Is My Sample Large Enough
Is My Sample Large Enough
Often, the voice of the person asking the following question is frantic. “Is the sample large enough to meet the goals of my study?” This is an age-old question, but relatively easy to answer if you know how to approach the question and have the right tools.
However, there may not be just one correct answer. The sample size required for a study depends on several variables. Some of these variables relate to your knowledge of the population being studied. Mostly, prior knowledge is hard to get, unless you are doing a kind of longitudinal study. Estimation is required.
As mentioned earlier, the answer to the sample size question is not necessarily a number. Obviously, there is a total sample size for the study, but it is likely that it is the sum of two or more parts. For example, suppose your study divides a priori into segments that you want to compare, and each segment or shift must be large enough.
In addition, the sample size depends on your need to detect changes of a certain magnitude. For example, suppose that you are satisfied that changes to a certain extent of 5% or more are real differences (ie, the probability is that the difference is not due to a measurement error). You also indicate that a confidence interval of 90% suits you (ie, you are willing to take the risk of being 10% of the time or 1 out of 10 times wrong), as your sample size requirements are likely to be relatively low (approx ) N = 270).
However, if you want to find differences of 3% or more and your risk tolerance is such that you do not want to be more than 1:20 (or 95% confidence) wrong, the sample will be much larger (about N = 1015).
An error rate of 5% is good in situations where there is a large difference in the proportions between groups (eg a group of respondents behave in 20% of cases and in 70% of cases in behavior). However, if the behavior or perception of your group of respondents is closer, in 45% of the cases one group may disagree with the positions of one political candidate and another group may need more precise measurements in 55% of the cases.
In summary, to obtain a lower error rate and / or a higher probability of correctness (or in other words, a lower probability of incorrectness), larger samples are required.
To get the best answers to your sample size questions, use performance statistics calculations. The performance statistics formulas are in several textbooks. However, you do not have to learn mathematics, but you can simply use one of the many readily available calculators available online. That being said, it would probably be to your advantage to learn a little about the various possibilities of performance statistics. There are calculations for proportions and others for mean tests. Just type one of the following queries in a good search engine to find one you like:
Sample size calculator, sample size calculations, sample size determination, sample size formula, performance, or sample size estimate.
If you need more information on sampling or help in any other area of market research, contact the author. Alternatively, visit the Ask-the-Expect page on our website, where you can access the knowledge of our entire community of experts.
Carey Azzara is a leading researcher with over 20 years of experience in the research community. He has two advanced degrees in market research disciplines. Director and Founder of AtHeath, LLC Mr. Azzara is a consultant, author and a highly respected researcher. Two of the key features of AtHeath are the Market Research Resource Center (MRRC) and the expert community that supports the MRRC.
The company name “AtHeath or At-the-Heath” is the point where forest and grassland meet – a metaphorical point of transformation or transition.