The sample size is an important feature of any empirical study in which the goal is to make inferences about a population from a sample. How many people to send it out to also depends on at what level you need to differentiate, i.e. tell if the results from different AOIs or media differ or not.
To get stable data, you should normally not go below 30 good recordings per media (minimum) in your experiment and preferably aim to get about 50 per media (recommended). The usable rate is depending on the panel you field your experiment in but usually varies between 20-40%.
The Sticky platform helps guiding you in the decision of the final sample size based on where you field it and on the design of your experiment. The formula used is:
Example:
Sending out an experiment in a panel with estimated usable rate of 45%. The playlist included a randomized list with 4 images of which 2 are shown to each respondent. The Sticky platform would give you the following options:
Recommended: (50/40%) x (4/2) = 250 completes
Minimum: (30/35%) x (4/2) = 150 completes
Tip: When interpreting the results the following simplification and a rule of thumb for confidence intervals (CI) can be used. Applies for AOIs larger than 200x200 px and sample sizes n=50 or larger:
Seen: Value +/- 10% at 95% confidence level
Time looking on AOI: Value +/- 0.2 sec at 95% confidence level
Time to first notice: Value +/- 0.4 sec at 95% confidence level