Chapter 3: Displaying and summarizing quantitative data ... mahinda/stab22/wk2.pdf Chapter 3:...
date post
21Aug2018Category
Documents
view
214download
0
Embed Size (px)
Transcript of Chapter 3: Displaying and summarizing quantitative data ... mahinda/stab22/wk2.pdf Chapter 3:...
1
Chapter 3: Displaying and summarizing
quantitative data p52
The pattern of variation of a variable is called
its distribution.
2
Histograms p53
The breakfast cereal data
Study collected data on nutritional content per serving
(and other things) of 77 different breakfast cereals, so
that different cereals can be compared.
 Calorie content varies from 40 to 180
 most cereals have between 100 and 120
calories/serving.
 A few have a lot more or a lot less, but only a few.  Shape symmetric: falls away from peak about same
both sides.
3
 Describing a distribution (p57)
 In any graph of data, look for the overall
pattern and for striking deviations from
that pattern (e.g.. gaps. P54)

 Overall pattern of a distribution can be
described by its shape, centre, and spread.

 An important kind of deviation is an
outlier, an individual value that falls outside
the overall pattern.

 For now we can describe the centre of a
distribution by the median (the midpoint)

 Some other things to look for in describing
shape are:

4
 Does the distribution have one or several
major peaks, called modes? A distribution
with one major peak is called unimodal.

