Displaying Quantitative Data Graphically and Describing It Numerically
description
Transcript of Displaying Quantitative Data Graphically and Describing It Numerically
Displaying Quantitative Data Graphically and Describing It Numerically
AP StatisticsChapters 4 & 5
Displaying Quantitative Data
• Histogram
• Stem-and-Leaf Plots
• Dotplots
• (Timeplots)
Histograms
• Bins and counts give the distribution of the quantitative data
• Bars touch—data is continuous
• Relative frequency histogram—useful and shows percentages, not counts
Stem-and-Leaf Plot
• Can see each individual data point
• Stem is like bin• Might need to “split”
3 34779 Key: 3 4 = 344 66777895 3567777776 00017 99
8 0222222 577799999 Key: 2 4 = 243 444443 6677899994 23333444444 577779
Dotplot
• Useful in seeing how many individual data points in bin
• Good for small sets of data
• Not used too often
Describing a Distribution
• Whenever you are describing a distribution you need to describe it by the– Shape– Center– Spread– Any Unusual points (outliers, gaps)
Shape
• Is the shape?• Uniform, Symmetric,
Skewed
• How many modes (high points)– Unimodal, bimodal,
multimodal
Center and Spread
• How we describe the center and spread of a distribution depends on the shape of the distribution.
Skewed Distribution
• Center: Median• Spread: Interquartile Range (IQR)
• Both of these are “resistant”• Both should include units
Skewed Distribution
How to find the IQR 1. Find median 2. Find the median of both halves of data
the lower median is 1st Quartilethe upper median is 3rd Quartile
3. Subtract the two quartile scores** 1st Quartile = 25th percentile** 3rd Quartile = 75th percentile
Symmetric Distributions
• Center: Mean
• Spread: Standard Deviation
• Both are not “resistant”• Both should include units
ny
y
Standard Deviation
• Takes into account how far each value in a data set is from the mean
Formula:
1
2
n
yys
Properties of standard deviation1. Only use with mean2. If s = 0, there is no spread and all data pieces
are same—otherwise s>0 and s gets larger as data pieces get more spread out.
3. A few outliers can really change the value of the standard deviation
Finding Standard Deviation by Hand
Find the standard deviation:10, 14, 15, 16, 20
Other information
• If distribution is symmetric, then mean=median
• If skewed right, mean>median• If skewed left, mean<median• Spread of distribution is just as important as
the center• How accurate: one or two decimal points
more than original data
Distributions with Outliers
• Really just data that seems unusual• Formally we compute fences and if data point
is outside the fences, we consider it an outlier• Always use common sense
• Upper fence:• Lower fence: IQRQ
IQRQ5.15.1
1
3
Distributions with Outliers• Tricky situation• Since outliers affect mean and standard deviation, it is usually
better to use median and IQR• If the mean and median are not similar in value, report the
median and IQR • If the mean and the median are similar in value, report the
mean and standard deviation.• Sometimes (especially if the mean and median are not
similar) it is a good idea to report your center and spread with and without the outlier and see what kind of effect removing the outlier has on the distribution.
Boxplots
• Complement histograms by providing more specific information
• Look at histogram and boxplot together
• Most useful when comparing distributions
Boxplots
5-Number Summary: Minimum, 1st Quartile Score, Median, 3rd Quartile Score, Maximum