A layout workshop.
Designed for your students, this workshop involves layout. Along the way, students will learn that revising layout forces them to rewrite text, redesign maps, and rescale graphs.
We give you a rewrite which you can photocopy and reproduce on both sides of an 8 1/2-by-11-inch sheet. When you fold it, you get a 5 1/2-by-8 1/2-inch booklet of 4 pages. Though we lack the space to provide its 48-page source, we do provide copy for a second 5 1/2-by-8 1/2-inch booklet. The latter contains not only the original text and (photo-reduced) illustrations, but also instructions for each of five groups which will comprise your layout workshop.
I find it ironic in this column to rail against edgewise art and art separated from the text it supports. On this page and the following three, I wrap discussion around the pages discussed -- the approach I prefer. In the past, alas, sometimes the pages wrapped about provided minimal source for comment. I then filled in with bits about production techniques: PageMaker[R]", FreeHand[R], and the like. Not all have approved. A judge for one of the STC competitions wrote of an earlier column: "Too long. Cut out the stuff about computer programs. Not everyone uses PageMaker." Hey, I'm flexible. This time you're getting an extra 6 pages of edgewise booklet with no adjacent comment instead of an extra 12 pages of booklet plus (sometimes windy) comment. As you shuffle and rotate pages, blame it on the judge!
Page 1 of the workshop booklet explains the exercise and provides a schematic of the original 48-page layout. The differing assignments of student Groups 1, 2, 3, 4, and S begin respectively on pages 2, 4, 6, 8, and 10. Page 12 supplies a grid onto which participants can sketch their layout solutions. Each group has 30 minutes to consider answers to the several questions asked it. Each has 10 more minutes to present its conclusions to the class. You wander from group to group, participating in and guiding discussions. (You're well ahead of participants because you have a 4-page "solution" in your pocket.)
The workshop is open-ended. Your job is not to show how to pack 48 pages onto one sheet of paper. It's to get students thinking about convenient layouts.
You'll want to look at the workshop booklet now. The rest of this introduction answers questions the booklet asks of each group of your students.
Question 5 asks, "How might you condense tables ... so as to better serve readers?" Table A (left) provides a solution: One table on part of 1 page and flanked by related text replaces 3 edgewise tables filling 3 pages and flanked by no text. An inferior combination might have placed the stub head and box head common to the 3 tables above a Table A body containing the successive bodies of those 3 tables, each body identified by a field spanner: "Daily for Smog Months ... Daily for the Year ... 3 Highest for Smog Months." Table A goes further. It combines each of those entries under the reporting city. It arranges them not in the 1-2-3 order of its source but in a 1-3-2 order: Smog Month Highest ... Smog Month Daily ... Annual Daily. This is an order of descending smog density: the 3 worst cases for the 3 worst months, the daily average for those same months, and the daily average for the year.
Question 6 relates to scale and is answered with Question 14.
Question 7 relates to those nine line graphs which depict three sorts of smog data for Azusa, Los Angeles, and Riverside. And to the three bar graphs which depict a fourth sort for Azusa, Los Angeles, and San Barnardino. The question asks: "Why San Bernardino rather than Riverside?" The original text is clear: "The trend ... was not shown for Riverside because around-the-clock oxidant data during 1968-1970 were not reported. ...Graph 13 shows the number of hours 20 pphm was equalled or exceeded each year in San Bernardino, which is 10 miles from Riverside." The strong implication is that proximity guarantees a similar graphic result. Our inset shows that it does not. For each of the 7 years of record, background Riverside smog towers above that of foreground San Bernardino. The authors get to choose their city. But the editor might ask if they have weighed the benefit of incomplete but reasonable consistency against that of complete inconsistency.
Question 8 asks: "What's wrong with the checked Lennox and Long Beach data?" Here's what. For Lennox in 1972 and for Long Beach in 1963, reported average smog concentrations are worse throughout the year than during the smog months -- surely not true. With its new arrangement, the table helps editors spot such typos! And helps readers better understand the graphic data drawn from Table A.
Question 9 asks how we might eliminate Map 3 by adding its isopleths to Map 1 -- without using dashed lines. Map B (right) shows how: white lines. To minimize confusion, we eliminated county lines and names previously shown on cover Map A. Frankly, we think it didn't work out. Too busy. If we return to three maps -- two on the cover -- the report will continue to fit into 4 pages.
