Survival Analysis

Question 1

Survival analysis

Answer 1

DV = duration until success/failure of an event

Cox model-assumption of proportional hazard rates between groups

Same as event history analysis

Question 2

Survival analysis assumption

Answer 2

Assumes multivariate normality; linearity and homoscedasticity not required, although more stable
with them

Question 3

Survival analysis- censoring

Answer 3

not being able to know how many subjects failed before or after sampling; accounted for in survival analysis but assumes that their behavior is not significantly different (right-censoring – occurs when study ends before the event has occurred, e.g., data on when an employee eventually leaves the firm is not known if employee has not left firm when study ended; leftcensoring is rare – occurs when the event of interest has already occurred before enrolment. This
is very rarely encountered

Question 4

Difference between survival analysis and logistic regression

Answer 4

Survival analysis looks at time to event while logistic regression is a binary outcome. Survival analysis accounts for problems caused by censoring via an algorithmic approach. This allows the gleaning of partial information from cases that have be censored and making adjustments
accordingly

logistic also is analysed in terms of odds ration, cox regression analyses hazard rates

With survival analysis, you can consider line items that you would remove in a logit- i.e. some people may not have experience with an event. If they have not yet experienced the event, they
would be zero, else 1.