This week is devoted to your midterms. The data set should be completed as a group--it takes things that you have learned throughout the course thus far. Turn this in in the usual fashion. It must be completed by Wednesday, March 27 at midnight.
Thursday, March 28 you will have a written midterm, similar to the quizzes. Some of it will be multiple choice. The rest of it will not be.
Question Zero: A survey
If you haven't taken the survey for Mr. Wilson yet, please do so. EACH PERSON IN THE GROUP SHOULD DO THIS!
Question One: Midterms
I was looking at some of my old midterms and was trying to determine if they were connected to how people did in the class as a whole (I certainly would hope so).
Below are the grades in the class and the grades on the midterm in the .csv file.
1) Find the correlation between the sets of data.
2) Do a full analysis of the data, creating appropriate graphs, lines, and anything else that’s necessary.
3) Are there any possible influential points in this data? Create a graph that has all possible influential points a different color than the ones you do not suspect.
4) Walk me through your process of deciding whether or not an influential point should be removed. (Hint: this should involve removing the influential data points individually and seeing what changes (if anything)). Explain to me your decisions on each influential point (whether or not you are going to choose to remove it).
5) If you removed any points, make a new line of best fit and tell me what it is. (you do not need to recreate graphs or anything else there).
6) If a student had a 70 in my class, what do I expect them to get on the midterm?
7) If a student received a 80 on the midterm, what do I expect the grade to be in the class?
8) Give the residual for the point (93.345, 81).
Question Two: Monkeys and Tones
The usual way to study the brains response to sound is to compare the response to a noise to that of a pure tone (for example, how quickly the brain recognizes speech vs. a pure tone).
Researchers are interested in whether or not monkey respond to monkey sounds quicker than pure tone, so they aenesthitize the monkeys, run a wire directly to their brain, and then play pure tones and monkey calls. They then check the electrical response of each by measuring the number of firing neurons. The more firing neurons, the more the monkey responds to the sound.
Below is a list of how a group of monkeys responded to a tone and then to a call. For example, the first monkey had 474 neurons fire when they heard the tone, and 500 neurons fire when they heard the call.
tone=c(474, 256, 241, 226, 185, 174, 176, 168, 161, 150, 145, 141, 129, 113, 112, 102, 100, 74, 72, 71, 68, 59, 57, 56, 47, 46, 41, 35, 31, 28, 26, 26, 21, 20, 20, 19,18)
call=c(500, 138, 485, 338, 194, 159, 341, 85, 303, 208, 42, 241, 194, 123, 182, 141, 118, 62, 112, 134, 65, 182, 97, 318, 201, 279, 62, 84, 103, 70, 192, 203, 135, 129, 193, 54, 66)
1) If there are any points you think need to me removed, please do so and explain why. Then create a plot and line of best fit for this data. Use tone as the explanatory variable and call as the response. Explain to me whether or not the data shows a strong correlation.
2) Perhaps a better way of dealing with this data is finding the difference between the call and tone data for each specific monkey (this is called a "matched pairs procedure"). First, make sure you remove any data considered removable in question one. Then create a new list of data that shows the difference between response to the call and the tone. (when done with this step you should have a single list of data. call it call.tone.diff or some such other creative thing).
3) Run an analysis on this set of data (call.tone.diff or whatever you called it).
4) Do you think that monkeys respond more to a monkey call as opposed to a tone? Back your assessment up with info from the previous three pieces of this question.
Question Three: some writing.
For each of the following, give an example and explain your choice. This may be from a set of data we have used or a more abstract idea or data set you have seen or used. Each example should be different for each piece (in other words, don't repeat). Be sure to FULLY explain why you think your choice is a good one.
a. A situation where you would want to use mean and standard deviation.
b. A situation where the five point summary would be a good description of the data.
c. A situation where looking at the mean and standard deviation would hide the true nature of the pattern or data.
Question Four: Sampling.
You would like to do a survey of BBA, finding out about what types of shoes people wear (both size and brand).
1) Create a survey. This survey should be no more than five questions long and should be as unbiased and clear as possible.
2) Explain how you would go about giving this survey to the population, again trying to avoid bias. You do not have to actually give the survey, but write out all the steps that would need to take place in order for you to do it effectively. If there is any code or other resources you would need, make sure to include them.