Unit 02 Multiple Regression Analysis

Economic Magnitude refers to ____ while Statistical Significance refers to ___.

Economic Magnitude refers to *slope* while Statistical Significance refers to *p-value* (% in tails, less than < 0.05 IS statistically significant while greater than > 0.05 is NOT stat. significant.)

Threats to External Validity

if there are limited duration (temporary and people may react differently than if it were permanent), experiment specificity (run in a certain area and may not have the same results elsewhere), Hawthorne effect (participants may make different choices because they know they're being watched), general equilibrium (too small to have unintended consequences but could indeed have them)

Threats to Power

if there are small samples

Selection Bias

in an experiment, unintended differences between the participants in different groups

Treatment Effect

the effect of the treatment on the treated group

Threats to Random Assignment

if participants influence their assignment, if the treatment spills into the control group

(HW04) Regression modeling is a statistical framework for developing a mathematical equation that describes how: a. one dependent and one or more independent variables are related b. several dependent and several independent variables are related c. one independent and one or more dependent variables are related d. All of these are correct.

(Answer) A

(HW02) The regression equation for predicting the # of speeding tickets (Y) from information about driver age (X) is Y = 5.57 - 0.065*X. How many tickets would you predict for a 20 year old? a. 6 b. 4.27 c. 5.57 d. 1

(Answer) B

(HW01) The line described by the regression attempts to a. pass through as many points as possible. b. pass through as few points as possible. c. minimize the squared distance from the points. d. minimize the number of points it touches.

(Answer) C

(HW05) If two variables, x and y, have a very strong estimated relationship via a simple regression, then a. we can say x causes a change in y. b. we can say y causes a change in x. c. there might not be any causal relationship between x and y. d. None of these alternatives is correct.

(Answer) C

(HW03) The regression equation for prediction sales in $1,000's (Y) from information on advertising in $'s (X) is Y = 80,000 + 5X. This implies that an a. increase of $1 in advertising is expected to result in a $5 increase in sales. b. increase of $5 in advertising is expected to result in a $5,000 increase in sales. c. increase of $1 in advertising is expected to result in a $80,005 increase in sales. d. increase of $1 in advertising is expected to result in a $5,000 increase in sales.

(Answer) D

Ordinary Least Squares Regression (OLS)

(Y = Xβ + ε) (Y = β0 + β1X1 + ε)

Costs to Experiments

financial and ethical costs

Natural Experiments (difference-in-difference)

identifying beta (β) by comparing the outcome before and after a discrete policy change

Fixed Effects Models

identifying beta (β) using changes in the X over time rather than at a single point in time

Regression Discontinuity Models

identifying beta (β) using the fact that a very small change in a certain policy will produce a very sharp change in X

Instrumental Variables Models

identifying beta (β) using variation from extra variables not correlated with the outcome but are correlated with the policy