Statistics Problems in a Funnel-In-Pattern - Essay Example

(a) There appears a non-linear relationship between the Human Development Index (HDI) and the number of cell phone subscribers. The relationship between HDI and the number of cell phone subscribers appears to be logistic (similar to exponential) that is increases quickly at first, and then slows down, and then finally reaches a saturation point. Therefore, fitting a linear model to these data might be misleading.(b) The scatterplot of residuals versus predicted HDI will look like a bend (curved) graph.

The errors (residuals) will be positive at the start (left) and then drop down in the middle and drop down further (negative) on the right.3) If that point were removed from the data, the correlation would become weaker (or no correlation). Removing this point there appears no relationship between x and y. This point is far from all other scattered points and thus tries to form a linear relationship.1) The point is unusual because it has high leverage.2) Yes, I think that point is an influential point.3) If that point were removed from the data, the correlation would become weaker (or no correlation).

Removing this point there appears no relationship between x and y. This point is far from all other scattered points and thus tries to form a linear relationship and thus increases the correlation.4) If that point were removed from the data, the slope of the regression line would decrease. The reason for this is that it tends to decrease the y-intercept by increasing the slope of the regression line.