# Effect of Color on Taste - Statistical Significance

#### Tastetest

##### New Member
Hi, for my 8th grade science fair project, I am testing to see if color affects one’s perception of taste. Specifically, I tricked test subjects into thinking that they were tasting different manufacturers’ chips and telling me which one tasted best, but in reality all the chips were the same and I was testing to see if the color of the bowl that the chips were in had an effect on taste. I tested 6 different bowl colors, so theoretically, each bowl should have a one in six chance of being picked as “best tasting”. I have a total of 64 taste test results. I tested over 5 sessions where the number of tests varied from session to session. My overall results for the colors that were picked as “best tasting” are (rounded to the nearest percent): Purple=11%, Blue=11%, Orange=13%, Yellow=16%, Green=19% and Pink=31%. I expected purple to be lower because previous research suggests that we associate purple with rotten, BUT is 11% statistically significant compared to the theoretical probability of 16.7%? Similarly, pink was most often chosen as “best tasting”, but is 31% statistically significant compared to the theoretical probability of 16.7%?
Can anyone help me with how to determine whether the deviation from the theoretical probability is statistically significant?
Thanks,
Matt

#### Karabiner

##### TS Contributor
If I understand you correctly, then you had 64 subjects,and each
subject showed 1 response (preferred chip/bowl). A global test
of whether the distribution of responses significantly differs from
an equal distribution would be a one-sample Chi square test.
If the null hypothesis of an equal distribution is rejected, then
you could calculate the standardized adjusted residuals in each cell.
Values below -1.96 or above +1.96 could be considered as indications
of "significant" deviations.

With kind regards

K.