Sample Size Estimation 1
TWO WAYS EMPIRICAL ANALYTICAL
2
REQUIREMENTS 1 – THE PARAMETERS (PROPORTION OR AVERAGE) 2 – THE DEGREE OF PRECISSION (α) 3 – THE DESIRED CONFIDENCE LEVEL (Z) 4 – THE ESTIMATED DEGREE OF VARIABILITY OF OBSERVATION (STANDARD DEVIATION) 3
EXAMPLE 1 ESTIMATING THE MEAN Q1 WHAT PARAMETER DO YOU WANT TO STUDY ANSWER I WANT TO ESTIMATE THE AVERAGE INCREASE IN BODY WTS OF INFANT RATS GIVEN TREATMENT A WITH A CERTAIN PERIOD OF TIME 4
Q2 WHAT IS EXPECTED S.D. • • • • • • •
S.D. = 20 GM Q 3 HOW MUCH ERROR TOLERABLE ANSWER A : 10 GM OR LESS ANSWER B 5% OF AVERAGE Q 4 WHAT IS AVERAGE ANSWER MEAN = 100GM HOW CONFIDENT DO YOU WANT TO BE ? • 95% = {Z} 5
IN CASE OF ANS A TO Q3 N= (Z )2 (S)2 e2
= 1.96x1.96x20x20
= 16
10x10
6
IN CASE OF ANS B TO Q3 N= (Z )2 (S)2 e2 = 1.96 x 1.96 x 20 x 20 = 62 5X5
7
Sample size es
8
IN CASE OF prevalence as 20% N= (Z )2(pq) e2 = 1.96x1.96x20x80
= 246
5x5
9
IN CASE OF prevalence as 20% N= (Z )2(pq) e2 = 1.96x1.96x20x80
= 246 X25 =
1x1
10
Objectives At the end of session participant will be able to: Understand the role of standard deviation in estimating sample size in continuous type of data Understand the role of proportion in estimating sample size in discrete type of data Effect of confidence interval on the sample size
11
Tutorial A study has been planned to estimate the Factors effecting the prevalence of hypertension in a community. The prevalence of hypertension from some previous study in the same area was 10%. Assuming the standard confidence interval as well as level of significance, What sample size should be selected to have an appropriate confidence interval regarding proportion. 12