Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 20

THE BASICS OF

MEASUREMENT
Topics Covered

 Introduction
 The Representational Theory of Measurement
 Measurement and Models
 Measurement scales and scale types
 Meaningfulness in measurement
Introduction

 How much must we know about an attribute before it is reasonable to


consider measuring it?
 How do we know if we have really measured the attribute we wanted
to measure?
 Using measurement, what meaningful statements can we make about
an attribute and the entities that possess it?
 What meaningful operations can we perform on measures?

The answers to above questions lies in Basics of Theory of


Measurement.
11/06/2024
THE REPRESENTATIONAL THEORY
OF MEASUREMENT

11/06/2024
Empirical Relations

 There is a reasonable consensus


about which pairs are in the
relation.
 These can be unary, binary, ternary
or n-ary.
 Define a mathematical mapping
that preserves the relationship.
 Measurement is a mapping from
the empirical world to the formal
(relational world)
 A measure is the number or symbol
assigned to an entity by this
mapping in order to characterize an
attribute.
11/06/2024
The Rules of Mapping

 We map some real world as the domain of mapping to a mathematical world as the
range of mapping.
Example- height of a person -> inches/centimetres/metres
 Rules for performing the mapping must be specified.

Example- shoes allowed or not/height of hair considered or not


 The mapping rules must be defined explicitly.

Example- while counting LOC, whether to include or exclude comments, blank


l lines etc..

11/06/2024
The representation condition of measurement

 Each relation in empirical relational system corresponds via the measurement to an


element in a number system.
 Representation condition asserts that a measurement mapping M must map entities
into numbers and empirical relations into numerical relations in such a way that the
empirical relations preserve and are preserved by numerical relations.
 The empirical relation “taller than” is mapped to the numerical relation “ >” with
representation condition ,
A is taller than B iff M(A) > M(B) for any M

UNARY is tall X > 70


BINARY is taller than X > Y
TERNARY x and y are combinely X+Y > Z
taller than
11/06/2024
Key Stages of formal measurement
Representation
Condition

Any measure that satisfies the representation condition is a valid measure.


The richer(large number of relations can be defined) the empirical relation system, the
fewer the valid measures.

11/06/2024
Examples of specific measures used in software engineering

11/06/2024
MEASUREMENT AND MODELS

11/06/2024
 Model – An abstraction of reality, allowing us to strip away detail and
view an entity or concept from a particular perspective.
Example- Cost Models permit us to examine only those project
aspects that contribute to the project’s final cost.
 Models come in many different forms: as equations, mappings or
diagrams.
 These show us how the component parts relate to one another, so
that we can examine and understand these relationships and make
judgements about them.

11/06/2024
Defining Attributes

 We must focus more on empirical world rather than formal, mathematical aspect.
Reality is reverse.

11/06/2024
Direct and Indirect Measures

 Direct Measure : Involves no other


attribute or entity.
Examples-
i. Length of source code,
ii. Duration of testing process,
iii. Number of defects discovered during
testing process,
iv. Time a programmer spends on a
project
 Indirect Measure : Involves other
attribute or entity.

11/06/2024
Measurement for Prediction

 A prediction system involves-

A Model
A set of procedures for determining the model procedures
Procedures for interpreting the results

11/06/2024
MEASUREMENT SCALES AND
SCALE TYPES

11/06/2024
 The measurement mapping M, along with empirical and numerical relation systems
is a measurement scale.
 Few Questions
 How do we determine when one numerical relation system is preferable to another?
 How do we know if a particular empirical relation system has a representation in a given numerical
relation system?
 What do we do when we have several different possible representations in the same numerical
relation system?

Types of Scales
i. Nominal
ii. Ordinal
iii. Interval
iv. Ratio
v. Absolute

11/06/2024
Types of Scales

 Nominal Scale : The elements are placed in different classes based on the
value of attribute. There is no ordering among classes. Distinct numbering is
acceptable here.
 Ordinal Scale : It is a nominal scale with information about ordering of classes
or categories. But this scale does not offer the difference between orders, and
do not use the basic math operations. It represents ranking. Any mapping
should preserve the ordering.
 Interval Scale : It preserves order, as with an ordinal scale. It also preserves
the difference between items, but not ratios between them. We can use addition
and subtraction operations only, and it is used to calculate the mean and
standard deviation.

11/06/2024
Types of Scales

 Ratio Scale : This is the high level of scales & allow more analysis. It
preserves ordering, size of intervals as well as ratios between entities. It has a
zero element that represents total lack of attributes. All arithmetic can be
meaningfully applied here.
 Absolute Scale : It is most restrictive scale. It counts the number of elements
in the entity set and has the form “number of occurences of x in the entity”
There is only one possible measurement mapping i.e. M=M’ i.e. identity
transformation.
 These scales are listed in order of increasing levels of richness i.e. all relations
of first is contained in second and so on..

11/06/2024
11/06/2024
Meaningfulness in Measurement

A statement involving measurement is meaningful if its truth value is invariant of


transformations of allowable scales.

 Statistical operations on measures


 Objective and subjective measures
 Measurement in extended number systems
 Indirect measurement and meaningfulness

11/06/2024

You might also like