Question 10 asks: "Why does Fig. 40, sentence H25, cite a 34 percent drop for Anaheim when both Map 3 and the original text cite a 29 percent drop?" The careful editor always checks author computations. Ask your students to check the percentages which Map 2 reports for the 15 cities. They'll find the following 9 are wrong:
Anaheim 29 34 Pomona 21 23 Los Angeles  30 Corona 13 15 Reseda 25 26 Azusa 13 12 Long Beach 21 26 La Habra ? 1 Burbank 19 25
Checking Table A Los Angeles data is especially tricky because the highest 3-year mean for that city involves not 1965-67 (as employed by the authors) but 1964-66. The mean for 1965-67 is 15.8; that for 1964-66 is 16.4.
(16.2 + 17.3 + 13.9) / 3 = 15.8
(15.7 + 16.2 + 17.3) / 3 = 16.4
The drop from 15.8 to 11.5 (the '70-72 mean) indeed is  percent; but the correct drop (from 16.4 to 11.5) is 30%.
(15.8 - 11.5) / 15.8 = 27%
(16.4 - 11.5) / 16.4 = 30%
As data changes, so also must illustration. Map 3 must be redrawn as Map B because it repeats the errors just discussed. The greatest change is the sink which surrounds La Habra. Shown on Map B but not on source Map 3, the "percent drop" plunges from a nearby 20 percent to 1 percent. Thus, by our reading of data presented in Table A, the Map B "percent drop" isopleths form a totally different pattern than that provided in Map 3.
What about cover Map A? Prior to rewrite, the "Basinwide" text tells why La Habra data is missing from Map 2 (source of A): the authors lack adequate data for 1967. Indeed. Table A reports that 2 to 3 months of data is missing for that year. But Map 2 does report smog concentrations for Redlands, though 10 to 12 months of data is missing for the same year -- and totally missing as well for each of the 2 previous years. Indeed, Map 2 reports smog concentrations for 4 other cities for which from 1 to 4 years of data is missing. Thus the stated reason for omission is false.
Can we postulate an unstated reason? We can. As Map A shows, the La Habra "highest year" was 1969. Study Map 2. Its omitted 1969 isopleth would pass southeast from Pasadena to La Habra, northeast to Pomona, and thence further eastward to Riverside and Redlands. In so doing it would cross the imaginary 1968 isopleth which the authors did supply. They resolved this problem by omitting the line. Such omissions are common to mapmaking.
In How to Lie with Map (University of Chicago Press, 1991), Mark Monmonier writes, "A single set of numerical data ... can yield markedly dissimilar maps. By manipulating breaks between categories..., a mapmaker can often create two distinctly different spatial patterns. A single map is thus just one of many maps that might be prepared from the same information..." (p. 123).
Nevertheless, in basing the Map A 63-64 isopleth on data from two cities (West Los Angeles and Corona), we stretch credulity to the breaking point. But in Map 3 our authors actually break it. Based on data from a single city, Corona, they drew a 60-mile isopleth across northern Orange and western Riverside counties. Indeed, from data collected from only 15 points, both the authors and I have drawn sweeping isopleths across an area of some 3,600 square miles. If the U. S. Geological Survey did this with contours we'd map plateaus above Lake Erie and vasty lakes atop Pike's Peak.
Question 11 asks: "What does the Fig. 40 rewrite drop or change from the original `Summary' and `Basinwide' texts? And why?"
At 48 pages, the original requires a formal summary. At 4 pages, the rewrite does not. Instead, it begins with background, measurement methods, and a terse conclusion: "Max-hour concentrations are declining." It thereby avoids that repetition of detail so favored by lengthy summaries.
As students study the text, they'll see (with helpful prompting by you) that "Summary" sentences F18 and H24-26 account for more than half the "Summary" text and simply repeat (rather than summarize) half the "Basinwide" text.
Elsewhere in "Basinwide," crossout type marks weak speculation by the authors: in effect, "It's possible, if we get heavy smog, we may exceed past smog highs." That's one I'd strike without consulting the authors. (And be honestly surprised should objections arise.)
Question 12 asks students to describe how the use of subjects and verbs in the Fig. 40 rewrite differs from that of its source and to tell how that change affects the reader. The rewrite employs 12 main clause subjects, all but 2 (weather and data) concrete. The original employs 26 subjects, all abstract. The rewrite employs 11 main clause verbs; the original 29. Almost half of the rewrite verbs (45%) are transitive; all employ the active voice. Few of the original verbs (7%) are transitive; only 76 percent employ the active voice.
As a result, a typical main clause core of the original vaguely reads, "Averages are used." That of the rewrite strongly reads, "Anaheim experienced drop." Clear writing demands concrete subjects and active voice verbs.
