Term talk.

Which Way Is Up?

Numbers are a precise way of communicating amounts, but words can be confusingly imprecise in communicating how amounts are changing. An example of the difficulty can be seen in the following sentence:

Although titanium dioxide production fell to 650,000 tons in 1982, about 30,000 tons more than originally predicted, the economic recovery this year is expected to boost output to near the 1980 figure.

Question: what was the originally predicted tonnage?

In ten years' worth of writing and editing classes, I've found about a 50:50 split among those who read the sentence to say that production fell 30,000 tons more than expected (the predicted tonnage being 680,000) and those who think that the production of 650,000 tons was 30,000 above a predicted 620,000 tons. You cannot, of course, determine between the two choices from the sentence, and in fact, you could not learn the correct figure from the entire article this sentence appeared in. This is clearly a structural problem within the sentence that can be easily corrected if it is noticed by the author or an editor, but some problems involving numbers aren't so tractable.

Heading in Opposite Directions

In some fields it is easy to produce a sentence or two full of words heading in opposite directions. For example, a characteristic of petroleum fuels is that their volume changes significantly with temperature, and it is usual to track the volume variations and their trends. As seasons and temperatures change, a report on these variations may contain sentences like these:

The refinery has experienced decreased volumetric gains over the past months. The trend of these lower gains is increasing.

In reality, the first sentence itself isn't too difficult to understand, simply saying that recent volume increases are not as large as those some months ago. But the use of decreased modifying the word gains is the first glimpse of the problem caused by the placing of pairs of opposites close together, a problem that is exacerbated by having one of them modify the other. Using words like increasing and decreasing with single variables makes fairly easy reading; when volume increases, it becomes larger. Increasing volumetric gains is also easy to understand, because the directional words, increasing and gains, send your thoughts in the same direction: something is growing larger, and the change with time is positive. Decreased gains, however, requires the reader to evaluate the negative direction of a set of positives, which presages more difficult mental effort required by the next sentence of the example.

Trends in Data

In the second sentence, we read that lower gains have an increasing trend. Here we must figure out the net result of three directional words, and even after four years in petroleum fuels work, I had a tough time recognizing what the writer was trying to say.

In discussing the trend of the data values, this writer could have used a more precise term. A trend described as increasing tells the reader that values are diverging from a reference level--becoming larger--but that trend can be either positive or negative. Furthermore, a trend can concern a variable that is either positive or negative itself, so that the trend of a positive variable (such as volumetric gains in fuels) can be decreasing (approaching a reference level). But applying these explanations to the second sentence of the example doesn't help much; how is a trend of lower gains (approaching zero gain or loss) increasing?

A better description of the trend would have used more common terms: upward or positive, downward or negative, and accelerating. The meanings of two pairs are obvious, while accelerating describes an increasing rate of change of the trend. When the writer in the second sentence states that the trend is increasing, the reader has to sort out the confusing directions of the words and realize that a trend of ever-lower volume gains cannot be positive. Once that is understood, the reader is able to recognize that the writer must mean that the rate of change is increasing, the slope becoming steeper with time. This could better be said as--

The negative trend of the gains is accelerating.

It is important to try to clearly and precisely describe the type of change data values are undergoing. Trying to sort out a page or two of concepts like "increasing trends of lower gains" is guaranteed to make your head hurt.

Even in areas other than refinery trends, it is often poor practice to place forms of the words increase and decrease adjacent to each other in a sentence. For example:

The mill's net profit decreases (losses) increased in the last quarter.

Here we have the juxtaposition of profit and losses as well as decreases and increased. Readers have to stop and think about how to reconcile pairs of opposites that are bound together within the meaning of the sentence. It is much easier on the reader to use words that are not direct opposites:

The mill's net profit decreases (losses) became larger in the last quarter.

Or, to get rid of both pairs:

The mill's net losses became larger in the last quarter.

Where Do We Start From?

Another important point is that readers need a baseline or reference point on which to base their understanding. When such a baseline is established and followed throughout a document, the reader's job of comparing, evaluating, and understanding the data is made easier. Not coincidentally, the writer's job is easier also, because errors in stating results are much more obvious when reviewing drafts in which similar information is presented in parallel form.

Recently I edited a manuscript in which the authors compared a number of test samples under a variety of experimental conditions. For much of the first half of the manuscript, the results were presented in this form:

Relative to sample X37, Y and Q were higher under the 45 |degrees~ conditions and virtually the same under the 37 |degrees~ conditions.

After two or three sets of results reported in this way came another statement, in this form:

X37 was lower under 60 |degrees~ conditions than was Y.

Now the writers might have thought that some variation was in order or, possibly, they didn't think about the particular form of expression at all when this was written; but in fact, this change in reference point causes a pause before the reader recognizes how this bit of data fits with those in previous paragraphs. Most technical subjects are complex enough on their own; it isn't necessary to add more work for the reader by varying the baseline from which results are presented.

The following passage is another manifestation of this type of baseline problem. Clear your mind, concentrate really hard, and see how few rereadings it takes you to understand this pair of sentences.

Samples containing no |radiolabel~ would be expected to exceed the average background by one standard deviation 16% of the time; experimentally they were found to exceed it about 20%. Samples containing enough |radiolabel~ to be expected to exceed the average background by 1.5 times the expected standard deviation would be expected not to exceed the average background by one standard deviation 30% of the time; experimentally they were found not to exceed it 20-30% of the time.

Summary

Science and engineering topics inevitably involve measurements, amounts, and the trends of variables. Care is needed in constructing written explanations of the data that don't cause the readers mental vertigo. Finally, once you have a sound, clear description of the data, proofread the text and the galleys carefully: You don't want to be tripped up by a small detail (for example, arrowheads pointing the wrong way).

The carbon content of lignin isolated from soda and sulfate cooking of R. communis and bagasse increased in the order acetone-soluble sulfate lignin |is greater than~ soda lignin |is greater than~ acetone-insoluble sulfate lignin, while methoxy content was in the order . . .
COPYRIGHT 1993 Society for Technical Communication
No portion of this article can be reproduced without the express written permission from the copyright holder.