 Is it approx. symmetric or skewed in one
direction.
5
Centre of the Distribution p 57
The mean and median (measures of centre)
Measuring center: mean Two common measures of center are the
mean and the median.
The two measures behave differently.
Example Find the mean of the following observations.
4, 5, 9, 3, 6
Solution:
4 6 9 3 6 28 5.65 5
mean + + + += = =
If there are n observations , , ,1 2
y y yn in a
sample, the sample mean (denoted by y ) is
given by
' ysum of y syn n
= = .
6
Example The annual salaries (in thousands) of a
random sample of five employees of a
company are:
40, 30, 25, 200, 28
40 30 25 200 28 323 64.65 5
mean + + + += = =
If we exclude 200 as an outlier,
40 30 25 28 123 30.754 4
mean + + += = =
Mean is sensitive to the influence of extreme observations. It cannot resist influence of the
extreme values. Mean is NOT a resistant
measure of center. (p62)
7
Measuring center: the median (p62)
The median is the midpoint of the
distribution, the number such that half the
observations are smaller and other the half
are larger.
To find the median of a distribution: 1. Arrange all observations in order of
size, from smallest to largest.
2. If the number of observations is odd the median is the center observation in
the ordered list.
3. If the number of observations is even the median is the average of the
two center observations in the ordered
list.
Examples 1. The annual salaries (in thousands) of
a random sample of five employees of a
company are:
40, 30, 25, 200, 28
8
Arranging the values in increasing order:
25, 28 30 40 200
median = 30
Excluding 200 median = (28+30)/2.
Note that the mean for this data set was 64.6 and the influence of the extreme
value 200 is much less.
StatCrunch commands Stat > Summary Statistics
StatCrunch output for the data in Example
above is as follows:
Summary statistics:
Column n Mean Median
salary 5 64.6 30
9
Mean versus median
The median and mean are the most common measures of the center of a distribution.
If the distribution is exactly symmetric, the mean and median are exactly the same.
Median is less influenced by extreme values.
If the distribution is skewed to the right, median < mean
If the distribution is skewed to the left, mean < median
10
Mean and median from cereal data
Summary statistics: Column n Mean Median
potass 77 96.07792 90
11
Distribution of a simulated data set (100 values)
y
Fre
qu
en
cy
100908070605040
20
15
10
5
0
Histogram of y
Descriptive Statistics: y Variable N Mean Median
y 100 82.69 86.19
12
Questions
1. You are asked to recommend a measure of center to characterize the following data:
0.6, 0.2, 0.1, 0.2, 0.2, 0.3, 0.7, 0.1, 0.0, 22.5, 0.4.
What is your recommendation and why?
2. The mean is ____ sensitive to extreme values than the median
a) more b) less c) equally d) cant say without data
3. Changing the value of a single score in a data set will necessarily cause the mean to change. (T/F)
4. Changing the value of a single score in a data set will necessarily cause the median to change. (T/F)
13
Spread: interquartile range p68
1st quartile Q1 has of data values below it and above
3rd quartile Q3 has of data values below it and above
Find a quartile by taking lower (upper) half of data, and finding median of that half.
Interquartile range is IQR=Q3Q1. Larger = more spread out.
Example: 3, 5, 7, 7, 8
lower half 3, 5, 7 (include middle), so Q1=5
upper half 7, 7, 8 so Q3= 7
IQR=75=2
IQR not affected by extremely high or low values, like median.
14
Measuring spread Standard deviation (p64)
The variance ( 2s ) of a set of n observations
, , ,1 2
y y yn is
2 2 2 2( ) ( ) ( ) ( )2 1 21 1
yy y y y y yn ysn n
+ + +
= =
.
The standard deviation(s) is the square root of
the variance ( 2s ).
i.e.
2 2 2 2( ) ( ) ( ) ( )1 21 1
yy y y y y yn ysn n
+ + +
= =
Example
Find the standard deviation of the following
data set: 5, 8, 2
3n= , Mean (x ) = 5 8 2 15 53 3
+ += =
2 2 2(5 5) (8 5) (2 5) 182 93 1 2
s + +
= = =
9 3s= = .
15
StatCrunch: cereal data(potassium)
Summary statistics:
Notes
s measures the spread about the mean x .
s = 0 only when there is no spread. This happens
only when all observations have the same value.
s, like the mean x , is not resistant.
Column n Mean Variance Std. Dev. Median Q1 Q3
potass 77 96.07792 5081.8096 71.28681 90 40 120
16
The fivenumber summary p69 The fivenumber summary of a set of
observations consists of the minimum,
the first quartile, median, the third
quartile and the maximum.
These five numbers give a quick summary of the both center and the
spread of the distribution.
StatCrunch commands:Stat > Summary Statistics
17
Example
Data 10, 11, 14, 15, 17, 19, 21, 28, 35:
find median
17, Q1= median of 10,11,14,15, 17 ie. 14
Q3=median of 17, 19, 21, 28, 35 ie. 21
find Q1 and Q3
find interquartile range
2114=7
find 5number summary min, Q1, median, Q3, max.
10, 14, 17, 21, 35
18
Boxplot (numbers from example above) p70
box goes down the page, with scale on left.
centre of box at median (17)
top of box at (21) Q3
bottom of box at (14) Q1
calculate R=1.5 x IQR= 1.5(2114)=10.5
upper fence at Q3+R=21+10.5=31.5
lower fence at Q1R= 1410.5=3.5
draw lines connecting box to most extreme value within fences
plot values outside fences individually. These are suspected outliers and deserve
to be investigated.
19
StatCrunch boxplot (select use fences to identify
outliers):
20
Chap 4: Understanding and Comparing Distributions p99
Cereals classified by shelf where found in grocery store:
1=top shelf
2=middle shelf
3=bottom shelf
Want to compare sugar/serving for shelves.
How about a histogram for each shelf, put results side by
side?
21
so where are the most sugary cereals?
maybe on shelf 2? Hard to decide.
how about sidebyside boxplots?
22
Median definitely highest
for shelf 2, lowest for
shelf 1.
Easier to see than on histograms.
Bonus: shelf 1 sugar right
skewed, shelf 2 sugar leftskewed.
where did median go for shelf 2 sugar?
23
Look at stemplot for shelf 2 sugar
a lot of the cereals had
sugars exactly 12.
so Q3 and median for shelf 2
sugars are the same.
do the means tell the same story as the medians?
does the skewness show up here as well?
yes: compare means and
medians
24
Choosing a summary p76 (in ch 3)
The fivenumber summary is usually better than
the mean and the standard deviation for
describing skewed distributions or distributions
with strong outliers.
Use mean and std. deviation for reasonably
symmetric distributions that are free of outliers.