Dear colleagues, I am looking for advice on the validation of a standard questionnaire that I intend to translate. The original version contains 40 items. Could you please tell me what sample size is required for validation? Thank you in advance for your help.
IYH Dear Diana Elena Lazar
IMHO minimum sample size of 200-300 participants; better 350 to 400 complete responses. Why? Validating a translated questionnaire demands careful deliberation wrt adequate sample size bc of competing priorities:
1. Power Requirements: Securing sufficient statistical power necessitates recruiting enough respondents to detect genuine differences and correlations among variables.
2. Precision Objectives: Attempting to achieve acceptable margins of error and confidence intervals guides calculations for requisite sample counts.
3. Domain Coverage Expectations: Representing various sub-populations and avoiding unduly skewed responses encourages gathering sizable samples.
Other factors come into play too wrt sample size:
* Item-to-respondent ratio: I would aim for five to ten times as many completed surveys as items. In your case, that amounts to somewhere between 200 and 400 replies.
* Anticipated Missing Values: Plan for item omissions, withdrawal, or dropouts by padding initial quotas accordingly.
* Population Homogeneity: Target relatively homogeneous participant pools when seeking fine-grained distinctions or specific effects, o/w cast wider nets to capture diverse segments.
More advanced considerations:
* Construct Complexity: Take stock of anticipated construct loadings, factor complexity, and intended confirmatory factor analysis models. Higher loads call for beefier samples to avoid underpowered analyses.
* Multiplicity: Contemplate multiple test correction techniques, such as Bonferroni or False Discovery Rate adjustments, to manage Type I errors.
Summa summarum: IMHO if you have the resources err on the side of caution: Moderately oversized samples afford greater flexibility, heightened precision, and enhanced robustness.
It is generally recommended that, in order to validate a scale, ideally, there must be at least ten respondents per item in the scale. This means that in order to validate a scale with 40 items, one must have a minimum of 400 respondents.
source- Article Best Practices for Developing and Validating Scales for Heal...
I have seen recommendations for either 5 or 10 respondents per item for this kind of work.
Hello, all depends of the level of communalities you have. See this article for example: https://www.tandfonline.com/doi/abs/10.1207/s15327574ijt0502_4#:~:text=There%20is%20no%20shortage%20of%20recommendations%20regarding%20the,and%20absolute%20ranges%20from%20100%20to%20over%201%2C000. or this one Article Sample Size in Factor Analysis
Soper’s sample size calculator may help you determine the right sample size for your research. This calculation method, developed for studies applying structural equation modeling, has recently been accepted by top-tier journals (e.g., https://doi.org/10.1080/13683500.2023.2301458).
Please follow the reference and the link below to benefit from the calculator: Soper, D. (2018). A-priori sample size calculator for structural equation models. https://www.danielsoper.com/statcalc/calculator.aspx?id=89
Good luck
you can use the Kline approach, in which the sample size is calculated based on the number of items. for example, if you have 40 items, then your sample size will be 400 (by multiplying 40 items x 10 per item).
the reference:
Kline, P. (2015). A handbook of test construction (psychology revivals): introduction to psychometric design: Routledge.
I agree with Professor David L Morgan, if this is for instrument development 5-10 can work.
You mention this is a "standard questionnaire". If it is a truly standard questionnaire the developer has already analyzed it for validity and reliability and that information is available to be cited in your methodology section of your research. Sample size for completing the questionnaire can be calculated as others have mentioned above or you can use different sample sizes and power calculations.
However if you are developing a new questionnaire yourself you have to deal with a lot of issues more complicated than can be covered briefly here, e.g. face validity, content validity, reliability coefficients and scale reliability (assuming you are utilizing a Likert-type scale), etc. If this is what you are doing, I recommend a fellow mentor-researcher to help you through the process. I have mentored my doctoral students with this text:
Designing and Conducting Survey Research: A Comprehensive Guide, 4th Edition by Louis M. Rea, Richard A. Parker
https://www.wiley.com/en-us/Designing+and+Conducting+Survey+Research%3A+A+Comprehensive+Guide%2C+4th+Edition-p-9781118767023
Hope this reply helps you move forward with your research.
A sample size of 5-10 would be adequate to validate the tool (usually 10% is recommended). Fully agree with the other researchers on 5-10.
Dear Diana Elena Lazar,
It depends on whether the questionnaire is designed to measure a latent variable or not.
If the questionnaire is not intended to measure a latent variable, then the minimum sample size required for validating a questionnaire will usually depend on the type of statistical analysis which is deployed for the measurement of its validity during the field-test.
On the other hand, if the questionnaire is intended to measure a latent variable, then the minimum sample size required for validating it can usually follow the rule of thumb of 10:1, in that at least five times the total number of items found in the questionnaire.
This is because any questionnaire which is designed to measure a latent variable will require exploratory factor analysis to determine its construct validity during its validation process.
Hence, if your questionnaire is designed to measure a latent variable and it also consists of 40 items, then it will require at least 40 x 10 = 400 respondents as a minimum sample size for its validation.
I sincerely hope the above has adequately addressed your query.
Note: You can refer to the attached e-book on validation of a questionnaire if you have any further questions on the questionnaire validation process.
My sincere apology: There is a typo in the answer I have just provided 5 minutes ago.
The above sentence "On the other hand, if the questionnaire is intended to measure a latent variable, then the minimum sample size required for validating it can usually follow the rule of thumb of 10:1, in that at least five times the total number of items found in the questionnaire." is wrong.
It should be corrected to "On the other hand, if the questionnaire is intended to measure a latent variable, then the minimum sample size required for validating it can usually follow the rule of thumb of 10:1, in that at least ten times the total number of items found in the questionnaire."