Regardless if the study retrospective or others, you need to calculate the needed sample and to have the minimum.

To do power analysis to estimate your sample size, you have to write your hypothesis, and based on that you decide what statistical test you will use. It should be one of the inferential statistics. so you need to determine the following: alpha {standard to be .05}, power [standard to be .80], effect size {small, moderate, or large, each test has its own value, you can find these values in the net}. Then download free programs to calculate the sample size such as G. power.

Hi,

Your question: Sample size in retrospective study

Sample size determination is a very nice and complicated procedure.  It is very interesting one also.  It involves many assumtions and considerations.  Sample size would be affected by many factors like study period, whether the researcher is a student or part - time researcher or full time researcher or prevalence (in %) and cases in a particular clinic or hospital or other organisations.

For retrospective study, you should not calculate the sample for your particular study.  But, you should mentioned the study period like in months or in years.  You have to take all the cases or items or subjects for your study and analyze it with suitable statistical analysis with clear knowledge and interest.  If you want to do research based on records (ie., retrospective data - secondary data) first you make a protocol for your study.  In that, you would mentioned the particulars of your study design, sample size, study variables, study period, study participants, study method, study tool, what are all the statistical analysis would be used and significance level (normally 5%).

Kindly see the following link, for your further clarifications:

http://www.jstor.org/discover/10.2307/30144196?sid=21105470048033&uid=2129&uid=2&uid=70&uid=4&uid=3738952

http://www.sjsu.edu/faculty/gerstman/StatPrimer/sampsize.pdf

http://www.uv.es/uvetica/files/McCrum_Gardner2010.pdf

Also, for reference, a good tool to calculate sample size is: http://www.statisticalsolutions.net/pss_calc.php   

For more information about retrospective designs, a good lesson can be found at: https://onlinecourses.science.psu.edu/stat507/node/54

The sample size calculator that Rhonda recommends assumes that you understand the concept of "statistical power." This is a central idea in selecting sample size, so if you are not familiar with "power," then you should read about that first.

I looked on the web and this seems to be a good source for beginners:

http://meera.snre.umich.edu/plan-an-evaluation/related-topics/power-analysis-statistical-significance-effect-size

You can use PS  (power and sample size calculation) a computer software application for power and sample size calculation developed by Dupont & Plummer (1990). It can be applied to retrospective and prospective cohort study, case-control study and other types of studies. It can be accessed through the following website: http://biostat.mc.vanderbilt.edu/wiki/Main/PowerSampleSize. 

thank you very much. So, in proposal in need to mention just the period of study and i do not need to calculate the sample size(in proposal)?

Regardless if the study retrospective or others, you need to calculate the needed sample and to have the minimum.

To do power analysis to estimate your sample size, you have to write your hypothesis, and based on that you decide what statistical test you will use. It should be one of the inferential statistics. so you need to determine the following: alpha {standard to be .05}, power [standard to be .80], effect size {small, moderate, or large, each test has its own value, you can find these values in the net}. Then download free programs to calculate the sample size such as G. power.

You can get formulae for sample size and power calculations for different study designs including prospective studies at the following web links:

1. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3775042/

Thank you,

Donath

It is also important to note that different studies are more power sensitive than the other. Therefore, the default 80% power level might not be rigid.

I found good discussion here. Solved my few queries.

Nice discussion. I had a similar challenge but found this discussion quite helpful!

use this web site, its helpful

http://www.statisticalsolutions.net/pss_calc.php

What is the preferred formula between using online calculators and sample size calculating formulae in research