13 - Data Marketing Experiments

Question 1

What is a marketing experiment?

Answer 1

Type of marketing research
eg: for data

Question 2

Three types of marketing research

Answer 2

Exploratory (insights)
Descriptive (market characteristics)
Causal (cause and effect)

Question 3

4 conditions for establishing causality

Answer 3

Association
Temporal precedence
Eliminating other causes
Mechanism between cause and effect

Need all four to determine causation

Question 4

Association

Answer 4

A statistical relationship must exist between two variables
eg: 5% decrease in price = 20% increase in sales

Question 5

Temporal precedence

Answer 5

The order in which the cause and effect variable occurs
eg: a 5% decrease in price must happen before noting the 20% increase in sales

Question 6

Mechanism between cause and effect

Answer 6

Identify a plausible mechanism that explains how the cause variable leads to the effect variable, such as science
eg: washing hands = less bacteria = less vulnerability to becoming ill

Question 7

History (other causes - causality)

Answer 7

Refers to any variables or events, other than the one manipulated by experimenter that occur between the pre- and post-measures and affect the value of the dependent variable
eg: 20% increase in sales may not be just affected by 5% discount, if say an alternative was recalled, this could affect the sales of the said offering

Question 8

Maturation (other causes - causality)

Answer 8

Any economic, biological, or psychological processes that systematically vary with passage of time
eg: sales people get more experienced and thus get improved performance

Question 9

Internal Validity in experiements

Answer 9

Ability to conclude that cause = effect (4 conditions)

Question 10

External validity in experiments

Answer 10

Ability to generalize cause = effect to a broader population
External validity is said to be high if:
Random sampling of participants from the population
Using the right sample frame that covers the population

Question 11

Threats to external validity

Answer 11

Surrogate situation
Measurement of dependent variable
Measurement timing
Sample selection

Question 12

Surrogate situation (threats to external validity)

Answer 12

Occurs when the experimented task is so “contrived” that it cannot be a reasonable proxy of reality
eg: forced attention to a single ad in copy testing experiments vs browsing a magazine

Question 13

Measurement of dependent variable (threats to external validity)

Answer 13

Occurs when the dependent variable does not “map on” to the behaviour or construct of interest
eg: use of SAT scores to measure intelligence

Question 14

Measurement timing (threats to external validity)

Answer 14

When pre- or post-measurements are made at an inappropriate time to indicate the effect of the experimental treatment
eg: there needs to be sufficient time between exposure and purchase intentions

Question 15

Sample selection (threats to external validity)

Answer 15

The sample of respondents may not be representative of the population of interest
eg: people are paid 25$ apple gift cards for a samsung survey