Sie sind auf Seite 1von 14

Fundamentals of Measurement

by

Michael Everton (mxe06u)


Fundamentals of Measurement
• Is part of many fields and subject areas.
• My example is based on ‘Measurement’ in conjunction with
statistics.
• The presentation will cover:
--A brief explanation of measurement and how it is used.
--A explanation of statistics.
--How measurement data is broken down into several
types or scales of measurement.
--The rules used to establish a consistent understanding
of measurement.
Measurement?
The word ‘Measurement’ as used in daily
life, implies the assignment of an exact
and quantitative number to an object, such
as £2, 4 metres, 5 amps or 70 MPH.
These quantitative terms by themselves
are of limited use.
After measurements are taken they need
to be made meaningful by relating them to
other variables.
Measurement:
Scientists use words and
numbers to
communicate the results
of their research.
Statistics are numbers
and results of tests
conducted on sets of
numbers.
Statistics:
• The term statistics is used to describe numerical
data used in research, reports or presentations
Examples: -The number of clients served in a
space of time e.g. an hour, day,
week, month.
-Performance ratios.
-Age and gender of a village
population etc...
• Statistics are also used to define many
mathematical techniques and procedures used
to collect, describe, analyse and interpret data
Statistics:
• In general statistics can be considered as both
numerical data and a variety of tools and
techniques used to process raw data to make it
more meaningful.
• There are two types of Statistics:
Descriptive statistics: used to summarise a larger
set of numbers called a dataset.
Inferential statistics: these are measurements of
a smaller group that are used to make
assumptions of a larger group of interest.
Statistics:
• The numerical values in statistics are
measurements taken with some kind of scale or
device.
• These scales provide different levels of
information, based upon the type of data they
intend to capture.
• The four types or levels of data:
Nominal, Ordinal, Interval and Ratio.
Each type has a body of tests appropriate for
that level.
Fundamentals of Measurement
• The key to ensure that everyone
understands the Measurements.
• Measurement and data.
• Variables.
• Data types also known as ‘scales of
measurement’.
• Each level has one or more rules.
Nominal Data
• Nominal data is the least powerful
• Of the four types.
It is only concerned with one rule.
The Rule states that “Different numbers must mean different things”.
In Nominal level data, numbers or labels are used to differentiate
between things.
Once a number has been assigned to a certain category all other
items with the same characteristics must receive the same number.

Typical examples of Nominal or categorical scales are:

• The values of “1” and “2” assigned to the categories of Female and
male.
• The count of times a head is shown when a coin is tossed up.
• Numbers used to denote different types of occupations, university
years etc...
Ordinal Data
• Two rules apply to Ordinal data:
1)“Different numbers must mean
different things”.
2)The second rule which states that “The things being measured
can be ranked or ordered along some dimension”.
Ordinal data can also be referred to as “Ranked data”.
• When things are ordered they are arranged in some logical
sequence, they may have ‘more or less’ of a particular characteristic
than others in a set.
• The primary limitation with Ordinal measurements is that the
numbers seldom state how much ‘more or less’ the difference exists
in two or more collections of data.
• A typical use of Ordinal data is to measure people’s preferences or
rankings for candidates (Political polls).
Interval Data
• The third class of measurement data is
‘equidistant interval’ or more simply known
as “Interval”.
Three rules apply to Interval scale
1)“Different numbers must mean different things“.
2)“ The things being measured can be ranked or ordered along some dimension”.
3)The third rule states that: “The differences between adjacent levels on a scale are
to be equal”.
The key requirement is that a single unit change always measures the same amount
of change in whatever is being measured.
The unit gradations within the scale must be as broad or as fine as needs be.
• For example on a five point scale the distance between “3” and “4” or “4” and “5”
might be measured in tenths , hundreds, thousands or even finer but they must apply
to every part of the scale equally.
• There are limitations to the information provided by Interval data.
• For example we cannot say that 100 degrees Fahrenheit is exactly twice as warm as
50 degrees Fahrenheit, nor can we say that 35 degrees is half as warm as 70
degrees. We can only say that the differences between the two single points on the
scale are equal.
Ratio Data
• Four rules apply to
Ratio scales.
1)Different numbers must still mean different things.
2)The data can be ranked or ordered along some
dimension.
3)The intervals between adjacent points must be equal.
4)The fourth rule: “The measurement scale must have
an absolute or fixed zero point”.
• Typical examples of Ratio scales are time, distance,
mass or temperature.
References:
1. Downie, N.M “Fundamentals of
measurement” Second Edition Published
by Oxford University press (1967) .
2. McNabb, D.E “Research methods for
Political Science” Published by M.E
Sharpe (2004).
Questions?

Das könnte Ihnen auch gefallen