# A Positively Skewed Distribution

## David C. Howell

An example that we will look at several times in the future comes from a study by Mireault (1990) investigating the effects of the death of a parent on the emotional well-being of college students. Among other things, she asked three different groups of college students to rate the perceived vulnerability to loss--i.e., how vulnerable did they feel about the loss of someone important to them. The three groups were (1) a group who had had a parent die before they started college, (2) a group whose parents had divorced, and (3) a group whose parents were both alive and still married to each other. Download these data from Mireault.dat . There are many variables here. They are, in order, ID, Group, Gender, YearColl, College, GPA, LostPGen, AgeAtLos, SomT, ObsessT, SensitT, DepressT, AnxT, HostT, PhobT, ParT, PsycT, GSIT, PVTotal, PVLoss, SuppTotl. We are interested in Group and PVLoss. The other variables will come up in other exercises.

• Plot the data on PVLoss for all groups combined. If your software gives you the option, superimpose a normal distribution on top of the plot as a reference.
• Calculate the descriptive statistics for these data and see how they are reflected in the plot.
• Obtain the descriptive statistics and histogram for each group separately.
• Create a boxplot for these data separated by groups.
• Do these data lead you to suspect that loss of a parent leads to emotional changes in the children?
• How does loss through death seem to compare to loss through divorce?
• How, can you relate the shape of the distribution of the complete set of data to the differences between groups?

Send mail to: David.Howell@uvm.edu)