I’ve always wondered what makes people like different flavors of ice cream. Could it be differences in their genetics that caused different flavors to taste better than others? Maybe familiarity of flavors played a part in which tastes best to different individuals? I never had a proper way of testing the popularity of different flavors of ice cream and if those who liked a certain flavor had some trait in common until now. I’ve decided to test for association/independence between ethnicity and favorite flavor of ice cream. Ethnicity is based off of the main four races: White, Black, Asian, and Hispanic. Flavors of ice cream individuals are able to choose from are the three basic flavors: vanilla, chocolate, and strawberry. My populations are all users on Myspace.

For the data to be used in a proper Chi-square test for association/independence certain conditions must be met. The data must be a SRS, each population must be at least ten times greater than their sample sizes, and all expected counts must be greater than one with no more than twenty percent of the expected counts less than five.

The data was collected from a simple random sample using users on Myspace.com, assuming that all users on Myspace can generalize to all possible people. The method for collecting the data was through getting users randomly generated via the site’s “browse people feature” then sending a message asking for two pieces of data, ethnicity and favorite flavor of ice cream from the available options. This meets the requirement of an SRS.

Expected: (Row total x Column total)/sample total

All expected counts are greater than 5 so the condition of all expected counts needing to be greater than 1 with no more than 20% less than 5 is met.

Observed: White = 100 Asian = 67 Black = 48 Hispanic = 51

There are definitely more than 1000 white people available, 670 Asians available, 480 black people available, and 510 Hispanic people available, so that meets the condition of all populations needing to be greater than 10 times their sample sizes.

All the conditions are met which means my test should be accurate and generate to the population. The Chi-square value is 23.6385 with a p-value of .00060855 and 6 degrees of freedom.

This means the probability of getting results like this if all flavors were liked the same is almost never by chance variation. So we have pretty good evidence to show that ethnicity influences favorite flavor of ice cream. It can be concluded that from my results white people generally prefer chocolate ice cream, Asians prefer vanilla, black people prefer vanilla, and Hispanic people are fairly evenly distributed. If I were to redo this project I’d probably try to gather a slightly larger sample and possibly offer more flavors, but this test was fairly good so there isn’t much that needs to be done to improve upon it.