Use of Statistics in Our Daily Life Type of Measurement

Chapter 2 Measurement in Statistics Qualitative (or categorical or attribute) data

(p.53-p.88) Values that can be placed into nonnumerical
categories. Example: genders (male/female);
Goals: districts of a city; flavors of ice cream
2.1 Be able to identify types and levels of
measurement Quantitative data
2.2 Understand the sources of errors Values representing counts or measurements.
2.3 Understand how to use percentage Example: weights; number of students; income
2.4 To study some index numbers

1 2

Example 1 Data Types (p 54) Working with Quantitative Data

Quantitative data can further be described by
¬ Classify each of the following sets of data as distinguishing between discrete and continuous
qualitative or quantitative types.
” Brand names of shoes in a consumer survey Discrete data can take on only particular values and
” Scores on a multiple-choice exam not other values in between (Example: The number of
books that a student has)
” Letter grades on an essay assignment
” Numbers on uniforms that identify players on a Continuous data can take on any value in a given
basketball team interval (Example: The amount of milk that a child
drinks in a day; e.g. 8 oz per day)

3 4
Example 2 Discrete or Levels of Measurement
Continuous? (p 55)
¬ For each data set, indicate whether the data
are discrete or continuous
” Measurements of the time it takes to walk a mile Another way to classify data is to use
” The numbers of calendar years (such as 2007, levels of measurement. Four of these
2008, 2009) levels are discussed in the following
” The numbers of dairy cows on different farms slides.
” The amounts of milk produced by dairy cows on a
farm

5 6

Nominal level of measurement Ordinal level of measurement

involves data that can be arranged in some order,
Characterized by data that consist of names,
but differences between data values either cannot
labels, or categories only, and the data cannot be
be determined or are meaningless or are different
arranged in an ordering scheme (such as low to
high) Example: Course grades A, B, C, D, or F;
Example: Survey responses: yes, no, undecided satisfied, unsatisfied or very unsatisfied

7 8
Interval level of measurement Ratio level of measurement
like the ordinal level, with the additional property Data in which both intervals and ratios are
that the difference between any two data values meaningful. Data at this level have a true zero
is equal or meaningful, however, there is no point (where zero indicates that none of the
natural zero starting point (where none of the quantity is present)
quantity is present)
Example: Distance, weight; Prices of college
Example: A temperature of 850 F is hotter than 830 F by
textbooks ($0 represents no cost)
the same amount that 250 F is hotter than 230 F. But it is
not true that 200 F is twice as hot as 100 F.

9 10

Summary - Levels of Measurement Data Type Level of Measurement

™ Nominal - categories only

Nominal
™ Ordinal - categories with order/ranking
Qualitative
™ Interval - equal differences but no natural Ordinal
starting point
™ Ratio - equal differences and a natural Interval
Discrete
starting point Ratio
Quantitative
Interval
Continuous
Ratio
11 12
Example 3 Levels of Measurement Types of Error
(p 57) ¬ Random errors occur because of random and inherently
unpredictable events in the measurement process (measurements
¬ Identify the level of measurement for each of differ from the true value in a random way, sometimes higher,
sometimes lower)
the following sets of data ¬ Systematic errors occur when there is a problem in the
” Numbers on uniforms that identify players on a measurement system that affects all measurements in the same
way (measurements differ consistently from the true value).
basketball team
” Student rankings of cafeteria food as excellent,
good, fair, or poor
” Calendar years of historic events, such as 1776,
1945 or 2001
” Temperatures on the Celsius scale
” Runners’ times in the Boston Marathon

13 14

Absolute and Relative Errors Example 2 Absolute and Relative

Error (p 62)
¬ The absolute error describe how far a ¬ Find the absolute and relative error in each
measured value lies from the true value case:
” Your true weight is 100 pounds, but a scale says
absolute error = measured value – true value you weigh 105 pounds.
” The government claims that a program costs $99
¬ The Relative error compares the size of the billion, but an audit shows that the true cost is
absolute error to the true value $100 billion.
measured value − true value
Relative error = x 100%
true value

15 16
Accuracy and Precision Example 3 Accuracy and Precision
¬ Accuracy describes how closely a measurement in Your Weight (p 63)
approximates a true value. An accurate measurement is
very close to the true value.
¬ Suppose that your true weight is 102.4
pounds. The scale at the doctor’s office,
¬ Precision describes the amount of detail in a which can be read only to the nearest quarter
measurement. (e.g. distance to the nearest tenth of a km, pound, says that you weigh 102.25 pounds.
to the nearest 0.01 cm) The scale at the gym, which gives a digital
Other precision specifications: readout to the nearest 0.1 pound, says that
Significant figures: Both 0.00235 and 5.83x106 have three you weigh 100.7 pounds.
significant figures
Decimal places: 78.23 is correct to two decimal places and 0.00235
” Which scale is more precise?
is correct to the fifth decimal places ” Which is more accurate?

