I obtain a residual plot against the fitted values and it does show some pattern for the residuals, so I suspect there may exists heteroskedasticity problem. Which kinds of test can be apply here to test if residuals are have constant variance or not? How do I apply those tests in R?

enter image description here

My nonlinear regression model is:

mod2 <- nls(Y~gamma_0+gamma_2*exp(-gamma_1*X), data = data,
            start = list(gamma_0 = 0, gamma_1 = 0.0006934571, gamma_2 = 0.6021485), trace = TRUE)

Here is the output:

Formula: Y ~ gamma_0 + gamma_2 * exp(-gamma_1 * X)

         Estimate Std. Error t value Pr(>|t|)    
gamma_0 4.823e-02  1.456e-02   3.313  0.00561 ** 
gamma_1 1.117e-03  9.207e-05  12.136 1.82e-08 ***
gamma_2 7.134e-01  2.277e-02  31.338 1.24e-13 ***
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.02332 on 13 degrees of freedom

Number of iterations to convergence: 6 
Achieved convergence tolerance: 1.156e-06
  • $\begingroup$ Not a lot of data, but my eyes see a stronger pattern of incorrect functional specification than heteroscedasticity. I see the latter but the former is a greater concern. $\endgroup$ – BigBendRegion Nov 22 '20 at 12:58

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.