I am studying a variation of regression model called segmented regression. For a response variable $y$ and covariate $x$, it is set that the first $m$ values of $x$ are less than threshold $z$ and the remaining are greater than $z$. Null hypothesis sets that given $x$, the responses $y_i$ follow a simple linear model $\alpha+\beta x_i$. Alternative hypothesis sets that response $y_i$ follows this scheme: $$E(y_i|x_i)=\alpha+\beta x_i,(i=1,\ldots,m)$$ $$E(y_i|x_i)=\alpha+\beta x_i+\delta(x_i-z),(i=m+1,\ldots,n)$$ Where $\alpha,\beta,\delta$ are unknown. The goal of this formulation is showing that expected response given $x$ does not change at $z$ but the slope $\beta$ does. I have tried to arrange the expectations in normal equations but I got large chains of values and I do not know how to show this. Could you please help me how to adress this question.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.