Which of the following is not an option for handling high leverage points?

In the context of managing high leverage points in statistical analysis, the correct response indicates that using a non-robust estimation method is not a viable option. High leverage points are data points that can exert substantial influence on the fitted model, potentially skewing the analysis and leading to misleading conclusions.

Robust estimation methods, in contrast to non-robust methods, are specifically designed to mitigate the effects of outliers and leverage points. They provide a more reliable analysis by reducing the impact that these influential observations have on the model. Therefore, opting for a non-robust estimation method is counterproductive when dealing with high leverage points, as it would likely amplify their influence rather than lessen it.

The other options reflect appropriate strategies for addressing high leverage points: deleting the observation might be warranted in certain cases, while including it and discussing its effects provides transparency in the analysis. Choosing another variable may also help in representing the information more accurately without being overly influenced by the high leverage point. Each of these methods acknowledges the potential impact of high leverage observations while seeking to maintain the integrity of the model.