I am running segmented regression using the R package 'segmented'. The original binomial logistic regression has two coefficients, approach_km (continuous), and sea (dichotomous) that explain the probability of the dependent variable occurring. I am interested in investigating if there is a break-point in the regression slope for approach_km.
Here are the coefficient estimates from the original non-segmented glm
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -1.15563 0.17352 -6.660 2.74e-11 ***
approach_km -0.03915 0.01140 -3.432 0.000598 ***
sea2 0.34501 0.20515 1.682 0.092615 .
Here are the estimates after running the model through the segmented function
Estimated Break-Point(s):
Est. St.Err
3.794 1.134
t value for the gap-variable(s) V: 0
Meaningful coefficients of the linear terms:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -0.08233 0.32692 -0.252 0.801
approach_km -0.49096 0.19776 -2.483 0.013 *
sea2 0.27182 0.21141 1.286 0.199
U1.approach_km 0.49085 0.19817 2.477 NA
A Davies' test indicates that the differences in the two slope estimates is significant.
The question I have is how do I interpret the coefficient sea? I see that the estimate (and the p value) changes in the segmented output. Does this mean that this is the best estimate for the value of the coefficient for both of the estimated slopes? Since the sea coefficient is not significant we can assume that this variable will not statistically alter the overall effect or interpretation of the model but I am interested in how I would interpret this result if it was in fact significant.
