Ch13 Zikmund Measurement and Scale

Business Research Methods

William G. Zikmund

Chapter 13:

Measurement & Scaling Concepts

Measurement in Survey ResearchMeasurement is the process of assigning numbers numbers or labelslabels to the attributes of objects or persons or states, or events under study in accordance with specific rulesrules

Measurement is the assignment of numbers to objects or events according to rules.

Measurement in Survey Research

• The numbers don’t mean anything until we interpret them• Statistics helps us in interpretation.

For example consider the people in this class as subjects and their height as the attribute.

The attribute height varies between objects, hence attributes are more collectively known as variables

Identify Concept / ConstructMeasurement begins by identifying a

concept of interest .

A concept is expressed in every-day terminology. This requires the researcher to generalize/categorize.

A concept also known as construct can be a theoretical abstraction that can’t really be observed (e.g., love, trust, social class, personality, power).

FOR EXAMPLE: What do sociologists want to measure?• Concepts or constructs –like

– e.g. social class– e.g. social capital– e.g. social attitudes

•But these aforesaid concepts can’t be directly observed, so how can we measure them?

• We use indicators that represent our concepts and measure those indicators (variables)E.g. to measure the concept social class, we use indicator Income

Identify Concept / Construct

‘objective’ attributes like level of education or traininglevel of education or training are not abstract concept.

• We can’t directly observe how educated someone is. So we use an indicator, e.g.

–Ask someone for their highest educational qualification (this is the variable under study)

– Or ‘how old were you when you completed fulltime education?’ (this is the variable under study)

Identify Concept / Construct

Concept

Concept which are concrete & easy to measure, such as:age, gender and number of children etc.

Concept which are abstract and difficult to measure, such as:Brand loyalty, job involvement etc.

For Example: Brand loyalty can be measured using a number factor such as,

attitude toward brand number of different brand purchases etc.

Concept: Operational DefinitionOperational Definition of concept specifies which observable characteristics will be measured and the process for assigning a value to the concept.

For example: abstract concept like grievances is difficult to operationalize.

Whereas, concrete concept like personnel turnover is less difficult to operationalize.

Concept: Operational Definition

ATTRIBUTES is a single characteristic or fundamental feature that is relevant to an object, person, or issue

people in this class are subjects and their height is the attribute

COMPOSITE MEASURE is a composite measure of several variables to measure a single concept; a multi-item instrument/questionnaire

To measure health, you can measure attributes like weight, Body Mass Index, Undergone diseases or medications.

VariablesWhen we measure the attributes of an object, we

obtain a value that varies between objects or subjects of study.

For example consider the people in this class as

objects and their height as the attribute

The attribute height varies between objects, hence attributes are more collectively known as variables

Variables can be measured on four different scales

Develop a Measurement Scale

SCALE– A scale is a set of symbols or numbers so

constructed that the symbols or numbers can be assigned by a rule for the individuals (or their behaviors or attitudes) to whom the scale is applied

Types of Measurement Scales

Nominal ScaleOrdinal ScaleInterval ScaleRatio Scale

Types of Measurement Scalesa) Nominal or Classificatory Scales

It is a Scale that categorize data into mutually exclusive and collectively exhaustive categories.

A nominal scale enables the classification of individuals, objects or responses based on a common/shared property or characteristic.

Typical Descriptive Statistics used for interpretation: is Frequency counts, percentages/modes.

Examples of Nominal Scales:

Gender Social status

Marital status Days of the week (months)

Geographic location Patrons per hour

Ethnic Group Types of restaurants

Brand choice Religion

Job Type: Executive, Technical, Clerical

a) Nominal or Classificatory Scales Examples

Coded as “1”

Coded as “2”

b) Ordinal or Ranking Scale

It is a Nominal scale that can order data. A ordinal scale enables the classification of

individuals, objects or responses based on a common/shared property or characteristic and it ranks the subgroups in a certain order.

Attributes / Characteristics can be rank ordered.

For example: for a variable ‘educational attainment’ following attributes/properties are ranked.

1=at least SSC2=at least HSC3=undergraduate degree4=postgraduate degree

b) Ordinal or Ranking Scales Someone with a postgrad has reached a higher educational than someone with only an undergrad.

• But distances between attributes have no meaning; no arithmetical comparisons or calculations can be applied to them Is the difference between SSC and HSC is same as between undergraduate and postgraduate degrees? No!

• Particular values (numbers) used as codes is arbitrary, but they must be in the correct order.

Typical Descriptive Statistics for interpretation: is Median

With ordinal data, it is fair to say that one response is greater or less than another.

E.g. if people were asked to rate the hotness of 3 chili peppers, a scale of "hot", "hotter" and "hottest" could be used. Values of "1" for "hot", "2" for "hotter" and "3" for "hottest" could be assigned.

b) Ordinal Scale: Example

The gap between the items is unspecified.

b) Ordinal or Ranking ScalesExample of Ordinal Scale: Please rank the following fax machines from 1 to 5 with 1 being the most preferred and 5 the least preferred.

_____ Panasonic _____ Toshiba _____ Sharp _____ Savin _____ Ricoh

It has all the properties of nominal and ordinal scales plus equal intervals between consecutive points to show relative amounts. It is a preferred measure for complex concepts or constructs

i.e. gaps between whole numbers on the scale are equal.

e.g. Fahrenheit and Celsius temperature scales

An interval scale does not have to have a true zero. e.g. A temperature of "zero" does not mean that there is no temperature...it is just an arbitrary zero point.

Statistics for interpretation: is frequencies, mode, median, mean, standard deviation

c) Interval Scale

c) Interval Scales:

Similar to interval scales except that the ratio scale has a true zero value.

e.g. the time something takes

allows you to compare differences between numbers.

If a train journey takes 2 hr , then this is half as long as a journey which takes 4 hr.

d) Ratio Scale

d) Ratio ScalesIt incorporates all the properties of nominal, ordinal, and interval scales It includes a meaningful zero point so that magnitudes can be compared arithmetically.The zero point of a ratio scale is fixed, which means it has a fixed starting point.

Typical Descriptive Statistics for interpretation: is Mean/variance plus a few higher order statistics

Example: Age– It makes sense to say that someone who is 40 years old is twice as old as someone who is 20 years old

0

1

2

3

4

5

6

7

Examples of Ratio Scaleheight, weight, age,

Length

time

Income

Market share

Use of Measurement Scales• Nominal

– Used to categorize objects

• Ordinal– Used to define ordered relationships

• Interval– Used to rank objects such that the magnitude of the

difference between two objects can be determined

• Ratio– Same as interval scale but has an absolute zero point

Nominal

Ordinal

Interval

Ratio

Win Place Show

1 length 2 lengths

40 to 1 long-shot pays $40

Error

Why do we care?

Characteristics of Good Measurement Scales1. Reliability

• The degree to which a measure accurately captures an individual’s true outcome without error; Accuracy

synonymous with repetitive consistency

2. Validity• The degree to which a measure faithfully represents the

underlying concept; Fidelity

3. Sensitivity• The ability to distinguish meaningful differences

between attitudes. The more categories the more sensitive (but less reliable)

4. Generalizability• How easy is scale to administer and interpret