called ‘Interval Estimates’

13
CONFIDENCE INTERVALS
 Suppose pregnant woman has recorded the weight within the ‘anticipated interval’ in
9 visit out of 10 visits (90%), then the doctor is 90% confident that the woman has
attained the interval of the weight fixed (i.e. 50-55 kg). Such intervals of sample
estimates that are fixed with certain degree of confidence to attain the ‘true interval’
of the population ‘Confidence Intervals’.

14
 SUMMARY
 The estimation of Mean, Median, S.D. etc. from the study sample, for
assessing its closeness to a specific true value (parameter) point of the
population is called 'Point Estimation'.
 If the anticipated 'true value of the population is given in an interval, then
such estimates are called 'Interval Estimates'
 The estimates which are specifying the interval in which 95% of the
observations may lie is called 95% Confidence Interval (95% C.I) and the
estimate specifying the interval in which approximately 99%of the
observations may lie is called '99% confidence Interval (99% C.I.)
 From only one sample estimate, the 95% confidence interval can be
estimated with the formula: Mean +-2 SE (approximately) and 99%
confidence interval can be estimated with the formula: Mean +- 3 S.E
(approximately) 15
16
Level of Significance

17
Definition of Level of Significance
• The level of significance ( also called as type-I error) is the
measurement of the statistical significance. It defines the probability
of rejecting the null hypothesis in a statistical test when it is true.
• This quantity ranges from zero (0.0) to one (1.0) and is typically
denoted by Greek letter ‘alpha’ (α)

18
Objectives of Level of Significance
• Define statistical inference.
• Describe the reasoning of test of significance.
• Describe the part of test of significance.
• State hypothesis.
• Define p-value and statistical significance.
• Conduct and interpret a test of significance for the mean of normal
population.
• Determine the significance from a table. 19
Test of Statistical Significance

For Qualitative data

Chi-Square test
Z test
20
Use of Level of Significance in practice
• Popular level of significance are 10% (0.1), 5% (0.05), 1% (0.01), 0.5% (0.005) and 0.1%
(0.001)
• If the test of significance gives a p- value ( The probability that the observed data would
occur by chance in a given single null hypothesis) lower than or equal to α (Level of
Significance); the null hypothesis is rejected at that level.
• Such results are informally referred to as statistically significant.
• To improve the level of significance:
- Increase the power of hypothesis test.
- Use the larger sample.
- Use the higher level of significance.(α)
- Choose the larger value for difference.
- Use a directional hypothesis ( Also called one tailed hypothesis). 21
Importance of Level of Significance
• Level of significance is used for new pharmaceutical trials.
• To test the effectiveness of new medical products including drugs,
devices and vaccines.
• To study the pathology for effectiveness testing.
• To inform investors on how successful the company is at releasing new
products.

22
Probability
It is the chance that something will happen.
It is the likelihood of something happening in the
future.

Probability is the way of expressing knowledge of

belief that an event will occur.
It is expressed as a number between zero to one(0 to 1) .

23
Events in probability

Exhaustive events
Favourable events
Mutually exhaustive events
Equally likely events
Complementary events
Independent events

24
 Measurement of probability

1. classical approach
In this, we assume that trial results in any of many possible outcomes.
p(E)= No of outcomes favorable to event E
exhaustive no of outcomes

25
2. Empirical or statistical approach.
Here probability is obtained by performing experiment
large no of times. As no of trial increases , we get more
accurate results.

26
3. Axiomatic approach
It is reasonably true and are accepted without seeking any
proof.
If S is sample space in random experiment, a be an event,
then P(A) probability of A if, p(A)>0
P(s)=1

27
Three types of probability
1. Theoretical probability
For this we assume that all n possible outcomes of particular experiment are equally
likely and we assign probability to each possible outcome.

2. Relative frequency interpretation of probability

it is based on observation.
Probability of event p(A).
how many times A occurs
how many trials

3. Subjective probability
These are values assigned by individual on how likely they think events are to occur.

28
Power of test

29
power of test
Power is the probability that a
test of significance will pick
up on an effect that is
present. Power is the
probability that a test of
significance will detect a
deviation from the null
hypothesis, should such a
deviation exist. Power is the
probability of avoiding a Type
II error.

30
In summary…

31
• The standard error enables us to
specify the limits within which
the parameters of the population
are expected to lie with a
specified degree of confidence.
Such an interval is usually known
as confidence interval

32
α

• α : Level of significance
• 1-β : Power of the test

33
 Type of Data Type of Data

Goal Measurement (from normal Rank, Score, Measurement (from

population) non normal population)

Describe one group Mean, SD Median, Interquartile range

Compare one group to a One sample t- test Wilcoxon test

hypothetical value

Compare two unpaired groups Unpaired t-test Mann- Whitney test/ Wilcoxon rank
sum test

Compare three or more unmatched One -way ANOVA Kruskal- Wallis test
group

34
Thank you

35