Hi. Basically i want to know how to implement a regression equation in R or JMP. its a multiple regression model, although with a slight twist (dummy variable)

the equation is

Y(i) = Beta0 + Beta1DP(i)+Beta2DPR(i) + Beta3LEVERAGE(i) + Beta4SIZE(i) + Beta5INVESTMENT(i) + E(i)

where Y is my dependant variable and DP,DPR,LEVERAGE,SIZE,INVESTMENT are my independant variables.

however heres the twist. the independant variable DP takes the value 0 or 1, and acts as a switch for the indpendant variable DPR. the reason is DP represents if a company pays out a dividend, and DPR represents the amount that they payout. therfore when DP is 0 the DPR is non applicable.

all other variables other than DP, are continous numerical variables.

How do i implement this equation in R or JMP? or STATA?

thanks in advance

Then you'll need a DP*DPR variable,then when DP=1 the payment will be DPR else it won't be applicable

Y(i) = Beta0 + Beta1[DP(i)*DPR(i)] + Beta3LEVERAGE(i) + Beta4SIZE(i) + Beta5INVESTMENT(i) + E(i)

Adding the DP in the regression would result to a variable intercept term

Y(i) = Beta0 + Beta1[DP(i)*DPR(i)] + Beta3LEVERAGE(i) + Beta4SIZE(i) + Beta5INVESTMENT(i) + Beta6DP(i)+E(i)

as, when DP=1, intercept=Beta0+Beta6

So, in the usual flavour of R write

toy.lm<-lm(y~dp.dpr+leverage+size+investment+dp),

where dp.dpr=dp*dpr

Then you'll need a DP*DPR variable,then when DP=1 the payment will be DPR else it won't be applicable

Y(i) = Beta0 + Beta1[DP(i)*DPR(i)] + Beta3LEVERAGE(i) + Beta4SIZE(i) + Beta5INVESTMENT(i) + E(i)

Adding the DP in the regression would result to a variable intercept term

Y(i) = Beta0 + Beta1[DP(i)*DPR(i)] + Beta3LEVERAGE(i) + Beta4SIZE(i) + Beta5INVESTMENT(i) + Beta6DP(i)+E(i)

as, when DP=1, intercept=Beta0+Beta6

So, in the usual flavour of R write

toy.lm<-lm(y~dp.dpr+leverage+size+investment+dp),

where dp.dpr=dp*dpr

Actually i think i may have spotted a slight error in my formula.

DP may not be a relevant independant variable in my regression model after all. All DP does is say whether a firm pays dividends, and DPR says how much. So DPR will show how dividends affect the dependant variable (firm performance) anyway ( if the regression coeffecient of DPR is near 1 it will mean high dividends are good, and if the regression is significantly negative it will mean high dividends are bad for the company). therfore i dont think i will need DP since the coeffecient of DP will tell me nothing new i.e. a significantly positive coeffecient will mean paying dividends is good for firm performance, and a significantly negative coeffecient will imply paying dividends is bad for firm performance.

i do need DP however to operationlise DPR. so DP should be used as a dummy variable. how with this affect the formula?
can a dummy variable be used in a regression formula without being an independant variable?
should i just keep the formula i have anyway? and not comment on the affect the coeffecient of the binary variable DP has on the regression?

as always many thanks

Actually i think i may have spotted a slight error in my formula.

DP may not be a relevant independant variable in my regression model after all. All DP does is say whether a firm pays dividends, and DPR says how much. So DPR will show how dividends affect the dependant variable (firm performance) anyway ( if the regression coeffecient of DPR is near 1 it will mean high dividends are good, and if the regression is significantly negative it will mean high dividends are bad for the company). therfore i dont think i will need DP since the coeffecient of DP will tell me nothing new i.e. a significantly positive coeffecient will mean paying dividends is good for firm performance, and a significantly negative coeffecient will imply paying dividends is bad for firm performance.

i do need DP however to operationlise DPR. so DP should be used as a dummy variable. how with this affect the formula?
can a dummy variable be used in a regression formula without being an independant variable?
should i just keep the formula i have anyway? and not comment on the affect the coeffecient of the binary variable DP has on the regression?

as always many thanks

ignore this last post plz ( the one above) i had a momentarily lapse in concentration lol; im using DP because it tells me whether dividends are good or bad for firm performance, and DPR is saying how much should be paid ( i.e. signifcantly postive coeffecient means high dividends are good, significantly negative co effecient low dividends are good.) therefore both DP and DPR are neccessary variables.