Given data in Table 6.6 (attached), the response variable Y is the number of injury incidents, and the predictor variable N is the proportion of flights.

1) First regress Y on N, plot residuals against the fitted values Y . Conclude if error is heteroscedastic, i.e., the "E" assumption is violated.

2) Use data transformation on the response variable, i.e., regress Y on N. The rationale behind this transformation is that the occurrence of accidents, Y , tends to follow the Poisson probability distribution, and the variance of Y is approximately equal to 0.25, see Table 6.5.

What is the regression line equation?

Plot residuals against the fitted values, and conclude if there is still evidence of heteroscedasticity.

table 6.6

Y	N
11	0.095
7	0.192
7	0.075
19	0.2078
9	0.1382
4	0.054
3	0.1292
1	0.0503
3	0.0629