Question

# Multiple regression model Consider a dataset for 500 students living in Melbourne with the following variables...

Multiple regression model

Consider a dataset for 500 students living in Melbourne with the following variables measured in 2019 --- expenditure on public transport, ?E, measured in dollars; number of days he/she attended school, ?S; number of days not attended school, ??NS (which is equal to 365−?365−S); income, ?I, measured in dollars; and age in year, ?A. A researcher is interested in estimating the following model using Eviews: ??=?1+?2??+?3???+?4??+?5??+??Ei=β1+β2Si+β3NSi+β4Ii+β5Ai+ei. Without any further information, we know for sure that one assumption of the multiple regression model is violated here. Which assumption is violated?

Select one:

a. The errors must be uncorrelated ???(??,??|?)=0Cov(ei,ej|x)=0 for all ?≠?i≠j

b. There can be no exact linear relationship between independent variables.

c. The errors ??ei must be normally distributed.

d. The errors must have constant variance ???(??|?)=?2Var(ei|x)=σ2 for all ?i.

Answer- b) There can be no exact linear relationship between independant variables.

Within the assumptions it has been stated that there shouldn't exist any Multicollinearity between the independant or explanatory variables. This happens so that the relationship between the independant variable and the dependant variable isn't hampered due to the relationship between the two independant variables. Further it can also happen that if Multicollinearity existed between the two independat variables, the two independant variables could show a very similar relation with the dependant variable, seperately. Hence, to avoid any such conflict between the relationship of independant and dependant variable, there shouldn't exist multicollinearity between the independant variables.