This procedure provides a simple calculation to determine sample size.
1. Choose ProcessMA > Statistics > Power and Sample Size > Sample Size Calculator (Simple).
2. In Type of data, choose the type of data you have.
· If your data is discrete:
- In Proportion, enter the population proportion.
- In Error/Difference, enter the difference you want to detect.
· If your data is continuous:
- In Standard deviation, enter the population standard deviation.
- In Error/Difference, enter the difference you want to detect.
3. In Confidence level, enter the desired confidence level.
4. Click Calculate.
Proportion: Numeric; Between 0 to 1.
Error/Difference: Numeric; For discrete data, difference must be between 0 and 1.
Standard deviation: Numeric.
Confidence level: Numeric; Between 0 and 100.
[Case 1]
You know from historical data that customer satisfaction index has a mean of 65 and a standard deviation of 15. You need to determine the sample size for the new survey at a confidence level of 95%, to determine whether customer satisfaction has changed by 3 units in the satisfaction index.
1. Choose ProcessMA > Statistics > Power and Sample Size > Sample Size Calculator (Simple).
2. In Type of data, choose Continuous.
3. In Standard deviation, enter 15.
4. In Error/Differences, enter 3.
5. In Confidence level, enter 0.95.
6. Click Calculate.
Interpretation
You need 96 samples to detect a difference of ±3 with a 95% confidence.
[Case 2]
You are a training manager and you want to conduct a post training survey. The trainees will give you a response of Yes or No to whether they found the training useful to their work. Historically, 70% of your participants responded Yes. Determine the sample size for the new survey at a confidence level of 95% to determine if training usefulness has changed with a difference of 10 units.
1. Choose ProcessMA > Statistics > Power and Sample Size > Sample Size Calculator (Simple).
2. In Type of data, choose Discrete.
3. In Proportion, enter 0.7.
4. In Error/Differences, enter 0.1.
5. In Confidence level, enter 0.95.
6. Click Calculate.
Interpretation
You need 81 samples to detect a difference of ±10 units with a 95% confidence.