Hello , both links may help you until an expert adding an answer :

http://www.openepi.com/SampleSize/SSCohort.htm

or

https://www.cdc.gov/epiinfo/support/downloads.html

----------------------------

You should know the prevalence of the point of your topic , Odds Ratio , Design effect

For example in Comparative cross-sectional study design :-

Sample Size :

The required sample size was calculated using the EPI info statistical package program Version 7.2.01 (CDC, Atlanta, GA, U.S.A). using two population proportion formula for this cross-sectional study design and the following assumptions: Confidence level=95%; Power (1-β): 80%; Design effect=2; Ratio=1:1; Odds ratio=2; P1=........ % (using the lowest acceptable prevalence of .............. reported in previous studies .......... Considering 10% possible non-response rate, ...... (non-exposed) and ..... (exposed) will be included.

Sampling procedures:

By using multi stage sampling techniques ( on 2 stages )

so Design effect=2

Dear Richa,

Once i noted these sampling techniques during my research project on patients of gall stones:-

1. Simple Random Sampling -Select patients at random from a complete list of all who had surgery or used a particular facility

• Advantage- Easy to do for small groups • Disadvantage- Labor intensive and small samples may be biased or not representative to the patient population

2. Systemic Sampling – Select every certain number on a list of patients who had surgery

• Advantage-Easy to select every 3rd or 5th patient • Disadvantage-Recurring pattern in potential patient group can introduce bias

3. Stratified Random Sampling -Divide the patient population into segments such as: Age, Gender, Health Issue, and then select a random sample from each of those segments

• Advantage-Sample is more likely to be truly random of your patient population than the “simple random sampling’ • Disadvantage- Can be labor intensive to calculate your patient population and sample size for each segment

4. Cluster Sampling – Random sample of clusters of patients. Data is collected from the entire cluster. Example: “All patients had surgery in the surgical center on Lime Street”

• Advantages-Convenient and in-expensive • Disadvantages-Clusters have variables and introduce bias

Furthermore, the same question was addressed on researchgate under this URL:-

https://www.researchgate.net/post/How_to_calculate_the_sample_size_in_a_hospital_based_stud

There are many sample size calculators available on the internet, and here is a very simple one:

http://www.raosoft.com/samplesize.html

It assumes that you are trying to estimate the percentage of the population that answers a yes-no question, and it begins with the assumption that the population is split 50-50 on this question. It also assumes that you want to have a 95% degree certainty about a margin of error of 5%. All of these are the typical default assumptions in calculating sample size.

Hello, Thanks Mr. Khaled. The resources are very helpful.I am unable to understand the meaning of power. Could you please reflect on it?

Power of the study means = 1 - beta = minimum 80 % to be good. It depends on Null Hypothesis and alternative hypothesis.

Power of study = 1 – β

Power must not < 80

So β value not > 20 ,, accepted as type II error

Power is the ability to detect an effect. So, if the effect (say, a significant t-Test) actually does exist in the population, power reflects whether a given sample size will show that result.

Thanks David and Khaled