Controlling formatting in a word processor
Microsoft Word can be made to behave on a long document once the writer sees what the program is acting on, formats through named styles inside the client's template and switches off the options that redefine those styles. Pasting, page numbering, figure placement and the contents list each go wrong for reasons that can be found and removed, and a file can lose information on its way to a reviewer who uses another version of the program.
Why a long document misbehaves
As of March 2014, the long, complex documents that medical writers often work with can place unusual demands on a word processor, and Word, although powerful, is not always easy or intuitive to use properly. Many of the problems writers meet come from not knowing exactly how Word works, what unintended effects its options may have, or the best practices for certain writing and editing tasks. Many are not bugs but features of the program, which may have their uses in certain circumstances but can cause problems when the user does not know how they work. These features are described in March 2014 for Word 2010, and the description is generally relevant to earlier versions.
In one medical editor's experience, several of the most common problems in Word arise in page numbering, tables of contents, tables and figures. Page numbering goes awry or has a mind of its own, figures disappear, tables break oddly and bullets refuse to indent. A writer who knows how Word thinks may then be able to make it do what is wanted. Her procedures are for Word 2013, which she finds very like Word 2010, 2007 and 2003.
A medical writer's job is to use writing and scientific expertise to create high-quality documents, and the writer should not be wasting time fighting the word processor.
Formatting marks and what Word sees
What Word sees in a document is sometimes very different from what the writer sees, and that difference causes many of the problems in Word. It has many smart features that are good to use, but it is not good at telling the writer how to fix a problem, and when formatting goes badly wrong the writer is largely on their own.
The show/hide button shows the document as Word sees it. In the editor's office of about 30 people, where she was the only editor, 90% of the problems she was asked to solve came from writers who had not turned the button on. It need not stay on, but turning it on while copying, pasting or moving text, tables or figures lowers the chance of a problem.
With the button on, spaces appear as dots between the words, arrows show where tabs are set, and every paragraph ends in a paragraph mark. Unexplained vertical space, a tab and more than two or three spaces between words also become visible, which the writer may not notice but other people will.
Word puts most of its formatting in the paragraph mark, so a mark deleted by accident takes with it some, and probably all, of the formatting of that line. A style, likewise, is embedded in the paragraph mark at the end of every paragraph.
Two further displays show what the program is doing. Word can be set to shade links and hyperlinks, and the shading also shows what Word has numbered automatically, so the writer knows what is already linked and what the program is numbering. Small black boxes in the left margin show that a paragraph has been given a command to be handled in a special way: to keep its lines together, to keep it with the next paragraph, or to break the page above its first line. The four commands are checkboxes on the line and page break tab of the paragraph dialog, which opens from the arrow in the corner of the paragraph group on the Home tab. The first of the four, widow and orphan control, keeps two lines together on the last part of a page or the first part of the next, and the editor leaves it alone because Word handles it well. She recommends trying out the other three in the dialog, so that the writer knows how they work before meeting them in a document.
Options that apply to every document
In Word 2010 for Windows, as one trainer described it in winter 2014, the settings that apply to all documents are reached from the Options link on the File tab, and the trainer recommends checking them first; similar options are available in other recent versions. Among them:
- Display sets which formatting marks appear on screen when show/hide is on, and general printing instructions.
- Advanced sets the default for pasting, which can remove the small drop-down box that accompanies pasted text, and the shading of fields such as cross-references and tables of contents. With field shading set to always, the fields on a page can be seen at a glance, and so can an intended field that has not been properly styled.
- The display group under Advanced sets the width of the style area pane in draft or outline view, which shows the style applied to each part of the document. It is not the Styles window, but it can show why a passage does not look as expected.
As of winter 2014, where a field code that shows a cross-reference misbehaves, selecting the code and holding down Shift+F1 opens a pane of details that may give clues to the coding that needs correcting. Right-clicking the code and choosing Toggle Field Code moves between the output, such as "Figure 3-1", and the formula that produces it.
An en dash is the mark for a range, as in 17–25 pg/ml. As of March 2014, when Word takes what is being typed for a range, its factory setting turns a hyphen typed with a space on each side into an en dash as soon as the space after the second number is typed. Word leaves the spaces in, although traditional typesetting sets the en dash without them. One writer who runs courses, writing in March 2014, has switched off other presets, such as drag and drop and smart cursoring, but not this one, has given up taking the spaces out and expects the en dash very soon to usually have spaces.
