I plan to study teachers' perception on M-Learning using a survey. How many teachers should be involved in the survey?
There are two basic approaches to determining samples size. The first involves setting limits on the amount of error in estimates of population values (such as the percentage who...). For this, you can search the internet for "sample size calculators."
The second approach is based on the power of your tests to detect significant results (e.g., the likelihood that a t-Test will be significant). For this, you should use the G*power program.
The appropriate sample size for a survey depends on various factors, such as the variability of the population, the desired level of precision, the level of confidence, and the available resources for conducting the survey.
In general, a sample size of at least 30 participants is considered sufficient for a survey, but this number may not be enough to ensure sufficient precision and generalizability of the results, especially if the population is large and diverse.
In the case of studying teachers' perception on M-Learning, a sample size of 100 to 200 participants is often recommended to ensure a representative sample of the population and sufficient precision in the results. This sample size is usually large enough to provide a stable and accurate estimate of the mean and standard deviation of the population.
It's also important to keep in mind that the sample size can be influenced by the response rate, which may be lower in a survey of teachers due to time constraints and other factors. To ensure a sufficient response rate, it may be necessary to use multiple methods of recruitment and follow-up with non-responders.
Mayur Wanjari What is your basis for saying "In general, a sample size of at least 30 participants is considered sufficient for a survey"? This is far too small to produce either an accurate estimate or a significant finding.
I apologize if my previous answer was not sufficient. In general, a sample size of 30 participants is considered to be the minimum threshold for obtaining an accurate estimate of the population parameters using statistical methods, such as mean, standard deviation, and correlation. However, this is a rough guideline and the actual sample size required depends on several factors, including the variability of the population, the level of precision desired, and the power of the statistical test being used.
In the case of hypothesis testing or estimating the effect size of an intervention, a sample size of 30 may not be enough to produce a significant finding or obtain a precise estimate of the population parameters. In these cases, larger sample sizes are recommended to increase the accuracy and power of the results. The appropriate sample size will also depend on the complexity of the research question, the variability of the outcome measures, and the desired level of precision.
The easiest way is to utilize an online sample size calculator, such as the one referenced below.
Momentive. (2023). Sample size calculator: Understanding sample sizes. SurveyMonkey. https://www.surveymonkey.com/mp/sample-size-calculator/
Good luck,
Typically, a larger sample size will provide more precise results, but also requires more resources and time to collect and analyze the data. A trade-off is made between the desired level of precision and the feasibility and cost of collecting a larger sample. Statisticians often use sample size calculators or consult with experts to determine the appropriate sample size for a study.
Zachary Farouk Chai
A sample size of 30 is also known as "the smallest large sample". That is the minimal sample size from which the sampling distribution is always normal regardless of the (shape of the) population distribution.
There are three main factors that affect the sample size:
1. Degree of heterogeneity in the population: if all population elements are perfectly similar (that is, perfect homogeneity), a sample size of 1 is enough. The more heterogeneity, the larger the required sample size. For proportions, maximal heterogeneity appears when p = q = 0.5
2. Level of confidence: usually 95% but it can be more or less
3. Margin of error: usually +/-3% but it depends from one variable to another.
For a so-called infinite population (in terms of the infinitely large number of possible samples of size n that can be drawn), requiring a 95% confidence interval, maximal heterogeneity and a 3% margin of error, the mathematically generated sample size is 1067.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey