Scribd
Upload
22 views14 pages

T Test

Uploaded by

beyondtheboxwithbhakti
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
22 views14 pages

T Test

Uploaded by

beyondtheboxwithbhakti
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 14

Parametric Vs Non

parametric tests
Parametric tests Non parametric tests

When the data lies on interval and ratio scale When the data lies on nominal and ordinal scale

Population is normally disitributed Population is not normally distributed / no assumption can be


made about population distribution

Homogeneity of population variance is required Homogeneity of population variance is not required

For eg., t test, ANOVA, Pearson correlation, etc. For eg., Chi square, Kruskal wallis test, Mann Whitney U test,
etc.
t test
• Appropriate for testing whether means of the two groups are significantly
different or not - one IV with two levels
• Two types: Independent t test and matched/paired t test/ dependent t test
• Independant t test is used when the two groups being compared are
independent of each others - random groups design
- for eg., comparing control group vs experimental group
• Paired t test is used when the two groups being compared are matched on
some basis - repeated measures design/matched design
- for eg., comapring pre test vs post test scores
Steps for calculating t test
I. Calculate the means of the two groups
II. Calculate the standard error of the difference between the two means
III.Find the calculated value of t
IV. Calculate df
V. Find the critical value of t
VI. Determine whether the null hypothesis should be rejected by
comparing the calculated value of t to the critical value of t
- when calculated t value is greater than critical t value, you reject the null
hypothesis
1) Calculated t value : 3.20
Critical t value: 2.10

2) Calculated t value: 1.99


Critical value : 2.10

t : 2.8 , p= 0.02
0.05
A B

-2.10 2.10
Independent samples t test

• When N1=N2
- df is N1 + N2 - 2
(2N-2)

• When N1#N2
- df is N1 + N2 - 2
(N1 -1 )+(N2-1)
N1+N2-2=20+20-2=38
N1=20
N2=20
20+15-2=33
Degrees of freedom
• Number of values that are free
to vary in a sample once we have
computed some statistics
- given that you know mean of the
set of N scores, the number of
scores that are free to vary in a
sample is one less than N, that is
N-1
• It is related to the sample size
1.3
2.2
3.4
4.12
5.4
Total of the scores = 25
Mean = 25/5=5
Dependent samples t test
• Number of participants in each group must be equal
Pre test Post D D2
(X1) test
(X2)
11 10 1 1
17 15 2 4
14 14 0 0
10 8 2 4
11 9
15 14
10 6
8 7
12 11
15 13
X1 = X2 =
N=10

s 12 s 22
t distribution
• If I take out multiple samples from a
population and calculate t value for each
sample, such a distribution will be called as
t distribution
• It is similar to normal distribution but its
fatter
• As the sample size increases it approaches
normal distribution
• Usually t distribution at df=30 isquite close
and similar to normal distribution. Hence
N=30 is considered to be a good number of
observations for t test
• The mean of t distribution is 0 and variance
is k/(k-2)

You might also like