DISCUSSION

by dena raheem & tia elkhatib

By using our sample size of only 50, it lead us to some problems with getting a p-value. From all the way back in the beginning, we started off with 50 randomly drawn names of teachers from a container. Not all the teachers were equally divided among subjects; we did not have an equal number of english teachers to math teachers, or science teachers to history teachers. We had more history teachers than any other subject, about 6, and only about 3 math teachers; this is one of the major weaknesses. To perform a chi-squared test, we needed to have a cell count of atleast 5 teachers per subject area. Since we had a few subjects of 5 teachers or more, and a few subjects with less than 5 teachers for a cell count it really threw of the chi-squared test and we couldn't get a p-value. So we tried to fix up the cell counts by condensing the classes together, to achieve cell counts greater than or equal to 5 per subject. So we took all the foreign languages and English classes, and condensed them into one subject, "languages". Then we condensed together health teachers, physical education teachers, and science teachers together as "sciences" subject. The math and the business classes were condensed as well.

 After all of that work, we still could not get a p-value, some of the expected cell counts were STILL too small.

So finally what we did, was we arranged the teachers according to the "Left Brain vs. Right Brain Theory. That is, we categorized the teachers according to analytical subject or creative subject. According to the theory, users of the left brain are more into the fields of logic such as that of business, math, and science teachers. The right brain controls creativity, that of foreign languages, English, history, and art teachers. So how did we go about doing this? Simple. We once again, recorded our sample of teachers in a list on Minitab. In the first column we labeled each of the teachers as either analytical or creative. Math, science, business teachers were assigned "analytical" and history ,art, and language teachers were of course "creative". Then we recorded his/her favorite color to the right of the first column. In a third column, we labeled each as cool or warm. We thought that maybe using "cool and warm colors" would maybe help us narrow down our results so we could get a more clear answer.

Now... We finally got our p-value! Our p-value is listed in the Raw Data link to the side.

Would Tia and I extrapolate our study to another population? Probably not. Overall there is truly, absolutely no relationship between teachers of certain subjects and their favorite colors. Just as teachers are ordinary people (most of them), whether you are a teacher or not, it makes no difference in your favorite color. There is experimentally no type of people that will definately favor a color so much over than another group of people. So if I were to extrapolate this study to any other type of place my results probably amount the same way as mine: no special association in any way.

Since the most prevelant favorite color among people is blue, no matter what population you use, blue will most likely still be the most popular favorite color.

The studies in my background I had read about prior to the study I conducted should have told me something about the results that were drawn. If you look at the background page, I have recorded a few studies about people and their favorite colors. It is PEOPLE that is the key word. I believe it has nothing to do with being a teacher that makes your favorite color, but your favorite color derives from the type of person you are. Once again, mentioning the studies on my background page, there is a difference between men and women when it comes to colors of all sorts that they like and dislike. So I think I should have stuck to doing a study that compared and contrasted men and women teachers and their favorite colors. That would have most likely lead me to results where there is actually a difference in something, finding a difference between two populations perhaps.

As dumb as it may sound, I also thought that maybe, if a study was performed on a city of great fans of a particular sports team, it may change the fact that there is no relationship of favorite colors. Lets take Michigan University for example. The colors of the school are blue and yellow. So maybe, most people there who happen to be fans of the great school may lead into favoring those colors of the team perhaps? Hmm... There's something to think about.

I don't necessarily think that the sample size would really play a big role in results. We had a pretty decent size anyway, but it does not change the results, or in other words, an extraneous factor.