What is your opinion on this?
CASE: If you need to include 100 people according to the power calculation and you expect 20% dropout, you need to include a total of 125 people.
QUESTION: Would you stop before 125 (100+20%) if you reach 100 participants that can be included in the analysis due to lower dropout than expected? Or would you continue to 125, i.e. include more than the power calculation?
Not in all studies do we get 100% response. literature has documented that any response rate above 70% is adequate enough for analysis. In your case, you expect response rate of 80% after droup out. In my opinion this is enough for analysis, hence you don't need to add. Besides, that can translate to about 80 study participants which again is enough sample size to give the desired parameter in your study population. According to Roscoe (1975), rule of thumb, any sample size between 30 and 500 is considered adequate for most of the studies. My friend, argue that your study belongs to the 'most'.
Considering a sample size of 125 and greater than that is good provided you have enough time, sources,logistics and good frequency of participants.
1) It helps if you are having subgroups such as control group, treatment group, and placebo group where equal proportion of participants are recommended else there is a risk of bias.
2) If research is based on repetitions /intervention with the participants (clinical trial), you have to be 100% sure that the same participant will come and meet you after the prescribed time which is always a risk.
3) It also helps if research outcome is based on factors such as age , medical conditions, and other demographics.
4) You may also want to consider the effect of the study on the sample size i.e., outcome may vary / greater /lesser than expected.
Having a higher sample size always keeps you on a safer side.