|
T. R. Girill
Society for Technical Communication/Lawrence Livermore Nat. Lab. (retired)
[email protected]
Technical Writing: How Text Supports Science Graphics
Most technical reports, papers, or articles contain 'visualizations'
--constructed graphs, charts, or diagrams--on which descriptive or
explanatory text supports the purely visual features. It is the
writer's job to draft and deploy such on-visualization text. A
recent analysis comparing empirical studies of on-graphics text
effectiveness by Marti Hearst reveals some often-overlooked writer
responsibilities here (M. Hearst, 'Show it or tell it?' Communications
of the Association for Computing Machinery, October 2023, 66:10,
68-75, doi: 10.1145/3593580). What do readers of technical text
really want in on-visualization textual annotations, and hence what
kind of extra planning and design effort do these needs impose on
(student) writers?
Hearst began her literature review with several anticipatory guesses
about the role of on-visualization technical text that may be widely
shared:
(1) science graphs and diagrams are so 'immediate' and vivid that
they are easy to interpret and need only light textual support, and
(2) hence most readers will prefer 'low text' visuals to heavily
annotated versions (p. 68).
Text Dominates Graphics
She actually discovered, however, that empirical studies of science
visualizations all went the other way. Comparisons of four versions
of an x-y line graph, for example, ranging from a plot with only
axis labels through increasingly annotated graphs, to a 'version'
with a text-only paragraph describing the plot without showing it
at all revealed that:
(1) 'as long as the text was relevant and not redundant,' pointing
out graph peaks and slope changes, for example, then more, 'thick'
on-graph text was preferred to less, 'thin' textual commentary
(p. 70), and
(2) a surprising 14% of respondents actually preferred an all-text
paragraph that described the plot (in about 100 words) but never
showed it visually at all (p.71).
The Two Literacies
This clear surprising dominance of explanatory text over
demonstrative graphics Hearst attributed to the difference of
'literacy' levels between reading and visual interpretation.
Everyone consuming a scientific paper or report has already learned
to read, and they have come to rely on what they learn through
such careful reading. Interpreting science visuals, however, even
just simple x-y plots of measured quantities, is a separate
learned skill that apparently many people never really master.
Their visual literacy is so much weaker than their textual literacy
that they prefer for someone else to just summarize the most
important features of or trends shown on a scientific visualization
rather than have to closely interact with the graph themselves.
As long as the textual comments are relevant and accurate,
relying on them is often a preferred alternative to doing their
own visual analysis.
The BLV Case
Hearst also discovered a very special case of this text-over-visuals
preference that calls for a different writer response. Blind or
low-vision (BLV) users of technical articles need a text alternative
to each plot that descriptively calls out the graph's visual
features rather than interpreting them ('a line plot of time
along the x-axis and pollen level along the y-axis, with a peak
about two-thirds along the hour-by-hour time span at about 4 p.m.').
This could also be useful ALT text for HTML display tools.
Another Responsibility
So including graphs and diagrams in a technical text can make it
more useful, but the writer must take on the added responsibility
of explaining those visualizations, not just injecting them into
the text unsupported. On-visualization comments must be carefully
planned and adjusted to meet reader/viewer needs by (1) calling
out salient visual features and then (2) stating overtly what
those features mean so that viewers are satisfied rather than
puzzled by what they see. Thus even visuals impose a textual
responsibility when one writes about science.
[For more more ways to place this in the context of usable
science-prose design, see
http://writeprofessionally.org/techlit/usabilty
For more on drafting good technical descriptions, see
http://writeprofessionally.org/techlit/analysisgd]
|