# Overlapping Samples T-test

#### ResearcherPat

##### New Member
Hi,

Could someone please tell me how to do an overlapping samples t-test? In my survey, respondents rate (0 to 10) any of 10 options that they're familiar with. So I want to compare the average ratings (significant difference or not) of the options to each other, keeping in mind that in each pair, some people rated 1 option, some rated only the other option, and some rated both. How would I do this in SPSS or Excel?

Thanks!

#### hlsmith

##### Less is more. Stay pure. Stay poor.
Please add more detail in your description, or an actual example. I am unsure what you are trying to do.

#### ResearcherPat

##### New Member
Suppose respondents are rating the taste of 10 fruits that they eat every week, where 10 is excellent, and 0 is terrible. The overall mean rating for apples is 7.9, and the overall mean rating for oranges is 8.7. Is the 8.7 for oranges significantly higher than the 7.9 for apples?

Now people are only rating fruit they eat every week, so the 8.7 is based on 100 people who eat oranges, and the 7.9 for apples is based on 70 people who eat apples, but out of those people, 50 eat both apples and oranges, thus 50 people eat only oranges and 20 people eat only apples.

So the parts of sample that eat both could use a paired-sample t-test, while the parts that eat only one could use an independent-sample t-test. So how do I combine these to compare the overall mean ratings?

Thanks!