We will be talking about the concepts of Standard Deviation and the 68-95-99.7 rule in class on Thursday.
I won't have time to go over the IQR Rule, so hopefully I can put a good explanation of it here.
The IQR Rule is a way we determine if we have outliers. As we know, the IQR is the inter-quartile range, or (Q3-Q1). Here is the rule:
If an observation is GREATER than (1.5 multiplied by IQR) plus Q3, then the observation is a high outlier. If an observation is LESS than Q1 minus (1.5 multiplied by IQR), then the observation is a low outlier. In mathematical form:
Observation < Q1 -1.5*(Q3-Q1), then the observation is a low outlier and
Observation > Q3 + 1.5*(Q3-Q1), then the observation is a high outlier.
Let's do an example to make sure it makes sense. Let's say that my five number summary is as follows: 2, 10, 15, 20, 50. I want to know if 50 or 2 is an outlier. Let's do the math.
(Q3-Q1) = (20-10)=10.
1.5* 10 = 15
To Check for the high outlier:
Q3+15=20+15=35. Since 50>35, 50 is an outlier.
To check for the low outlier:
Q1-15=10-15=-5. Because 2 is NOT < -5, then 2 is NOT an outlier.
Hopefully that helps with that concept.
Questions 9-13 should not be that difficult. These all deal with concepts we discussed last Thursday. Particularly the mean and median ones: Remember, the median cuts the data in HALF (not necessarily where the peak is). Then remember where the mean goes based on skewed data.
Hope that helps,
Hillary
Tuesday, January 24, 2012
Assignment 5
We went through most of the assignment already in class (the ones we didn't are very similar in nature to the rest of the questions).
The only question I have advice about is Question 13. You can do this by sight. Look at the graphs. When we say "clear", we mean, is there a data point that is further away from the others by quite a margin? Don't over think it.
-Hillary
The only question I have advice about is Question 13. You can do this by sight. Look at the graphs. When we say "clear", we mean, is there a data point that is further away from the others by quite a margin? Don't over think it.
-Hillary
Tuesday, January 17, 2012
Assignment 4
We will go through the other difficult ones in class. Here is some help on the others:
Question 5-7: This is a really long, hard problem if done by hand. So, I already plotted the data for you to make your life easier :). Be sure you understand HOW I plotted this data: be sure you can interpret the graph. (You will be required at times to plot your own data, like for example on a written portion of an exam. So please be sure you understand how to graph).
See how I "split the stems" as it asked? This allows one line in a chart not to be super long. Splitting means having two "stems" (like double 2's or double 1's as shown above). Realize that my scale is 1|7 = 1700. I also rounded to the nearest hundred like asked.
Use this to find the rest of the answers. I understand it may be difficult to say the exact shape "is it only "slightly" or "totally" skewed?) But the medians are different, so you should be able to determine the answer. (You may ask the quesiton: if you rounded, how am I to know the actual median? Let's say the median fell on a 2|8 (this is incorrect). If it was the first 2|8, I would use the number "2828" because it is the lower of the 2800 numbers. Hope that makes sense).
Hope that helps :)
-Hillary
Question 5-7: This is a really long, hard problem if done by hand. So, I already plotted the data for you to make your life easier :). Be sure you understand HOW I plotted this data: be sure you can interpret the graph. (You will be required at times to plot your own data, like for example on a written portion of an exam. So please be sure you understand how to graph).
| 0 | 4 |
| 0 | 66777888 |
| 1 | 11233 |
| 1 | 78999 |
| 2 | 012334 |
| 2 | 578899 |
| 3 | 0001 |
| 3 | 88 |
| 4 | |
| 4 | |
| 5 | |
| 5 | 7 |
See how I "split the stems" as it asked? This allows one line in a chart not to be super long. Splitting means having two "stems" (like double 2's or double 1's as shown above). Realize that my scale is 1|7 = 1700. I also rounded to the nearest hundred like asked.
Use this to find the rest of the answers. I understand it may be difficult to say the exact shape "is it only "slightly" or "totally" skewed?) But the medians are different, so you should be able to determine the answer. (You may ask the quesiton: if you rounded, how am I to know the actual median? Let's say the median fell on a 2|8 (this is incorrect). If it was the first 2|8, I would use the number "2828" because it is the lower of the 2800 numbers. Hope that makes sense).
Hope that helps :)
-Hillary
Assignment 3
Sorry this is so late- the next assignments will be ON TIME (meaning at least a day or two before they are due). I was sick this weekend and just didn't do as much as usual!
We talked about this assignment already, but just some reminders.
Make sure you have an actual experiment. This means applying a treatment (be sure you aren't just observing).
Answer the whole question. (This is especially true for #2. When can we infer?)
Careful on the response variable. Make sure it is measurable and on the individual.
Good Luck!
-Hillary
We talked about this assignment already, but just some reminders.
Make sure you have an actual experiment. This means applying a treatment (be sure you aren't just observing).
Answer the whole question. (This is especially true for #2. When can we infer?)
Careful on the response variable. Make sure it is measurable and on the individual.
Good Luck!
-Hillary
Wednesday, January 11, 2012
Welcome! and Assignment 2
Welcome to Stat 121! Hopefully this blog will be useful to you throughout the semester. As a reminder, come here for homework questions we didn't cover in lab as well as helpful reviews for all of the exams.
Some reminders on housekeeping duties at the beginning of the semester:
- Be sure to register your iClicker (using your netID! NOT your 9-digit number!)
- Buy StatsPortal, if you haven't
- Put your netID into StatsPortal at the top of the page
Now, onto some help with Assignment 2:
Subscribe to:
Posts (Atom)