Scribd
Upload
9 views

Lec 7 Week 1

The document describes a study that tested the effect of a drug on diabetic rats. It reports the mean and standard deviation of untreated (XതU=326.2, sU=15.81) and treated (XതT=287, sT=9.74) rat populations. It performs a t-test to compare the populations, calculating the t-statistic (4.7178) and p-value (0.0015). Since the p-value is less than the significance level of 0.05, the null hypothesis that the populations are the same is rejected, indicating the treated and untreated populations are statistically different.

Uploaded by

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

Lec 7 Week 1

The document describes a study that tested the effect of a drug on diabetic rats. It reports the mean and standard deviation of untreated (XതU=326.2, sU=15.81) and treated (XതT=287, sT=9.74) rat populations. It performs a t-test to compare the populations, calculating the t-statistic (4.7178) and p-value (0.0015). Since the p-value is less than the significance level of 0.05, the null hypothesis that the populations are the same is rejected, indicating the treated and untreated populations are statistically different.

Uploaded by

A A
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 28

Concepts of Statistical Tests

Effect of a drug on diabetic rat model

𝑋ത𝑈 = 326.2 𝑋ത𝑇 = 287


𝑠𝑈 = 15.81 𝑠𝑇 = 9.74
Effect of a drug on diabetic rat model

𝑋ത𝑈 = 326.2 𝑋ത𝑇 = 287


𝑠𝑈 = 15.81 𝑠𝑇 = 9.74

Are the treated and untreated


populations different?
Effect of a drug on diabetic rat model

𝑋ത𝑈 = 326.2 𝑋ത𝑇 = 287


𝑠𝑈 = 15.81 𝑠𝑇 = 9.74

If two population distributions are same, then what


is the probability that the difference between the
sample means be greater-equal to observed
difference of sample means?
Testing alternate hypotheses

Null hypothesis, Ho

Two population distributions are same →


𝜇𝑈 = 𝜇 𝑇

Alternate or Experimental hypothesis, Ha

Two population distributions are different →


𝜇𝑈 ≠ 𝜇 𝑇
Testing alternate hypotheses

Null hypothesis, Ho

Two population distributions are same →


𝜇𝑈 = 𝜇 𝑇

Alternate or Experimental hypothesis, Ha

Two population distributions are different →


𝜇𝑈 ≠ 𝜇 𝑇
Calculate the probability,

𝑃 Difference in sample means ≥ 𝑋ത𝑈 − 𝑋ത𝑇 |Ho = 𝑝


Calculate the probability,

𝑃 Difference in sample means ≥ 𝑋ത𝑈 − 𝑋ത𝑇 |Ho = 𝑝


t-distribution
t-distribution
Effect of a drug on diabetic rat model

𝑋ത𝑈 = 326.2 𝑋ത𝑇 = 287


𝑠𝑈 = 15.81 𝑠𝑇 = 9.74

Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇
Alternate hypothesis, Ha : 𝜇𝑈 ≠ 𝜇 𝑇
Effect of a drug on diabetic rat model

Test statistics:

Sample statistics − Population parameter


𝑡=
Standard Error of sample statistics

Sample statistics: 𝑋ത𝑈 − 𝑋ത𝑇


Population parameter: 𝜇𝑈 − 𝜇 𝑇
𝑋ത𝑈 = 326.2 𝑠𝑈 = 15.81 𝑋ത𝑇 = 287 𝑠𝑇 = 9.74
𝑛𝑈 = 𝑛 𝑇 = 5
Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇
Sample statistics: 𝑋ത𝑈 − 𝑋ത𝑇
Population parameter: 𝜇𝑈 − 𝜇 𝑇

Sample statistics − Population parameter


𝑡=
Standard Error of sample statistics
𝑋ത𝑈 = 326.2 𝑠𝑈 = 15.81 𝑋ത𝑇 = 287 𝑠𝑇 = 9.74
𝑛𝑈 = 𝑛 𝑇 = 5
Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇
Sample statistics: 𝑋ത𝑈 − 𝑋ത𝑇
Population parameter: 𝜇𝑈 − 𝜇 𝑇

Sample statistics − Population parameter


𝑡=
Standard Error of sample statistics

𝑋ത𝑈 − 𝑋ത𝑇
𝑡= 𝐷𝐹 = 𝑛𝑈 + 𝑛 𝑇 − 2
1 1 2
+ 𝑠
𝑛𝑈 𝑛 𝑇

2 2
𝑛𝑈 − 1 𝑠𝑈 + 𝑛 𝑇 − 1 𝑠 𝑇
𝑠2 =
𝑛𝑈 + 𝑛 𝑇 − 2
𝑋ത𝑈 = 326.2 𝑠𝑈 = 15.81 𝑋ത𝑇 = 287 𝑠𝑇 = 9.74
𝑛𝑈 = 𝑛 𝑇 = 5
Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇

