Fichas
Non-linear relationships
Fichas del conjunto (24)
- 1What are the 4 non-linear relationships that can be used?- logs
- quadratics
- interaction terms
- dummy variables - 2Given an example of a logged regressionanswer on a physical flashcard
- 3Why are logs used?make models closer to OLS
- 4What can using logs reduce? When is this the case?Can reduce heteroskedasticity and makes the conditional distribution more normal when a dependent variable is strictly positive (often has heteroskedasticity and/or skewed)
- 5Why are quadratics used?to allow for non-constant effects of x on y
- 6When does the relationship between x and y have an inverse U shape?When β ̂_1 > 0 and β ̂_2 < 0
- 7What is the turning point?where the marginal effect of x on y = 0, after this point marginal returns become negative
- 8How is the turning point calculated?answer on a physical flashcard
- 9When is an increase in x never reducing in y?With non-negative x and β1 >= 0, β2 >= 0
- 10When is an increase in x never increasing in y?With non-negative x and β1 =< 0, β2 =< 0
- 11When combining logs and quadratics, which way round do you apply them?Take the quadratic of the log
- 12Why are interaction terms used?They let the marginal effect, elasticity or semi-elasticity of y with respect to an explanatory variable depend on another explanatory variable z
- 13Give an example of a regression that uses reaction termsprice = β0 + β1squmtr + β2bdrooms + β3squmtr x bdrooms + u
- 14Why are dummy variables used?When you want to take into account qualitative information: gender, race, martial status, industry, firm location etc
- 15How many dummy variables do you include?Classify variable in n groups, need to include n-1 variables (colinearity)
- 16How is the coefficient of dummy variables interpreted?Interpret with respect to the variable represented when the dummy = 0
Interept shift (no change in slope) - 17What is the effect of ZCM on dummy variables?Causeds downwards bias
- 18How do you find the direction of ZCM bias?Work out the effect between the dummy and omitted variables and between the dependent and omitted variables.
Multiply the signs together to find the direction of the bias - 19Can dummy variables be logged? Why/why not?No, either 0 or infinite
- 20With multiple dummies, what is the interpretation of each coefficient?the interpretation of each coefficient is with respect to the unrepresented variables
- 21What type of data can dummies also model?Ordinal, all induviduals can be listed, and Categorical
- 22What issue is encountered?With ordinal data it is hard to interpret a one unit increase
- 23What is the effect of interacting dummies with continuous variables?Allows for a difference in slope
- 24What is the interpretation of the dummy interaction coefficient?The returns to the continuous variable with respect to the change in the dummy
Información
- Descripción
- Year 2 ECN21004 - Econometrics Fichas sobre Non-linear relationships, creado por Imogen Donoghue el 11/01/2023.
- Creado por
- Imogen Donoghue · ene 2023
- Actualizado
- 18 ene 2023
- Visitas
- 0
- Valoración