**Question 4**

In a recent article published in Administrative Science Quarterly, a top journal inmanagement, the authors argue that state ownership affects a firm’s innovation output through two underlying mechanisms. On the one hand, following an institutional logic,state ownership of a firm increases its R&D input. On the other hand, however, following anefficiency logic, state ownership reduces the efficiency at which the firm turns R&D inputinto innovation output (i.e. R&D efficiency).

Based on the above argumentation, how wouldthe authors hypothesize the effect of state ownership on the firm’s innovation output?

Note innovation output = R&D input × R&D efficiency. State ownership may range from 0%to 100%.

a. State ownership of a firm has a U-shaped effect (first decreasing and then increasing)on its innovation output.

b. State ownership of a firm has an inverted U-shaped effect (first increasing and thendecreasing) on its innovation output.

c. State ownership of a firm has a positive effect on its innovation output.

d. State ownership of a firm has a negative effect on its innovation output.

**Question 5**

If you predict that the dependent and independent variables do not move in parallel to eachother, then the appropriate null hypothesis is that the coefficient on the independentvariable should be:

a. 0

b. -1

c. +1

d. None of the other three options

An example of “move in parallel”: Y doubles when X doubles, or Y increases by 1% when Xgrows by 1%. Think about elasticity in Economics.

**Question 8**

You first run the following regression:

. sysuse auto.dta, clear

. generate lweight=log(weight)

. regress mpg lweight foreign

Then you test for heteroscedasticity:

. hettest lweight foreign

You obtain the following output:

Breusch-Pagan / Cook-Weisberg test for heteroskedasticity

Ho: Constant variance

Variables: lweight foreign

chi2(2) = 26.91

Prob > chi2 = 0.0000

What is the null hypothesis being tested here?

a. There is heteroscedasticity in mpg conditional upon lweight and foreign.

b. The conditional variance of mpg given lweight and foreign is not constant.

c. mpg is homoscedastic conditional upon lweight and foreign.

d. lweight and foreign have constant variances at different levels of mpg.

**Question 10 ( choose one only)**

In Tutorial 3 slide 23, I show how to run an OLS regression that is equivalent to anova. Now suppose I run the following anova:

. webuse manuf, clear

. anova yield temp##chem##meth

Which OLS regression below is equivalent to the above anova?

a. regress yield i.temp i.chem i.meth

b. regress yield i.temp#i.chem i.temp#i.meth i.chem#i.meth

c. regress yield i.temp i.chem i.meth i.temp#i.chem i.temp#i.meth i.chem#i.meth

d. regress yield i.temp i.chem i.meth i.temp#i.chem i.temp#i.meth i.chem#i.methi.temp#i.chem#i.meth