2
2
8
2
p<0.05
2
2
.009
.009
8
Example 4
2 TPS scores for veterans were
114% higher than for new
recruits (Figure A).
reader. Tables are also preferable
and provides verbal answers to
in some instances. As Edward
specific questions being addressed
This gets the reader thinking
Tufte suggests in his 1983 book,
by the researcher.
about relative differences and abso-
The Visual Display of Quantitative
lute scores, which are presented in
Information, complex and large Communicate It Once
an accompanying figure.
sets of summary statistics can be
In the text, you will tell the reader
recorded in well-organized tables.
what can be found in the figures
Respect Your Reader
Tables can allow the reader to
and tables. Don’t list the same
Keep the inferential statistics in
explore the data when there are
statistics in the text as are already
the background. Some authors
many relevant combinations of
plotted in a graph. Do describe
use a repetitive pattern in which
comparisons among variables.
the general patterns and details
the first sentence declares an
not directly plotted in the figure.
Integrate Nontextual
effect significant and the second
Consider the following examples:
Results and the Text
sentence describes the direc-
tion of the effect. This couplet is
You should be able to read the
Example 3
repeated for each effect. Here is
text in the results section and
Figure A illustrates that the
an example:
know the answer to each question
mean TPS score was 21 (SE
raised in the introduction without
= 3.4) for new recruits and
Example 5
looking at tables or figures. This
45 (SE=5.5) for veteran
The logistic model predicting
recommendation suggests you
employees.
relapse was significant (LR
narrate the story in the graphs by
x
2
parenthetically referencing your
If Figure A is a plot of the
(3)
= 21, p<0.05). Level of
social supports was found to
graphs when describing patterns
means, the text is redundant. If
be significantly related to sub-
in the data graphics. Consider
the text is sufficient to recreate
stance use relapse (p<0.05).
the following examples:
the graphs, dump the graphs or
Higher levels of abstaining
use the text to present another
Example 1
social supports decreased the
perspective on the data or focus
rate of relapse. Joint diagnosis
Watermelon production was on the take-home message.
with internalizing and exter-
23% [95% CI: 20–26%]
nalizing psychiatric disorders
higher in plots inoculated with
was significantly related to
mycorrhizal fungi as compared
to controls (Figure 2).
Example 1 is much more
informative than writing, “The
(See Example 4)
results of the mycorrhizal inocu-
lation are presented in Figure 2.”
Example 2
Subjects with low self-efficacy
relapsed sooner than those
with high self-efficacy (Figure
3). Relapse risk was 2.1 [95%
CI: 1.8–2.3] times higher for
those with low self-efficacy.
Example 2 is a better use of
text than, “The relationship
between self-efficacy and relapse
is presented in Figure 3.”
Figure A. Mean TPS scores by employee type. Error bars are 95%
The examples demonstrate
confidence intervals.
how the text highlights patterns
that can be seen in the graphics
40 AMSTAT NEwS SEPTEMbER 2009
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80