HC 10 Research methods 2

Question 1

Conducting research

Answer 1

1. Research design
2. Analysing results
3. Endogeneity

Question 2

Specific (measure)

Answer 2

SPin: Hoge Specificity rules IN.
- True Negative: Wie hebben het echt niet?

• Vergroten Specificiteit: No or few false positives, but many false negatives (verkleinen sensitiviteit).
• Hoe groot is de proportie die het niet hebben (actual negatives) die correct zijn geïdentificeerd?

SPin en SNout: Hoge SPeciviciteit rules IN! hoge SeNsitiviteit rules OUT!

Question 3

Sensitive (measure)

Answer 3

SNout: Hoge Sensitivity rules OUT.
- True Positive: Wie hebben het echt wel?

• Vergroten Sensitiviteit: No or few false negatives but many false positives (verkleinen specificiteit)
• Hoe groot is de proportie die het wel hebben (actual positives) die correct zijn geïdentificeerd?

SPin en SNout: Hoge SPeciviciteit rules IN! hoge SeNsitiviteit rules OUT!

Question 4

Sample size (Type 1, Type 2?)

• Too Small
• Too Big
• Rule of Thumb

Answer 4

• Too small: Unable to pick up subtle effects (Type II error: wrongly not rejecting the H0)
• Too big: almost everything becomes significant (Type I error: wrongly rejecting the H0)
• Rule of thumb: 20 participants per cell

Question 5

Analyses Study types (4)

Answer 5

• Discriptive study
• Correlational study
• Experimental studie
• Quasi-experimental study

Question 6

Discriptive study (2)

Answer 6

• can give ecological valid picture of constructs in a situation
• Cannot draw conclusions about relationships or causes

Question 7

Correlational study (2)

Answer 7

• Can give insight into the relationship between constructs

- Cannot draw conclusions about causality

Question 8

Experimental study (3)

Answer 8

• Can give insight into the causal relationship between independent and dependent variabels
• Because of extremely high control of environment low ecological validity

Question 9

Quasi-experimental designs (4)

Answer 9

• Field experiment
• Usage of intstrumental variables (e.g. smoking, health and tax as independent variable)
• Granger causality (longitudinal data; corruption and bonusses)
• combination of correlational and experimental design

Question 10

Endogeneity

Answer 10

True model: Y is completely explained by X1 and X2 ( no error)

Question 11

E.g. Selection bias (AH)

Answer 11

• Participants are generally more compliant than those who refuse to partake
• Shop lifting is not equally distributed over all times of the day (eg. More during school lunch hour)

Question 12

E.g. Information bias (AH)

Answer 12

•Eg. compliance and risk officers report more economic crime in organizations than other
employees

Question 13

Type 1 error

Answer 13

• Alpha
• False Positive
• Reject H0 while it is actually true

Question 14

Type 2 error

Answer 14

• Beta
• False Negative
• Accept H0 while it is actually false

Question 15

3 Sources of Endogeneity

Answer 15

1.Omitted variables

2.Common method variance /bias
•Harman’s single factor score (if factor score

Question 16

Granger Causality

Answer 16

Meting in tijd
X1 > Y1
X2 > Y2

X1 op Y2 gecontroleerd voor Y1
Je moet geen effect heben van Y1 op X1 gecontroleerd voor X1