(a) Median can be negative for the fact in univariate data, each piece of data can be negative, hence the median can potentially be negative.
(b) Common sense dictates that for a residual graph there are points plotted above and below the zero line, which indicates a positive or negative residual.
c) Standardized score. This is a difficult one actually, only those people that have done methods would definately know that a standardized score can be negative. We realise that for a standardized score, it is the conversion of scores into a form where we have 0 as the mean, and 1 as the standard deviation, hence as standard deviation is added and subtracted from the mean a negative standardized score can exist
(d) Interquartile range is the easiest to prove that there cannot be negative.
i.e We know that IQR = Quartile 3 - Quartile 1, and that Quartile 3> Quartile 1. We can let Quartile 1 be either postive or negative
Therefore Quartile 3- Quartile 1 is positive ( if quartile 1 is positive, it makes logical sense, from 2 positive numbers a higher - a lower one is positive). However even if Quartile 1 is negative, it is Quartile 3 -- Quartile 1, which is Quartile 3 + Quartile 1 from simple addition. Hence d is always positive.
(e) Correlation co-efficient as stated in the book tells us the strength of a relationship, whether it is positive or negative. Hence a negative correlation co-efficient tells use that there is negative relationship between 2 variables