Mock exam

Question 1

Part-worth utility

Answer 1

–> used to calculate the total utility vakue per each object and the relative importance of the attributes

• accord. to the the empirically determined ranking of the stimuli
• for all attributes of each variable

Question 2

Importance

Answer 2

= Difference between the highest and the lowest par-worth utility of the attribute values

Question 3

Cluster combined

Answer 3

which clusters are combined to a new cluster in the respective step

Question 4

Cluster analysis - Coefficients

Answer 4

showes the euclididan distance measure - the values correspond to the error sum of squares if ward’s method is used

Question 5

what if coefficients increase?

Answer 5

A large increase is an indicator for a reasonable cluster solution

Question 6

stage cluster first appears / next stage

Answer 6

shows in which steps clusters are merged to already existing clusters

Question 7

Reasons for insignificant explanatory variables in a regression mdoel?

What effects does insignificance have?

What actions have to be taken in order to solve insignificance

Answer 7

1. no correlation with the dependent variable / or no other variable takes care of it
3. you dont need to do anything

Question 8

R² goes … with every added variable

Answer 8

goes up

Question 9

How do multicollinearity, omitted variable bias, and endogeneity correlate?

Answer 9

Multicollinearity: highly correlated explanatory variables
–>increase sample size, eliminate variable but lose info, leave it as it is, aggregate into factors

Omitted variable bias: Problem of multiple regression; causes high SE, RE is unbiased, but inefficient –> use robust estimates

endogeneity: explanatory variabl is corrletaed with the error term, caused by omitted variables, autoregression and measurement errors

Question 10

Z-test
T-test
Differences?

Answer 10

T-test:
One: compare sample means with pop mean
H0: b =0
Two: compare (in)dependent samples

• use for N < 30
• follows Students T-distribution
• unknown SD!

Z-test:
compare sample & pop means

• use for N > 30
• normally distributed (for large N a t-test becomse similar to a z-test)
• known SD!

Question 11

Right-skewed:
gm median and mean

left skewed:
gm median and mean

Answer 11

right:
gm > 0
median < 0 and < mean

left:
gm < 0
median > 0 and > mean

Question 12

whats the p-value?

Answer 12

The probability of obtaining the observed value, or even larger ones, if H0 was true

Question 13

advantage of spearmans rho over bravais pearson ?

Answer 13

it decreases the effect of outliers

Question 14

conceptualization

Answer 14

measuring of a concept that is not directly observable

Question 15

basic idea log regr

Answer 15

log. regression determines the probability of a certai result & analysis which factors infuence this probability

DV: dummy