As random sampling means that everyone in the population have an equal chance to be chosen, can it have inclusion criteria?
My research participants have been chosen based on the inclusion criteria (age 18-35, living in area X). At first, I had distributed my research among my friends and then they spread mine to their social circle. Also, I had posted my research in Twitter to reach the target participants who are out of my social circle and their participation was based on their willingness to volunteer. Can this be regarded as random?
Hi,
Sample selection is based on your research criteria i.e Inclusion and Exclusion.
Population are all those eligible units in a defined area.
Since sample is a sub set of population, you haif you are collecting from different sizesve to choose from this population according to the sampling frame.
There are different methods of sampling in random sampling and you have to follow one of these.(you can find in sampling methods book.
Population proportion to size depends on size of population
Please go through to choose a sample according to the method.missingness.
Keep these in mind.
Good luck
Procedure of data collection should be a true representaion of what you are aiming at. There could be falsifications,
Thank you so much Mr. @Bhogaraju Anand ! Finally I can clear my confusion now.
It's not a random sample. I would call that a convenience sample using inclusion criteria. But it's a quite legitimate recruitment technique for many research questions. You would, of course, have to describe the recruitment procedures for publication.
My point of view concerning of random sample is extremely controversial...
Random sample should consider the point of view of the subjects. Individuals should be able to chose the group in which they will be included. This a very ethical way to consider random sample...
Simple random sampling (SRS) is when every element of a population has an equal non-zero probability of selection (like names in a hat, or cards in a well-shuffled deck). There are many departures from SRS that yield probability-based samples that can be analyzed if we can adjust for the departures from SRS. The key feature of a probability-based sample is that the p of selection of any element can be calculated. If you had started with a list of everybody aged 18-35 in your Area X reference population, and sampled from that list randomly, you would have an SRS of that population.
Any sample in which you can't calculate the p of an element being sampled is a non-probability sample. It sounds like you used a form of 'snowball sampling' (here is a good site: https://www.questionpro.com/blog/snowball-sampling/)
supplemented by convenience sampling using social media.
So no I would not call your sample 'random'. Hope this helps.
So we have to choose between ethics and statistics... That’s a big deal!!!
Francoise - the views of the research participant are very important and there are even studies called 'preference trials' where people are asked if they have a preference for Treatment A or B. So they can refuse the study from the beginning, or agree then refuse if they don't get their preferred treatment.
But a lottery is a form of simple random sampling. What would happen if we allowed everyone who bought a ticket to decide if they wanted to be included in the winner or the loser group? :-)
It might seem obvious that no one should ever be 'forced' to participate in any research study, but sometimes there is no way of allowing people to opt out: in studies that look for traces of Covid virus (or fentanyl) in a city's sewage, everybody is included because there is no practical way to exclude any individuals (although I suppose they could leave town during the sampling periods).
Rhianon Allen Mark R Speechley Françoise Léon Thank you so much for the brilliant explanation! I appreciate them deeply.
Mark R Speechley yes, I had planned to use snowball sampling and even wrote it in my research proposal but my supervisor doesn't allow me to use snowball sampling and keep on telling to opt for random sampling instead. Thank you for clarifying it for me!
Fix the background of your target audience, qualification, age, gender, income level, type of work, geography area, etc.
Then you can distribute the questionnaire to the sample at random based on statistical adequacy.
- Badges
- Science topic
- Similar topics
- Psychology