What is Pre-requisites for calculation of sample size?
In medical research, it is very important to enroll sufficient number of samples in a study. For a study samples should not be smaller or larger. Results become invalid if study done on smaller size than required. Un-necessary large sample size should be avoided. So sample size should be calculated using appropriate formula. For calculation of sample size following Pre-requisites is essential.
Pre-requisites for calculation of sample size:
1. Study Type:
First requirement for calculation of sample size is study type. Study type me be observational study, Clinical Trial, Case Control Study & Cohort Study etc.
2. Information on Variable:
Information on outcome variable is required i.e. Proportions, Mean & Standard deviation, Proportion of exposed and unexposed cases etc.
3. α error, and Confidence Level :
α error (Alpha error ): α is Significance level of a test. It is probability of rejecting a true null hypothesis known as “Type-I Error”. Commonly accepted level of alpha error is: 0.05 or 0.10 = 5% or 10%.
Confidence Level:
The probability of correctly accepting Null Hypothesis.
Denoted by 1- α or (100 x 1- α)
When α is decided, Confidence Level is automatically fixed.
4. ß error, and Power of Test:
Beta (ß) :
The probability of rejecting a true alternate hypothesis. It is also called “Type-II Error”. Commonly accepted levels of Beta error are: 0.05, 0.1 & 0.2. In terms of % : 5%, 10% & 20%
Power of a Test:
The probability correctly accepting an alternate hypothesis.
It is denoted by 1- ß . In terms of % it is 100 x (1- ß).
When ß is selected, Power of Test is automatically fixed.
5. Z values:
z 1- α /2 and z z 1- α are the functions of the confidence level,
While z 1- ß is the function of the power of the test.
Two sided test: z 1- α /2 = 1.96 (95 % Confidence)
One Sided test: z 1- α = 1.65 (95 % Confidence)
Two sided test: z 1- α/2 = 1.65 (90 % Confidence)
One sided test: z 1- α = 1.28 (90 % Confidence)
z 1- ß = 1.28 (90 % Power )
z 1- ß = 0.84 (80 % Power)
6. Precision:
Measure of how close are the estimates to the parameter. Expressed in absolute terms or relative points.
