Part II: Two - Variable Statistics
description
Transcript of Part II: Two - Variable Statistics
![Page 1: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/1.jpg)
Part II: Two - Variable Statistics Statistical studies often involve more than one variable. We are interested in knowing if there is a relationship between the two characteristics for the same subject.
Example: A person's age and the time spent using a mobile phone.
When the data is quantitative (numbers), the variables can be written as an ordered pair (x, y).
Correlation is the study and description of the relationship (if any) that exists between the variables.
![Page 2: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/2.jpg)
A) Qualitative Interpretation of Correlation
Data can be organised and displayed in a scatterplot (Cartesian plane) or a contingency table.
By inspection, we will describe the type, the direction, and the intensity (or strength) of the relation between the variables.
![Page 3: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/3.jpg)
Type: Refers to the function that best fits the relation between the variables. We will be using linear correlation.
Direction: If both variables move in the same direction (increase together or decrease
together), then the direction is positive.
If both variables move in opposite directions, then direction is negative.
Intensity: Strength may be categorised as... Zero, weak, moderate, strong or perfect.
![Page 4: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/4.jpg)
Example: In a gym class students were required to do push-ups and sit-ups. Each student's
achievements were recorded as an ordered pair. The first number refers to push-ups
and the second to sit-ups.
(27, 30), (26, 28), (38, 45), (52, 55), (35, 36), (40, 54), (40, 50), (52, 46), (42, 55), (61, 62), (35, 38), (45, 53), (38, 42), (63, 55), (55, 54), (46, 46), (34, 36), (45, 45), (30, 34), (68, 62)
![Page 5: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/5.jpg)
Contingency Table - A table of values (often involves classes) - Listed down the left side is one variable; listed across the top is the other variable. - Each row and column is added and the totals are displayed; the bottom right cell shows the total frequency of the distribution.
![Page 6: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/6.jpg)
(27, 30), (26, 28), (38, 45), (52, 55), (35, 36), (40, 54), (40, 50), (52, 46), (42, 55), (61, 62), (35, 38), (45, 53), (38, 42), (63, 55), (55, 54), (46, 46), (34, 36), (45, 45), (30, 34), (68, 62)
![Page 7: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/7.jpg)
When the majority of the values fall into a diagonal of a contingency table, and the corners contain mostly zeros, then the correlation is said to be linear and strong.
Direction is positive if the diagonal is Direction is negative if the diagonal is
![Page 8: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/8.jpg)
(27, 30), (26, 28), (38, 45), (52, 55), (35, 36), (40, 54), (40, 50), (52, 46), (42, 55), (61, 62), (35, 38), (45, 53), (38, 42), (63, 55), (55, 54), (46, 46), (34, 36), (45, 45), (30, 34), (68, 62)
Scatterplot- A Cartesian graph - Each ordered pair is
represented as a point
![Page 9: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/9.jpg)
B) Quantitative Interpretation of Correlation
The correlation will be represented by a number, called the correlation coefficient. This coefficient will range from to .
Its symbol is r.
![Page 10: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/10.jpg)
(27, 30), (26, 28), (38, 45), (52, 55), (35, 36), (40, 54), (40, 50), (52, 46), (42, 55), (61, 62), (35, 38), (45, 53), (38, 42), (63, 55), (55, 54), (46, 46), (34, 36), (45, 45), (30, 34), (68, 62)
1) Draw a scatterplot
2) Draw a line that "best fits" the points. This line passes through the middle of the scatterplot.
3) Around the points, draw the smallest rectangle possible. Two of the sides should be parallel to the line.
4) Measure the dimensions of the rectangle.
![Page 11: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/11.jpg)
+ if it's increasing, if it's decreasing
r = + 1 -
r = +(1- 0.3)
r = 0.7 moderate
2.47.9 )(
5) Calculate the correlation coefficient, r.
real r = 0.87
![Page 12: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/12.jpg)
r MeaningNear 0 Zero correlationNear ± 0.5
Weak correlation
Near ± 0.75
Moderate correlation
Near ± 0.87
Strong correlation
Near ± 1 Perfect correlation
![Page 13: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/13.jpg)
The Regression Line
Also called a line of best fit, a regression line is one that best represents (or passes through) the points of a scatterplot. It passes through as many points as possible, going through the middle of the scatterplot.
There are several different methods of drawing a regression line.
The regression line may be used to predict values that do not appear in the distribution.
![Page 14: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/14.jpg)
Using the regression line, we can predict the value of one variable, given the value of the other. The reliability of the prediction depends on the strength of the correlation.
Examples:
a) Predict the number of b) Predict the number of sit-ups a student can push-ups a student can do if he can do 49
do if she can do 70 push-ups. sit-ups.
Determine the equation of the line.
![Page 15: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/15.jpg)
predicted
actual
Recall:
Since the correlation is moderate/high, we can be confident that our predictions are good.
![Page 16: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/16.jpg)
Interpreting a Correlation
A strong correlation indicates that there is a statistical relationship between two variables.
It does not, however, explain the reason for the relationship or its nature. There are other things to consider...
![Page 17: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/17.jpg)
![Page 18: Part II: Two - Variable Statistics](https://reader034.fdocuments.net/reader034/viewer/2022050802/56816194550346895dd13564/html5/thumbnails/18.jpg)
The outlier any point that stands alone away from the group is left outide the rectangle.DON'T INCLUDE THIS POINT IN THE RECTANGLE
OUTLIER
Is there a strong correlation?What is the correlation coefficient?What would be the value of y when x = 2? Make a predictionwhat is the equation for the line of best fit?