next up previous
Next: Conclusions Up: Analysis Previous: Stability

Sources of Variation of the Times

Figure 3 displays the times versus groups (``cr'' is the group with Nightcrawler as releaser and Robin as timer, for example). There are two sources of variation revealed here: the variation of times within each group and the differences in measurements from group to group.

From this plot we can see that the variation of the times within group cr is greatest, with the variation within the other groups being smaller and about equal from group to group. Further investigation revealed that the smallest observation in the cr group was caused by the helicopter losing lift after its rotor brushed the releaser's shirt. Thus, it makes sense to omit that observation. If we do so, the variation appears to be about equal within each group.

The other component of variation in the data shown in Figure 3, is the differences in the centers of the measurements in the different groups. Groups cr, cw, rc and rw all have times scattered about a central value of 25 or so (if we ignore the outlier in group cr, it's center is a bit higher). The other two groups have centers which are considerably lower-around 15 or so.

Joseph D Petruccelli
Thu Aug 29 11:44:31 EDT 1996