Samenvatting
Onderzoeksproject
(Blok 8) Tentamen 2
,Lecture 8 - Introduction to experimental design
Today’s plan
- Causation
- Experimental design
- Developing experimental stimuli
- Experimental validity
Causation
The question you should ask yourself; why do we want to do experiments?
Correlation vs. causation
It is easy to find correlation, it is however more difficult to determine whether the correlation
also indicates causation.
Correlation: a mutual relationship or connection between two or more things.
Causation: the relationship between cause and effect; causality.
VB; zo is er bijvoorbeeld een positieve correlatie tussen het aantal bezoekers aan Disney land
Parijs en de neerslag in Louisiana -> de grafieken lijken op elkaar (correlation kan ook negatief
zijn-> dan zijn de grafieken het tegenovergestelde van elkaar). Maar er is geen causality
tussen deze twee, there is no logic behind this correlation-> meer neerslag in Louisiana hoeft
niet tot minder bezoekers in Disney land te leiden. Het is eerder een coincidence dat de
grafieken zo op elkaar lijken.
“Correlation does not imply causation”
Example: why is shoe size related to reading ability? Is there a causal relationship? Of is het
enkel correlation?
-->Shoe size is related to age (older= grotere maat) en age is related to reading ability. So
there is one possible additional factor behind the correlation. Deze factor (age) is not included
in this relationship. Age is an omitted variable.
Several reasons why X and Y can correlate:
- X causes Y
- Y causes X
- Z causes both X and Y
- Spurious correlation
The Difficulty of Causal Inference
Reverse Causality: X appears to cause Y, but it is actually Y that causes the change of X
and because X can also change Y, Y causes the change of Y.
,Examples: Diversification and profitability. Diversified firms(IV) tend to profit(DV) more. Maar
kan ook zijn dat profitability de diversity beïnvloed want als een bedrijf veel winst heeft, kan
het niet gaan zitten op deze winst. Ze zullen dan in verscheidene projecten investeren en op
deze manier meer diversified worden.
Sales(DV) of a brand of soda are higher during weeks of heavy advertising (IV).
- However, advertising is allocated when the stakes are greatest (e.g., during holidays and
summer months). Thus, anticipated sales(IV) may actually cause advertising(DV).
Third Variable: X appears to cause Y, but both X and Y are actually caused by Z. Z is de
omitted variable → voorbeeld over shoe size and reading ability.
On the average, the more toys a child has, the higher his or her IQ.
- Both the number of toys and IQ may be caused by family resources such as income (i.e.,
better nutrition and education).
On average, students who sit in the front of the class end up with higher grades.
- However, students choose where they sit. Motivation may influence both sitting and grades.
Causation
When can we infer that X causes Y?
--> Three conditions for causality:
Relationship between X and Y--> X and Y vary together
Time order --> X cannot happen after Y
Elimination of other possible causal factors --> All other possible causes held constant or
controlled. VB; income in VB hierboven is de variable die je constant houdt. Hier gaan we
regressie voor gebruiken.
Experimental design
We want to do experiments to find out the causation. How to design experiments?
Basic features of a between-subjects design (van experiments)
• Independent variable that is “manipulated” across groups or “between-subject”.
– Between-subject: one participant is assigned to one experimental condition of the
IV (e.g., ad A or ad B)
–Within-subject: one participant is assigned to several experimental conditions (e.g.,
ad A and ad B)
• Dependent variable that is measured
• Context: laboratory, online survey, field, etc.
• Controlling extraneous factors
• All things but the independent variable are the same
• Participants randomly assigned to groups
• Measurement of other variables for statistical control; zoals het meten van inkomen
in het voorbeeld hierboven.
, Example A: Framing Coupons
• The experimenter develops two or more conditions
-->Two different discount coupons to existing customers
• Conditions differ in one characteristic (“independent variable”)
-->“Obtain a 25% discount” vs. “Pay 75% of the price” → framing
• Everything else is constant across conditions
-->Otherwise, coupons are exactly the same
• Participants are randomly assigned to one condition
-->50% of customers are mailed one coupon, 50% the other
• Do participants behave differently across conditions?
--> DV: percentage of participants who used the coupon
Example B: Default Choice
Example C: Nutritional labeling
What is stimuli
A stimulus is the event or object to which a response is measured. The plural of stimulus is
stimuli. Stimuli can be visual, textual, verbal, space, etc.
