StatBasics
< Back DiffTest Page 5 of 8 Next >
Site highly rated by Schoolzone.co.uk
How-to-use button View-calculator button Show-glossary button

 

Amount of data

Amount of Data Panel

 

The Amount of data panel is useful when planning an investigation that aims to give reassurance about samples being similar. A section of this site is devoted to testing for similarity.

For the task described in the Task panel and the data (number correct, out of...) described in the Frequencies panel, the Upper Bound estimates the largest difference that may plausibly exist. But the upper bound will have the value shown only if the number of correct trials is exactly the number shown in the Frequencies panel.

When designing an investigation, we can specify the task, the total number of trials and the degree of confidence we require in the estimated bounds but it is not possible to predict what the number of correct trials will actually be. The number in the Amount of data panel is the probability that if the investigation proceeds with the specified total number of trials, the upper bound will not exceed the value shown in the Confidence bounds panel.

If we arrange for this probability to be small, we gain reassurance that we are unlikely to get a result with worse bounds than those shown in the Confidence bounds panel if the investigation proceeds as planned.

We can make this probability smaller by increasing the hypothetical frequency in the Number correct window, though doing so also increases the values in the Upper Bound windows.