Consumer Psychology

Chapter one: Correlation and Causation

How do we “know” something? People will say learn from class,books or experiences etc.

Today I am going to share with you “knowing” something by OBSERVATION-THEORY-HYPOTHESIS-FALSIFIABILITY（观察-理论-猜想-证伪，简称猜想与反驳）Procedure.(OTHF) For instance, you observed that the Heartthrob bar is always busy on Thursdaynights when they offer half price burgers. Then you may conclude that Heartthrobis busy on Thursday BECAUSE it offers half-price burgers. Wait, hold on. Isthat true? We could make a HYPOTHESIS here first: Heartthrob won’t be busy on Thursdaysif we remove the promotion. Then we falsify the Hypothesis by 2 ways:

1.If we remove the promotion, is it still busy on Thursdays?

2.If we switch to Friday nights is it going to be busy too?

If it could be disapproved, then we could not induce thecausation between the promotion and business, it maybe just correlated.

So, what is difference between correlation and causation.Correlation means the relationship between things that happen or changetogether, simply put: “X and Y change together”. Causation means “Change in X CausesChange in Y”. (X=Independent Variables/ the Cause; Y=Dependent Variable/the Effect).Correlation alone doesn’t indicate a causal relationship, but causation means Xand Y are correlated.

Now let’s focus on causation. There are 3 necessaryconditions for causation:

1.Correlation: X and Y “change together”

2.Temporal Antecedence: X happens before Y

3.There is no ‘Confounded Factor’ (简称CF、混淆变量)whichdrives both X and Y, X and Y are spuriously correlated because of the CF

·CF drives both X and Y. In thefollowing example, we know that it seems it is smoking cigarettes causes thediminished work performance, but we could find a third variable or theconfounded factor STRESS which affects both. Therefore, smoking and workperformance has not causal relationship.

·CF causes a spurious correlation.Here we assume sunrise causes our hunger, but the real reason is we haven’teaten for 12 hours. Therefore, CF cause a spurious correlation.

In the above-mentioned contents,it is the simple causation: X causes Y, we could also have multiple causes,like air pollution is caused by cars emission/manufacturing/cows and sheep etc.

Since we already have a clearunderstanding of correlation and causation, let’s focus on ExperimentalResearch: how to design an experiment and disapprove a hypothesis.

1.Identify IV and DV, making sure IV causes change in DV.

2.Random Sampling & Random Assignment.

·Random sampling is “chooseparticipants randomly from the entire population. It allows generalization tothe population;

·Random assignment is “the membersof your sample are randomly assigned to different groups or conditions.

3.Between-subjects Designs: different groups or conditions are created. (e.g.experimental vs. control)

Now we could design a BASEHYPOTHESIS:E.g. on average,people will be less likely to buy a cold soft drink if it costs $2 vs. $1.

However, there is always interaction involved.

4.Interactions: an interaction occurs when the effect of IV on the DVchanges because of another IV. (example will be given later)

5.The classic 2X2 Between-Subjects Design with 2 factors(IV) and 2 levels(IV levels), like Mood as a factor, its levels could be “happy vs. sad”. Allthe LEVELS must meet 3 requirements:

·Clearly defined, for time level,we could not say ‘long vs. short’

·Mutually exclusive, happy vs. sad

·Different enough. $100 vs. $5000

Now we coulddesign an INTERACTION HYPOTHESIS:

E.g. on average, people will be less likely to buya cold soft drink if it costs $2 vs. $1.but only when it is on cold days. Onhot days, the higher price will not reduce the likelihood of purchase.

Marketing Implications:

1.“Know” something by Observation is not enough, we

should follow steps of Observation-Theory-Hypothesis-Falsifiability.

2.Design Interaction Hypothesis Experiment with

randomly sampling & random assignment, to scientifically conclude variables’

relationship and projecting trends.

3.Marketing is a science and art, use research and

data to find consumers’ psychology behind and apply it to reality.

(Keep writing and thinking:)