Onderzoeksproject
(Blok 8) Tentamen 2
,Lecture 8 - Introduction to experimental design
Today’s plan
- Causation
- Experimental design
- Developing experimental stimuli
- Experimental validity
Causation
The question you should ask yourself; why do we want to do experiments?
Correlation vs. causation
It is easy to find correlation, it is however more difficult to determine whether the correlation
also indicates causation.
Correlation: a mutual relationship or connection between two or more things.
Causation: the relationship between cause and effect; causality.
VB; zo is er bijvoorbeeld een positieve correlatie tussen het aantal bezoekers aan Disney land
Parijs en de neerslag in Louisiana -> de grafieken lijken op elkaar (correlation kan ook negatief
zijn-> dan zijn de grafieken het tegenovergestelde van elkaar). Maar er is geen causality
tussen deze twee, there is no logic behind this correlation-> meer neerslag in Louisiana hoeft
niet tot minder bezoekers in Disney land te leiden. Het is eerder een coincidence dat de
grafieken zo op elkaar lijken.
“Correlation does not imply causation”
Example: why is shoe size related to reading ability? Is there a causal relationship? Of is het
enkel correlation?
-->Shoe size is related to age (older= grotere maat) en age is related to reading ability. So
there is one possible additional factor behind the correlation. Deze factor (age) is not included
in this relationship. Age is an omitted variable.
Several reasons why X and Y can correlate:
- X causes Y
- Y causes X
- Z causes both X and Y
- Spurious correlation
The Difficulty of Causal Inference
Reverse Causality: X appears to cause Y, but it is actually Y that causes the change of X
and because X can also change Y, Y causes the change of Y.
,Examples: Diversification and profitability. Diversified firms(IV) tend to profit(DV) more. Maar
kan ook zijn dat profitability de diversity beïnvloed want als een bedrijf veel winst heeft, kan
het niet gaan zitten op deze winst. Ze zullen dan in verscheidene projecten investeren en op
deze manier meer diversified worden.
Sales(DV) of a brand of soda are higher during weeks of heavy advertising (IV).
- However, advertising is allocated when the stakes are greatest (e.g., during holidays and
summer months). Thus, anticipated sales(IV) may actually cause advertising(DV).
Third Variable: X appears to cause Y, but both X and Y are actually caused by Z. Z is de
omitted variable → voorbeeld over shoe size and reading ability.
On the average, the more toys a child has, the higher his or her IQ.
- Both the number of toys and IQ may be caused by family resources such as income (i.e.,
better nutrition and education).
On average, students who sit in the front of the class end up with higher grades.
- However, students choose where they sit. Motivation may influence both sitting and grades.
Causation
When can we infer that X causes Y?
--> Three conditions for causality:
Relationship between X and Y--> X and Y vary together
Time order --> X cannot happen after Y
Elimination of other possible causal factors --> All other possible causes held constant or
controlled. VB; income in VB hierboven is de variable die je constant houdt. Hier gaan we
regressie voor gebruiken.
Experimental design
We want to do experiments to find out the causation. How to design experiments?
Basic features of a between-subjects design (van experiments)
• Independent variable that is “manipulated” across groups or “between-subject”.
– Between-subject: one participant is assigned to one experimental condition of the
IV (e.g., ad A or ad B)
–Within-subject: one participant is assigned to several experimental conditions (e.g.,
ad A and ad B)
• Dependent variable that is measured
• Context: laboratory, online survey, field, etc.
• Controlling extraneous factors
• All things but the independent variable are the same
• Participants randomly assigned to groups
• Measurement of other variables for statistical control; zoals het meten van inkomen
in het voorbeeld hierboven.
, Example A: Framing Coupons
• The experimenter develops two or more conditions
-->Two different discount coupons to existing customers
• Conditions differ in one characteristic (“independent variable”)
-->“Obtain a 25% discount” vs. “Pay 75% of the price” → framing
• Everything else is constant across conditions
-->Otherwise, coupons are exactly the same
• Participants are randomly assigned to one condition
-->50% of customers are mailed one coupon, 50% the other
• Do participants behave differently across conditions?
--> DV: percentage of participants who used the coupon
Example B: Default Choice
Example C: Nutritional labeling
What is stimuli
A stimulus is the event or object to which a response is measured. The plural of stimulus is
stimuli. Stimuli can be visual, textual, verbal, space, etc.
