Casual Research

Exploratory

Methods: case studies, focus groups, qualitative research Characteristics: flexible, versatile, but not conclusive Useful for: discovery of ideas and insights

Causal

Methods: experiments Characteristics: manipulation and control of variables Useful for: determine cause and effect relationships

Descriptive

Methods: surveys, panels, scanner data Characteristics: preplanned, structured, conclusive Useful for: describe market characteristics or functions

Research Design Types

Exploratory Descriptive Casual

Simpson's Paradox

A trend that appears in different groups of data disappears or reverses when these groups are combined

True Experimental Design

Always involves: - Random assignment to groups - At least one manipulated variable - At least one measured variable Randomization - Random assignment of test units to groups - Ensures prior equality of experimental group

Independent Variables

Controlled Monitored Manipulated

Nature of Experimentation

Experimentation involves the manipulation of one or more variables (cause, treatment) by the experimenter in such a way that its effect on one or more other variables can be measured.

Experiments

Experiments provide insight into cause-and-effect by demonstrating what outcome occurs when a particular factor is manipulated

Effects of causes

Forward causal questions - What might happen if we do X? Examples: • The effect of smoking on health • The effect of schooling on earnings • The effect of campaigns on election outcomes?

Test/Experiment/Treatment group (EG)

Group receiving treatment

Control group: (CG):

Group without treatment

Treatment

Manipulated variables

Dependent Variables

Measured

Reliability

Observed Score (O) vs. True Score (T): O = T + Error*systematic + Error*random • Extent to which we can be certain that the observed score is free from random error. • "Will I get the same result if I measure again?"

Measurement

Recording of a response

Causes of effects

Reverse causal inference - What causes Y? • Why do more attractive people earn more money? • Why does per capita income vary some much by country? • Why did the economy collapse?

Establishing Causality

Three elements of evidence for causality • Concomitant variation • Precedence • Systematic elimination of other possible factors

Empirical Causality

Two Type of Data: Observational Data - Observe subjects and their data of interest without assigning treatments Experimental Data - • First, apply treatments to subjects, • Then, proceed to observe their data • There is time sequence of treatment and data observation

Goals of Marketing Research

Two different types of goals: • Causal effects Casuality) are answers to 'what if' questions: - What would happen to our sales if we * Advertise online? * Offer a discount? • Forecasting - Only need observed independent variables for best prediction - Don't worry about causality

Validity

Two principal goals: Internal Validity External Validity

Field Experiments

• 'Natural' setting • Within the routine of ordinary life • Difficult to control variance of other independent variables • Respondents not aware of being in a test e.g. exposure to a new product in a supermarket

External Validity

• Ability to draw conclusions about a larger population. • Refers to the extent to which the results of the experiment can be generalized from the experimental environment to real world • "Can I generalize the result?"

Internal Validity

• Ability to draw conclusions about variable being measured • Refers to the ability of the experiment to unambiguously show a cause and effect relationship, • "Am I measuring what I think I am measuring?"

Experiments in Marketing

• Aims to conduct experiments in marketing - Field experiments - Test markets - Lab experiments • Key question: "When have we identified a true causal effect?" • Focus on experiments that are possible to execute and then make additional assumptions that - if valid - guarantees that we have uncovered a causal effect • Usually take the form of a comparison between a test and (at least one) control group • Experiments are usually run as field experiments

Lab Experiments

• Artificial setting • Isolated from the routine of ordinary life • Variance of other independent variables kept to a minimum • Respondents aware of being tested - testing effect e.g. exposure to a new product in a simulated supermarket

Causal Research

• Obtain evidence of cause-and-effect relationships - Identify cause and effect of a phenomenon - A change in one variable will produce a change in another variable • Highly structured design - Manipulation of one or more independent variables - Control other variables • Usually takes the form of experiments - Lab experiments - Field experiments - Test markets

Experiments and Evidence of Causality

• We can use: - lab experiment - field experiment • In experiments, we can easily control some conditions and factors • Experiment design can help eliminate other possible extraneous variables