17 18

Absolute and Relative Change Example 2 World Population

Growth (p 69)
¬ The absolute change describe the actual increase or ¬ World population in 1950 was 2.6 billion. By
decrease from a reference value to a new value the beginning of 2000, it had reached 6
billion. Describe the absolute and relative
absolute change = new value – reference value
change in world population from 1950 to
¬ The Relative change describes the size of the 2000.
absolute change in comparison to the reference value

new value − reference value

Relative change = x 100%
reference value

19 20
Absolute and Relative Difference Example 3 Russian and American
Life Expectancy (p 70)
¬ The absolute difference is the difference between the ¬ Life expectancy for American men is about
compared value and the reference value 75 years, while life expectancy for Russian
men is about 59 years. Compare the life
absolute difference = compared value – reference value
expectancy of American men to that of
¬ The Relative difference describes the size of the Russian men in absolute and relative terms.
absolute difference in comparison to the reference
value
compared value − reference value
Relative difference = x 100%
reference value

21 22

Of vs More Than (or Less Than) Example: Of versus More Than

(p 70)
¬ If the new or compared value is P% more ¬ A subtlety in dealing with percentage statements comes from
the way they are worded. Consider a population that triples in
than the reference value, then it is (100+P)% size from 200 to 600. There are two equivalent ways to state
of the reference value this change with percentages:
” Using more than: The new population is 200% more than the
¬ If the new or compared value is P% less than original population. Here we are looking at the relative change in
the reference value, then it is (100-P)% of the the population
reference value Relative change=(600-200)/200*100%=200%
” Using of: The new population is 300% of the original population,
which means it is three times the original population. Here we
are looking at the ratio of the new population to the original
population
Ratio=600/200=3=300%

23 24
Example 4 World Population (p 71) Index Numbers
¬ Index numbers provide a simple way to
¬ We found that world population in 2000 was compare measurements made at different
about 131% more than world population in times or in different places. The value at one
1950. Express this change with an “of” particular time (or place) must be chosen as
statement. the reference value (or base value).
¬ The index number for any other time (or
¬ A store is having a “25% off” sale. How does place) is
a sale price compare to an original price?
value
Index number = x 100
reference value

25 26

Examples 1 & 2 Gas Price Index (p 76) Change of base year of an index
Average Gasoline Prices (per gallon) (Source: US Department of Energy) numbers
Year Price Price as a percentage of 1975 price Price index (1975=100)
1955 29.1¢ 51.3% 51.3 The base period of an index number can be changed by the
1965 31.2¢ 55.0% 55.0 following formula:
1975 56.7¢ 100.0% 100.0 Index number for period Cbase period A
Index number for period Cbase periodB =
1985 119.6¢ 210.9% 210.9 Index number for period Bbase period A
1995 120.5¢ 212.5% 212.5 Example:
2000 155.0¢ 273.4% 273.4
Using May 2005 as base, the CPI for July 2008 is 104.3; for July
2005 231.0¢ 407.4% 407.4
1998 is 113.6.
¬ Suppose the cost of gasoline today is $3.20 per gallon. Using the 1975 price as the
reference value, find the price index number for gasoline today So using July 1998 as base,
¬ How much more expensive gas was in 1995 than in 1965? the CPI for July 2008 is 104.3/113.6 = 91.8.
¬ Suppose that it cost $7.00 to fill your gas tank in 1975. How much did it cost to buy
the same amount of gas in 2005? That is, the average price 2 months ago is only 91.8% as much as
¬ Suppose that it cost $20.00 to fill your gas tank in 1995. How much did it cost to buy average price 10 years ago.
the same amount of gas in 1955?

27 28
Example 3 CPI Changes (p 78) Some Common Indexes

¬ Consumer Price Index (CPI) Changes ¬ Consumer Price Index

Suppose you needed $30,000 to maintain a ” www.censtatd.gov.hk
particular standard of living in 2000. How much ¬ Gross Domestic Production Index
would you needed in 2006 to maintain the same ” GDP.pdf from Blackboard
living standard? Assume that the average price of
your typical purchases has risen at the same rate as ¬ Gini Coefficient and Lorenz Curve
the CPI, and CPIs for 2000 and 2006 with base of ” Gini Coefficient and Lorenz Curve.pdf from Blackboard
1983 are 172.2 and 201.6 ¬ Centa-City Index and Hang Seng Index
” CCI and HIS.pdf from Blackboard

29 30