The population where I did the survey is 100. Thus, based on Krejcie & Morgan, the sampling size shall be 80. I have received 82 responses but there are 8 errors. Thus, the usable responses are 74. Is this acceptable and how should I explain this in the thesis?
Thank you
Be careful about the definition of a population!
Are there really only 100 people in the population? And does your research only apply to these 100 people?
For example, there might be only 100 patients currently attending the day hospital, but your population is "people attending the day hospital" – so the current attendees are a sample of this wider population who have attended in the past and will attend in the future. For sure, past and future attendees may differ from the current attendees, but if you treat the current attendees as the population the assumption is that you are not interested in anyone who might attend in the future.
Does this apply in your case? There are very few 'closed' populations, and we are mostly interested in generalising to people who have not yet entered the population.
Very true what Ronan says
https://scholar.google.es/scholar?hl=es&as_sdt=0%2C5&q=How+do+we+calculate+margin+of+error+in+sampling%3F&btnG=
No matter what the population is, a small sample is likely to lead to relatively large standard errors, which can make it difficult to detect significant effects. Instead of relying on Krejcie & Morgan, I would look into approaches to sampling that are based on the "power" of your tests (i.e., your ability to detect an effect with your sample size). The most common tool for assessing power is G*Power.
Having said that, however, if you really do have data from 80 of the total population, I think most people will accept such a sample simply on the grounds of "face validity." The exception would be if there is some evidence of systematic bias that makes the non-respondents different from those who did supply data.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size