Styles and formatting by hand
The Home tab gives a long stretch of its ribbon to styles because Word expects them to be used, and the editor promises that a writer who gets used to them will wonder how they managed without them. A style is a one-stop definition of the formatting of a piece of text, and it makes formatting consistent and saves time. Bullets that change from square to round, or in size or indentation, and paragraphs spaced differently from each other may not bother the writer, but they bother other people: a messy, inconsistent document makes a bad first impression, and a client may then read it with a more critical eye than a neat one would have met.
A rejected manuscript that goes to a second journal may have to change from 12-point Arial to indented 10-point Times New Roman. Changing the paragraphs one at a time, or with the format painter or Ctrl+Y again and again, might work for a 40-page manuscript and will not work for a 400-page dossier. A style makes almost any such change almost effortlessly, leaves the paragraphs consistent, and removes the worry that a page or a part of the text was missed. Reconstructing the formatting of a bullet by hand means looking at it, noting everything about it and possibly spending several minutes finding the settings, where applying a style takes one or two clicks.
Word keeps track of styles in a style list, and even a new document has styles on the ribbon. With text selected, hovering over the styles on the ribbon shows the text in each style, and double-clicking applies one. In a larger document with many kinds of text, such as paragraphs, bullets, headings and references, the editor finds a style easier to find in the drop-down list, which opens from the arrow in the corner of the style ribbon. The drop-down list is probably the more powerful of the two, because more can be done with each style in it. The list can be docked by dragging it, and Word remembers where it was put. Docked on the left, it does not cover the comments, which appear on the right.
Three icons at the bottom of the list create, inspect and manage styles. To create a style, the writer selects the text, clicks the create button and adds a memorable name. The editor recommends a name that says which element of the document the style formats, as in a first-level bullet, cover text, a table title, table text and table bullets. A client's style named with an abbreviation and a number leaves the writer wondering what the style is for, and a name memorable at least to the writer makes the style simpler to find. Where a document lacks the styles a writer wants, they can be taken from other documents: an element such as a bullet is copied into a blank document and a style is created from it.
The inspect button shows at once what type of style a style is, for instance that an opaquely named client style is a bullet and not a paragraph. The manage button is where comments that appear in a one-point font are corrected: the writer chooses manage, sorts the list alphabetically, selects balloon text, which looks greyed out but can be clicked, and increases the font size. The editor, who thinks eight points is Word's default size for comments, makes them a little bigger there.
The options link at the bottom of the list is small and easily missed, and it sets how long the list is. It offers four lists: the styles in use in the document at that moment, which was the editor's usual choice; the styles in the current document, used and unused; all styles; and the styles Word recommends, which she did not use. The same link sorts the list alphabetically, and its checkboxes show how a style's format has been changed, so that a first-level heading given a different font is listed as that heading style plus the font. With the boxes unchecked the list is clean.
The client's template
Before a regulatory writing project starts, the writer should find out whether a standardised template is expected. As of 2012, many companies have their own regulatory templates, which are often governed by standard operating procedures. The writer should check whether the file supplied is a document (.doc) or a template (.dot) and, where it is a document, ask whether a template file also exists, so that the styles used are the ones the company has approved.
A template file is downloaded to the appropriate template folder and then attached to the document. In Word 2003 and Word 2011 the route is Templates and Add-Ins on the Tools menu, where the Add button opens the choice of template. Once the template is attached, the document has its styles and macros, and the Normal template remains attached as well. In Word 2007 and 2010 the route is the Options dialog, then Add-Ins, where the Manage list is changed from COM Add-ins to Templates before Go opens the Templates and Add-Ins dialog and its Attach button. Selecting Automatically Update Document Styles there changes the styles of the document to those of the attached template.
As of 2012, a writer checks that the styles in a document are the compliant ones by choosing the normal view, which shows the style of each paragraph in the left margin. Choosing print layout closes the display.
Changing a style throughout a document
With the style list open, a bullet that has the wrong format is corrected by selecting it and clicking the intended style. The list also reaches every instance of a variant at once. Where it shows a paragraph style with an indented first line, the command to select all instances selects them, and one click on the basic paragraph style corrects them all. Headings that have drifted apart are corrected the same way.
A change to the style itself changes the whole document. When the spacing after one heading is adjusted and the change is applied to the style, every heading in the document changes with it, and a change of that kind can be undone. Numbered references work the same way: one entry is changed, the style is updated, and the rest follow, which shows little with three references and comes in really handy with ten pages of them.