Question 13 asks about graphic scales: "Would you combine the three graphs in each arrowed column? Or those in each bulleted row? Why?" To follow the bullets would be to group by data type, subgrouped by cities. In each group, the three city lines would tangle. Our inset (left) depicts the tangle of "3 highest" lines for Azusa (solid), Los Angeles (dashed), and Riverside (grayed).
Graph A follows the arrrows. Its Azusa data combines Graphs 1, 5, and 6; its Los Angeles, 2, 8, and 9; and its Riverside, 3, 11, and 12. For each city, the data is presented in order of lessening smog density, just as in Table A: in effect, 9 graphs (9 pages) as 1 graph (1/5 page). A single sweeping glance, and readers learn that in Azusa smog swoops up and down; that in Los Angeles it decreases, and that in Riverside it increases. Further (surprise!) throughout the 10 years, Los Angeles -- that synonym for SMOG -- proves far less smoggy than either Azusa or Riverside.
Questions 14-15 ask what we must do before we can combine graphs. We must assign identical scales to each. The original provides differing scales for cities and differing scales for each of the three measured concentrations within a single city. The authors scaled their graphs to fill the identically-sized rectangles they provided to frame them. Neat. But not helpful!
Question 16 asks why we moved original "Data Presentation" text to the rewritten "Three Cities" and "Smog Alert" texts. And why we moved original "Three Cities" text to the same "Smog Alert" text. To show these moves, and others, our inset outline lists sequential headings of both texts and keys original headings to alpha callouts of the rewrite. The rewrite chooses to group discussion of smog levels apart from smog alert hours and thus can provide a complete smog alert discussion beside the graph it illustrates.
B: Summary A: Introduction A: Data Measurement C: Data Presentation C/E: Three Cities D: Temperature Effects A: Basinwide
A: South Coast Smog 63-72 B: Smog Trends C: Three Cities D: Temperature Effects E: Smog Alert
Question 17 restates Question 7.
Question 18 asks why we eliminate the original's initial "Temperature Effects" paragraph. Well, without it, the text moves smoothly from [paragraph] K (The lower concentration may have resulted from cold air aloft) to [paragraph] L ([As) temperatures... dropped, ...concentrations [dropped].
Question 19 is answered by that deletion. The rewrite describes concentration decreases as a factor of temperature decreases because the second sentence now follows logically from the first. The original contrast, though accurate, proves disconcerting: Lower concentrations resulted from cold air; as temperatures rose, concentrations increased.
Question 20 requires us to consider how the text treats original Graph 4 (now Graph B). Two of the original sentences relating to this graph (L44-45, L45-46) imply that its correlations cover only 1957-70. In fact, as we later learn, correlations cover 16 years (through 1972); the regression line employs only the first 14 of those years (1957-70). The rewrite makes this clear.
Question 21 asks about sentence sequence. The rewrite is chronological, moving from 1968 to 1971-72. The original begins with 1971-72, then returns to 1968. When all else is equal, choose the more logical sequence.
Question 22 asks the source of rewrite Sentence N55. The source is Graph 5 itself: a note describing 1971 and 1972 points as "Not used to establish regression line." Though it does belong on the graph, the note requires fuller explanation. Thus, the rewrite text (N55) states: "[Points] fall so low [on Graph B] that we did not use them in plotting the regression line."
Question 23 asks about a line missing from original Graph 5 -- and from our look-alike Graph B. This line is a curve which would show actual concentrations: those from which the correlations were constructed. We can't ourselves draw it. In Table A, the authors provide combined data for the smog months (July, August, September) but not for August alone. We've made the vertical scale of Graph B match that of Graph A. We must continue to report 16 years rather than the 10 consistently reported elsewhere. Otherwise insufficient points would exist to construct the regression line.
Question 24 relates to scale and is answered with Question 14.
Question 1 relates to scale and is answered with Question 14.
Question 2 asks students to compute readability. The Gunning Fog Index of "Introduction" is 14.0; that of "Data Measurement" is 14.1 -- neither considered plain English. That of their Fig. 39 rewrite is plain English: 10.3.
Question 3 asks students if they would like to retain anything which the Fig. 39 rewrite omits. Marked by crossout, such omissions mostly cite broad trends which the maps, graphs, and later text detail better. Specific mention of sulfur dioxide generates no further discussion in a study which does not report it. Specific mention of an all-time high (56 pphm) may refer to Riverside in 1970 where the smog-month 3 highest readings -- reported in Table A and depicted in Graph A -- averaged 53 pphm. Except in Graph B (where it shouldn't) the text cites no other monthly data. Why begin here?
Question 4 allows students to critique the workshop. If you've helped each group work out answers, you'll find the critique favorable.