𝑋ത𝑈 − 𝑋ത𝑇
𝑡= 𝐷𝐹 = 𝑛𝑈 + 𝑛 𝑇 − 2
1 1 2
+ 𝑠 =8
𝑛𝑈 𝑛 𝑇

= 4.7178
𝑋ത𝑈 = 326.2 𝑠𝑈 = 15.81 𝑋ത𝑇 = 287 𝑠𝑇 = 9.74
𝑛𝑈 = 𝑛 𝑇 = 5
Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇
Alternate hypothesis, Ha : 𝜇𝑈 ≠ 𝜇 𝑇
t = 4.7178 DF = 8

DF = 8
𝑋ത𝑈 = 326.2 𝑠𝑈 = 15.81 𝑋ത𝑇 = 287 𝑠𝑇 = 9.74
𝑛𝑈 = 𝑛 𝑇 = 5
Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇
Alternate hypothesis, Ha : 𝜇𝑈 ≠ 𝜇 𝑇
t = 4.7178 DF = 8

p = 2 × 𝑃(𝑡 > |4.7178|) = 0.0015

DF = 8
𝑋ത𝑈 = 326.2 𝑠𝑈 = 15.81 𝑋ത𝑇 = 287 𝑠𝑇 = 9.74
𝑛𝑈 = 𝑛 𝑇 = 5

Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇

Alternate hypothesis, Ha : 𝜇𝑈 ≠ 𝜇 𝑇

t = 4.7178 DF = 8

p = 2 × 𝑃(𝑡 > |4.7178|) = 0.0015


Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇

Alternate hypothesis, Ha : 𝜇𝑈 ≠ 𝜇 𝑇

p = 2 × 𝑃(𝑡 > |4.7178|) = 0.0015

Level of significance 𝛼 = 0.05


Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇

Alternate hypothesis, Ha : 𝜇𝑈 ≠ 𝜇 𝑇

p = 2 × 𝑃(𝑡 > |4.7178|) = 0.0015

Level of significance 𝛼 = 0.05

As, 𝑝 < 𝛼

Reject Ho
Null hypothesis, Ho : 𝜇𝑈 = 𝜇 𝑇

Alternate hypothesis, Ha : 𝜇𝑈 ≠ 𝜇 𝑇

p = 2 × 𝑃(𝑡 > |4.7178|) = 0.0015

Level of significance 𝛼 = 0.05

As, 𝑝 < 𝛼

Reject Ho

The treated and untreated


populations are different.
Testing hypotheses

Null hypothesis, Ho

Alternate or Experimental hypothesis, Ha

Define a relevant test statistics

Calculate the probability,


𝑃 Test Statistics ≥ Observed test statistics |Ho = 𝑝

Compare p with respect to Significance Level 

Reject or accept H0
Parametric Tests for
Hypothesis testing
t-test
Z-test
ANOVA

Non-parametric Tests for


Hypothesis testing

Chi-square
Mann-Whitney U-test
Kruskal-Wallis H-test
‘Statistically Significant’ ≠ ‘Significant’

‘Statistically Significant’ does not mean


that the observed changes/differences
have Important/Relevant/Big
biological/clinical implications

‘Statistical Significance’ is all about


rejecting or accepting null hypothesis.

‘Statistical Significance’ depends upon


sample size.
Key Points
Purpose of hypothesis testing: does the
observation made in sample hold true in
population
Key Points

Null hypothesis, Ho
Alternate or Experimental hypothesis, Ha
𝑃 Test Statistics ≥ Observed test statistics |Ho = 𝑝
Compare p with respect to Significance Level 
to reject or accept H0
Key Points

t-test: to compare means, to test significance of


parameters in a model
Key Points

‘Statistically Significant’ does not mean that the


observed changes/differences have biological/clinical
implications
Number of control samples, n1 = 5
Number of +Inhibitor samples, n2 = 5
Degree of freedom = n1 + n2 − 2
t-value = 3.855

Is the difference between these two


groups statistically significant at the level
of 0.01?

Online t-distribution calculator:


https://homepage.divms.uiowa.edu/~mbognar/applets/t.html

You might also like