Pilot testing

Ensure you properly pilot test your instrument (Boynton & Greenhalgh, 2004; Story & Tait, 2019). 

Sample size

A power analysis is recommended if you plan to use inferential statistics. Regardless, you should consider what sample size might be viewed as adequate by your audience (Story & Tait, 2019). 

Consider (Story & Tait, 2019): 

• Expected response rate

• Oversampling

• Response bias

Selecting a distribution method

Choose a distribution method that suits your needs and is the best fit for your topic and available funding (Boynton & Greenhalgh, 2004; Story & Tait, 2019).

Surveys can be an efficient way to understand perspectives of large populations (Boynton & Greenhalgh, 2004; Story & Tait, 2019). Every effort should be made to obtain a representative sample. Unfortunately, cost and resource limitations often limit survey representativeness (Story & Tait, 2019). Oversampling (collecting extra data from predetermined groups) can be used to capture perspectives of underrepresented groups (Boynton & Greenhalgh, 2004; Story & Tait, 2019). Sample size should be large enough that results approximate responses from the entire population with 95% confidence (Story & Tait, 2019).