Authoring a PhD Thesis How to Plan, Draft, Write and Finish a Doctoral Dissertation by Patrick Dunleavy

HANDLING ATTENTION POINTS◆ 187

Stem-and-leaf analysis goes a little bit further because it
retains more of the information given in the original numbers.
Each number is divided into two parts, the larger ‘stem’ part
and the smaller ‘leaf’ or unit part. We choose what to set as
the stem in relation to the range of the data being analysed
(the variation from top to bottom score). Here we could set the
stem as equal to 10s, just as in the frequency table above. But
since we want to look a little deeper we could set the stem as
fives instead, with (for instance) one stem running from 20 to
24, and another stem running from 25 to 29. On this basis the
first number in the set is 25, which would separate into an
upper 20s stem and a leaf of 5. The next number 46 would sep-
arate into an upper 40s stem and a leaf of 6. The next number
52 would separate into a lower 50s stem and a leaf of 2, and
so on. Working through the whole set of numbers above would
give a stem-and leaf analysis as follows:

Stem Leaf (1s)

(5s)

52

46

4

3

3 1223344

259

2 00122223

1 5788999

12

0

03

Upper outlier

Upper outlier

Upper quartile 33

Median 22

Lower quartile 19

It is clear here that there is not just a single-peaked curve (one
bell curve). Instead there is a main bulge of observations scoring
from 15 to 23 (including 13 data points), and then another
smaller bulge from 29 to 34 (including 7 data points). Since
there are 27 observations we can find the median by counting