(9.1) Quant methods

Question 1

3

Describe the relationship between quant and qual methods

Answer 1

• Quant: greater breadth
• Qual: greater depth
• More complementary than substitutable

Question 2

Describe population (QM)

Answer 2

• Complete group we want to learn about
• All observations across time and space
• e.g. all eligible voters in Germany

Question 3

Describe sample (QM)

Answer 3

• Subset of population we actually study
• e.g. 1200 German voters surveyed by phone
• Used to make inferences about population — any measure comes with degree of uncertainty

Question 4

What are the 4 main variable types?

Answer 4

• Dummy (binary)
• Categorical
• Discrete
• Continuous

Question 5

Describe dummy variable

Answer 5

• Can only take 2 values (0 and 1)
• 1 if property met; 0 if not
• e.g. voted for Biden, occurrence of civil war

Question 6

Describe categorical variable

Answer 6

• Finite number of categories
• e.g. gender (7 categories), Freedom House index (free/partly free/non-free)
• Ordinal: can rank categories but intervals not meaningfully interpretable

Question 7

Describe discrete variable

Answer 7

• Countable number of increments
• Can only assume finite number of real values within interval
• e.g. number of terrorist attacks, number of active interest groups

Question 8

Describe continuous variable

Answer 8

• Can take any value (infinitely small increments)
• Intervals between values meaningfully interpretable
• e.g. age, income, GDP

Question 9

What are the 3 levels of measurement?

Answer 9

• Nominal
• Ordinal
• Interval

Question 10

Describe nominal level

Answer 10

• Values classify cases into categories with no logical order
• Cannot rank answers
• e.g. party choice, country, gender

Question 11

4

Describe ordinal level

Answer 11

• Values classify cases into ranked categories
• Intervals between values not meaningfully interpretable
• Can only make ‘greater than’/’less than’ statements
• e.g. edu level, 5pt rating scale

Question 12

3

Describe interval level

Answer 12

• Values indicate actual quantity with equal distances between them
• Intervals are meaningfully interpretable
• e.g. income, age, temperature

Question 13

What are the 3 measures of central tendency?

Answer 13

• Mode
• Median
• Mean

Question 14

3

Match variable type to level of measurement

Answer 14

• Dummy → nominal
• Categorical → nominal/ordinal
• Continuous → interval

Question 15

3

Match variable type of central tendency

Answer 15

• Dummy → mode
• Categorical → mode/median
• Continuous → mean/median

Question 16

What are the 2 measures of dispersion?

Answer 16

• Variance
• Standard deviation

Question 17

Describe variance

Answer 17

• Measures variability of observations around the mean

Question 18

What does a small variance indicate

Answer 18

• observations concentrated around mean
• mean is informative

Question 19

What does a large variance indicate

Answer 19

• observations spread far from mean
• mean less informative

Question 20

Why do we use standard deviation and not variance?

Answer 20

• so unit matches variable (e.g. £ not £²)

Question 21

What are the 4 ways to display descriptive statistics?

Answer 21

• Bar charts
• Histograms
• Kernel density functions
• Boxplots

Question 22

When do we use bar charts?

Answer 22

• dummy + categorical variables
• See mode + dispersion

Question 23

When do we use histograms?

Answer 23

• Continuous/discrete variables
• See mode + dispersion

Question 24

1 + photo

Describe kernel density functions

Answer 24

• Same info as histogram but with a smoother applied

Question 25

When do we use kernel density functions

Answer 25

* continuous + discrete * see mode + dispersion

Question 26

# 4 When do we use boxplots

Answer 26

* continuous + categorical * denote outliers * see median * bigger box = larger dispersion

Question 27

What are 2 ways to analyse relationships between variables

Answer 27

* Mathematically: correlation * Visually: scatter plot

Question 28

Describe correlation

Answer 28

* Measure of extent to which 2 variables tend to occur together * Coefficient from -1 to 1 * larger number (+ or -) = stronger correlation

Question 29

Describe the intuition behind the correlation coefficient

Answer 29

* For each observation, ask: is deviation from mean in x accompanied by deviation from mean in y? * Divide by sd

Question 30

Why do we divide by sd in the correlation coefficient?

Answer 30

* large deviations mean little if variable is very disperse * small deviations matter more if variable has low variability

Question 31

When do scatter plots work?

Answer 31

two continuous variables