## Adjusting Sample Size

There are many situations where the sample sizes estimated using the prescribed formula will not hold good for some practical reasons. In such case, an adjustment in the sample size will become necessary. Also when we depart from Simple Random Sampling (SRS), the sampling error increases. This happens very much in the case of cluster sampling. Instead of drawing sampling units at equal intervals, we draw a group of sampling units at equal cluster intervals.

In order to compensate for the increase in sampling error, a design effect has to be calculated and used as a factor for enlarging the sample size primarily estimated using the recommended formula. There are also other scenarios where adjustment in the sample size is needed.

In order to compensate for the increase in sampling error, a design effect has to be calculated and used as a factor for enlarging the sample size primarily estimated using the recommended formula. There are also other scenarios where adjustment in the sample size is needed.

**For Unequal Groups**

**For Drop-Outs**

**For Cluster Effect**

**Finite Population Correction**

When you have to draw a sample from a finite small population, it may either not be possible, or not be practical to draw the estimated number of sampling units. For example, if the relevant formula results in a sample size of 1800 children while there are only 1500 children in the community where the survey is to be done, you can use the below formula to adjust the sample size which will make more sense