Psychology of Measurement and Attitude

Measurement Attitude

• Attitudes are mental states that structure how individuals perceive and respond to their environment.
• Attitudes are used to predict behavior.
• Attitudes are psychological constructs, meaning they cannot be directly observed or measured, but inferred.

Components of Attitude

• Attitudes are measured to understand and influence behavior.
• Measuring attitudes is more accurate than observing and interpreting behavior.
• Measurement provides a capacity for diagnosis and explanation.
• Measurement helps understand the strengths and weaknesses of competitive alternatives.

Concept of Measurement

• Measurement is a standardized process of assigning numbers or symbols to quantify characteristics of objects.
• Measurement uses pre-specified rules to represent quantities or qualities.

Concept of Scaling

• Scaling creates a continuum to locate objects based on their measured characteristics.

Examples of Measurement & Scaling

• Weather Temperature Measurement:
  • Object: Weather
  • Attribute: Temperature
  • Measurement: Kelvin, Celsius, Fahrenheit
• Consumer Intent to Buy Measurement:
  • Object: Consumer
  • Attribute: Intent to Buy
  • Measurement: Likelihood to Buy Scale, Willingness to Buy

How To Capture Attitudes? Nominal Scale

• Assigns objects to mutually exclusive categories.
• No ordering or spacing is implied.
• Only arithmetic operation is counting.
• Example: Categorizing consumers by age group.

Ordinal Scale

• Ranks objects by a common variable.
• Does not indicate the difference between objects.
• Arithmetic operations are limited (e.g., median or mode).
• Example: Ranking brands based on customer preference.

Interval Scale

• Numbers represent equal increments of the attribute being measured.
• Differences between objects can be compared.
• All common statistical operations can be used.
• Example: Measuring temperature on a Celsius scale.

Ratio Scale

• Naturally defined zero point.
• Objects can be compared in terms of absolute magnitude.
• Only scale allowing comparisons of absolute magnitude.
• Example: Measuring weight in kilograms.

Summary of Measurement Scales

• Nominal: Categorization (gender, location, social class).
• Ordinal: Rankings (preference, class standing).
• Interval: Index numbers, attitude measures.
• Ratio: Sales, income, age, quantities.

Classification of Attitude Scales

Attitude Rating Scales

• Rating scales offer a continuum of numbered categories representing attitude levels.
• Two types: Single-item scales and Multiple-item scales.

Single-Item vs Multiple-Item Scale Example

• Example (Measuring Loyalty):
  • Single-item: How likely are you to recommend our brand?
  • Multiple-item:
    • Item 1: How likely are you to come back to our store in the future?
    • Item 2: How likely are you to purchase our products in the future?
    • Item 3: How likely are you to recommend our brand?

Classification of Single Item Scales

Item-Category Scale

• Most widely used single-item scale.
• Respondent selects from a limited number of categories.

Pictorial Scale

• Categories are depicted pictorially.
• Commonly used types of pictures: thermometers and funny faces.
• Format must be understandable and allow accurate responses.

Rank-Order Scale

• Respondents compare and rank two or more items.
• Requires knowledge of all options for accurate comparison.
• Random ordering can reduce bias.
• Limitations: Requires attention and effort, not suitable for more than 5 objects.

Paired-Comparison Scale

• Respondents choose one of two items based on a specific criterion.
• Advantages: Easier to select than rank, removes bias from ordering.
• Disadvantages: Number of paired combinations increases geometrically (for n objects, n(n-1)/2 comparisons are required).

Constant-Sum Scale

• Respondents divide a given number of points (typically 100) among attributes based on their relative importance.
• Advantages: Relative distance between rankings can be determined, both rank and relative magnitude can be compared.
• Disadvantages: Respondents may struggle to allocate points accurately for a larger number of categories.

Designing Single Item Scale

• Number of scale categories: Depends on the format of the interview/survey and the nature of the object.
• Strength of the anchors: Stronger anchors reduce use of extreme values.
• Balance of the scale: Balanced scales provide an easy escape for respondents when asked to take a side.
• Labeling of the categories: Labeling all categories reduces ambiguity, but may be taxing for the respondent.

Multiple Item Scales

• Used to measure attitudes toward complex objects with many facets.
• Developed to measure a sample of beliefs about an attitude object and combine the answers into an average/sum score.
• Detects attributes determining consumer's overall attitudes.

Classification of Multiple-Item Scales

• Likert scale: Respondents indicate level of agreement/disagreement with statements, expressing favorable or unfavorable attitude.
• Thurstone scales: Compensates for limitations of Likert scale by accounting for the strength of individual items in computing the attitude score.
• Semantic differential scales: Describes a set of beliefs comprising a person's image of an organization or brand.

Semantic Differential Scales

• Examines the strength and weaknesses of a concept described by several attributes.
• Respondents rate each attribute on a scale of opposing attributes.
• Usually have 5 or 7 points, allowing for a neutral position.
• Advantages: Efficient way to examine brand personality, attitudes toward advertisements, celebrities, or retail stores.
• Disadvantages: Halo effect (overall impression dominates the rating), Ordering effect (location of previous judgments influences subsequent judgments).

Likert Scales

• Respondents express a level of agreement/disagreement with statements.
• Each item measures some aspect of a single common factor.
• Usually 5, 7, or 11 point scales, producing interval data.
• Include both reversed and non-reversed items.
• Advantages: Scores on individual items can be summed up/averaged to produce a total score for the respondent.
• Disadvantages: Favorable/unfavorable attitude toward each item has equal weight, does not provide information on the more influential items.

Thurstone Scales

• Compensates for the limitation of the Likert scale by taking into account the strength of individual items.
• Two-stage process:
  • Stage 1: Subject matter experts determine final scale items and corresponding weights.
  • Stage 2: Collect ratings from target respondents.
• Advantages: Easy to administer, requires minimal instructions.
• Disadvantages: Time-consuming and expensive to construct, scale values and weights may depend on the attitudes of the original judges, complex scoring system.

Designing Multiple Item Scales

• Develop a clear conceptual framework or purpose for the scale.
• Select relevant and representative items to measure the attitude.
• Ensure items are clear, concise, and unambiguous.
• Use a balanced number of positive and negative items.
• Choose an appropriate scale format (e.g., 5-point, 7-point Likert).
• Conduct a pilot test to assess the reliability and validity of the scale.

Accuracy of Attitude Measurement

• Validity: Measures what it is supposed to.
• Reliability: Consistent results across different comparable populations.
• Sensitivity: Discriminate between respondents with opposed attitudes.
• Generalizability: Applicable to different research settings or situations.
• Relevancy: Relevancy = Reliability x Validity

Can Scales be Generalized Across Countries?

• Scales may be affected if not properly adjusted for different cultures.
• Factors influencing responses across countries:
  • Literacy and education levels.
  • Culture (semantic differential scale is closest to pan-cultural).
  • Adapted response formats and calibration for specific countries and cultures.

Scales Across the Globe

• Responses can be affected by:
  • Low literacy and educational levels.
  • Culture (semantic differential scale is closest to pan cultural scale).
  • Adapted response formats, particularly their calibration, for specific countries and cultures.