Strategic Management: Quantitative Measurement

Concepts in Business Science

• A theory gives a clear explanation of individual concepts, playing a central role in management research.
• Concepts represent generalized abstractions, providing a general idea associated with an understanding of certain concepts.
• Concepts are hypothetical, meaning they are only descriptions and ideas of reality that exist only in the minds of people and researchers.

Quantitative Measurement

• Quantitative measurement involves assigning values (numbers) to characteristics of objects or subjects of measurement according to certain rules.
• The value determines different levels of characteristics of individual measurement subjects, objects, or events, or sometimes if the subject or object has a specific characteristic.
• In social sciences, many characteristics are not directly measurable (they are purely theoretical concepts).

Basic Measuring Scales

• Nominal scale
• Ordinal scale
• Interval scale
• Proportional scale

Example: Measuring Satisfaction

• Measuring satisfaction can involve one or more questions using a Likert scale with several questions, such as:
  • I'm pleased to visit this hotel.
  • Visiting this hotel exceeded my expectations.
  • I will recommend this hotel to my friends and acquaintances.

Terminology

• Construct: a conceptual term used to describe theoretical phenomena (concepts).
• Constructs can be simple (have one dimension) or assembled (have several components or dimensions).
• In measurement models, constructs are latent variables (usually several) and are not measured directly.
• Indicators (variables, items) are observed scores (usually quantified) and represent manifest variables, which are measured directly.

Characteristics of Reflective Constructs

• Constructs affect indicators.
• Indicators of the same construct should be as interconnected as possible.
• Indicators of different constructs should not be related to each other.
• Omission of indicators is not problematic, and indicators may be excluded.

A Good Measurement Scale

• Valid: measures what is consistent with the definition of the concept.
• Reliable: consistent with repeated measurements and internally consistent.
• Generalizable: can be used repeatedly for similar concepts.
• Parsimonious: shorter and simpler measuring scales are preferred.

Content Validity

• Assessed before data collection, with a prerequisite of a good literature review.
• Represents the level to which items measure what we want to measure.
• The various items of the scale cover all aspects of the characteristic we want to measure.
• Each item has a logical relationship with the concept we measure.

Convergent Validity

• Tested on test or final sample data.
• Measures the relationships (level of correlation) between items that represent the same construct.
• Items that are related from the point of view of the theory should also be related in practice.
• Methods for assessing convergent validity include Pearson correlation coefficient, exploratory factor analysis (EFA), and confirmatory factor analysis (CFA).

Discriminant Validity

• Tested on test or final sample data.
• Measures the relationship (level of correlation) between questions that represent different constructs.
• The relationships between items measuring different constructs must be weaker than the relationships between items measuring the same constructs.
• Methods for verifying discriminant validity include Pearson correlation coefficient and confirmatory factor analysis (CFA).

Nomological Validity

• Represents the constructs that are measures and the relationships between constructs.
• The relationships between constructs must be consistent with the theory.
• Nomological validity is tested by hypothesis testing.

Reliability

• Possible reliability testing methods include double testing and internal consistency testing.
• Internal consistency testing involves measuring the internal consistency of a set of items with the same measurement scale using the Cronbach alpha coefficient.

Quantitative Measurement

• Quantitative measurement assigns values (numbers) to characteristics of objects or subjects of measurement according to certain rules.
• The value determines different levels of characteristics of individual measurement subjects, objects, or events, or if the subject or object has a specific characteristic.

Basic Measuring Scales

• Nominal scale
• Ordinal scale
• Interval scale
• Proportional scale

Measurement Definitions

• A concept is a generalized abstraction that represents a general idea associated with an understanding of certain concepts.
• Concepts are hypothetical, meaning they are only descriptions and ideas of reality and exist only in the minds of people and researchers.
• Constructs are conceptual terms used to describe theoretical phenomena (concepts) and can be simple (one-dimensional) or assembled (with several components or dimensions).

Measurement Models

• Constructs are latent variables (not measured directly) and are represented by indicators (variables, items) which are observed scores (usually quantified).
• Indicators represent manifest variables and are measured directly.

Characteristics of Reflective Constructs

• Constructs affect indicators.
• Indicators of the same construct should be as interconnected as possible.
• Indicators of different constructs should not be related to each other.
• Omission of indicators is not problematic, and indicators may be excluded.

Good Measurement Scale

• Valid: measures what is consistent with the definition of the concept.
• Reliable: consistent with repeated measurements.
• Generalizable: can be used repeatedly for similar concepts.
• Parsimonious: shorter and simpler measuring scales are preferred.

Content Validity

• Assessed before data collection.
• Prerequisite: good literature review.
• Represents the level to which items measure what we want to measure.
• Items should cover all aspects of the characteristic we want to measure.
• Each item has a logical relationship with the concept we measure.

Convergent Validity

• Tested on test or final sample data.
• Measures the relationships (level of correlation) between items that represent the same construct.
• Items measuring the same constructs must be strongly interrelated.
• Methods for assessing convergent validity: Pearson correlation coefficient, exploratory factor analysis (EFA), and confirmatory factor analysis (CFA).

Discriminant Validity

• Tested on test or final sample data.
• Measures the relationship (level of correlation) between questions that represent different constructs.
• The relationships between items measuring different constructs must be weaker than the relationships between items measuring the same constructs.
• Methods for verifying discriminant validity: Pearson correlation coefficient and confirmatory factor analysis (CFA).

Nomological Validity

• The nomological network represents the constructs that are measures and the relationships between constructs.
• The relationships between constructs must be consistent with the theory.
• Nomological validity is tested by hypothesis testing.

Reliability

• Possible reliability testing: double testing and internal consistency testing.
• Internal consistency testing: a measure of the internal consistency of a set of items with the same measurement scale, such as Cronbach alpha coefficient.