Pallavi - 

Did you actually assign a number to each member of the population, and select the sample at random?  Regardless, if you say you have a random selection, it actually has to be random. 

You mentioned different levels of experience.  Perhaps you should have had a stratified random sample.  You might still research the term "poststratification." 

Note: If you have another study later, this one might be used to help determine an estimate for sample size needs. 

Cheers - Jim 

You are almost correct and Jim hinted at the problem. A simple random sample is not representative but a stratified sample is.  See the links below for details.  Let me know if you have questions. This is the only thing that I believe that you need to change if you really want your sample to be representative. Best wishes, David

https://en.wikipedia.org/wiki/Sampling_%28statistics%29

https://www.amazon.com/Elementary-Survey-Sampling-5th-Fifth/dp/B001NURIMO/ref=sr_1_1?ie=UTF8&qid=1491210937&sr=8-1&keywords=elementary+survey+sampling+by+richard+l.+sheaffer

James and David seem to have answered the question. However, we come from a diversity of backgrounds and we do not all use exactly the same language or textbook. If there is still a question, maybe you could write out the definitions as you would like to use them.

 Hello Pallavi,

First what should be clear, which is the research question and the scope of the result.

If the research question addresses something QUALITATIVE, then the sampling is NOT PROBABILISTIC, (theoretical sampling, by quotas, case selection, intentional or directed sampling). What he seeks, is not the generalization of his results, but the richness of the discourse in the investigated (meaning, relations between concepts or ideas according to the personal construction of the subject).

If the question is quantitative, then it must be framed within some Quantitative sampling paradigm, since what it is going to show is possible to take it to populations with the same essential characteristics.

And here things are complicated, since it must consider that if everyone has the same probability of responding the same (same probability of being selected), then the question about the sampling is that characteristic makes the person can have a different answer . Then the stratified sample or by clusters appears, and also the units of investigation are assembled (sampling in two, three stages or more stages).

According to your theme, it is possible to think about the business item (a restaurant may have a different answer to a fashion store),

According to what you say Your justification, is actually a feature to consider to see, how to sample or classify to sample

Remember that there is sampling for infinite and finite populations

Allthough you could get statistics from a company (which you may consider a case), but you should be clear that the scope of the results are only for that reality.

read book Foundations of Behavioral Research (author Kerllinger)

https://books.google.cl/books?id=3QQQAQAAIAAJ&q=Foundations+of+Behavioral+Research&dq=Foundations+of+Behavioral+Research&hl=es&sa=X&redir_esc=y

There has been good attempts by earlier scholars. I just want to add that both stratified and random sampling techniques are both probability based hence, prevents bias. By random sampling, there should be a complete listing of the population from which the sample is to be drawn. It is from that sampling frame that the sample will now be randomly selected. On that basis, you should select 450 respondents from a known population size.

I hope this clarifies.