A style that seems to be missing may be in the document unused. One document had a title style for figures in one appendix and seemed to lack the one for another, and listing the styles in the current document showed that it was there to apply.
A page break can be made part of a style. Where a figure title sits too close to the foot of a page, a manual break can be inserted there, or the style can be modified through its paragraph format to break the page before it, and the page then breaks at every instance of that style in the document.
Options that redefine styles
As of March 2014, two options under Advanced in Word's options change styles without being asked. With Keep track of formatting turned on, Word watches the writer and defines a new style for each piece of formatting applied to an existing style. Where the base style is Body Text in 12-point Arial, making one paragraph bold creates a style called Body Text + Bold and reducing another to 10 points creates Body Text + 10 pt, and this goes on until the style list is a metre long and the styles the writer needs are slow to find. The option is turned off from the File tab, under Options and then Advanced. In a document inherited with such styles already in it, deleting them is possible, but hiding their names is easier: the options link at the bottom of the style list opens the Style Pane Options dialog, where three options are turned off.
Formatting that changes seemingly for no reason is likely to be the work of Prompt to update style. It updates a style's definition to match a manual change, so that reducing one Body Text paragraph from 12 points to 10 gives every Body Text paragraph, current and future, a 10-point font. That is probably not what the writer wants, and it can be particularly problematic in a client's template, where a writer can change one of the client's carefully planned styles without meaning to. The option should be turned off and left off. It stands in the same dialog as Keep track of formatting.
Spacing and breaks made with repeated keystrokes
Three keystroke habits each have a proper replacement. Pressing Enter twice for white space between paragraphs adds a blank paragraph that can cause problems later, and the space belongs in the paragraph's style as space before or after it. Pressing Enter repeatedly to move an item to the top of the next page may look fine at first, but things shift as text is added or deleted, and the proper ways are a hard page break, made with Ctrl+Enter before the item, or the keep-with-next option in the paragraph dialog. Pressing the spacebar repeatedly to move a word to the next line is replaced by Shift+Enter, which starts a new line without starting a new paragraph.
Spaces used for alignment fail in two further cases. A tab, not spaces, goes between a heading's number and its title, so that the entries in the contents list and the lists of tables and figures align: in one example sections 6 and 6.1 align in the contents list and section 6.2, numbered with spaces, does not. In an in-text table, content is indented within a cell with the tab ruler at the top of the window and not with the spacebar, because the ruler's indents survive in a clean PDF and the spacebar's may not.
Pasting without the source's formatting
Copied text may arrive with the formatting it had in its source document. That is sometimes wanted, but more often it causes problems because the destination's formatting is the one wanted, and the text can also bring style definitions from the source that conflict with the document's own styles. Pasted text can arrive with its own fonts, sizes and links, and straightening it out afterwards, even with the format painter, wastes time that can be saved when the text is pasted.
To paste it unformatted, the writer copies the text, navigates to the place where it belongs, clicks the arrow under the Paste button on the Home tab and chooses unformatted text, from the Paste Special dialog or from the paste options. The text is pasted as if it had been typed in, and the formatting in effect at that place is applied to it by default, although it can then be formatted differently.
Show/hide should be on while the writer navigates to the place for the text. The writer then knows exactly where the text is going and which formatting or style it will take, and without it the text may land in a bullet, in a footnote or, worse, in a numbered heading. Show/hide controls where the text goes, and styles control what it looks like.
As of March 2014, pasting as unformatted text has a shortcoming. Non-standard symbols such as ≤ and © are lost, as are non-breaking spaces and hyphens, and they are typically replaced by small boxes in the pasted text and have to be put back by hand.
When copying from one document to another, a regulatory writer, writing in 2012, selects everything except the paragraph mark in which the style is embedded and then uses Paste Special, typically with unformatted text, so as not to carry unwanted styles into the template. The pasted text then takes the style embedded in the destination document and not the one it was copied from.
The other paste options serve other purposes. Merge formatting adapts the text to its destination automatically, and frequently used commands can be set up on the quick access toolbar. Keep text only strips out any links, and keep source formatting is for the occasions when the original appearance is needed. The choice deserves particular care with material taken from the web.
Page numbering, section breaks and linked headers
In the editor's view, page numbering is the commonest complaint of anyone who has handled a large document with several parts. One document can number its pages in several ways: no numbers on some pages, Roman numerals for the introductory parts, Arabic numerals for the text, a scheme such as A-1 for the appendices, or the form "page 1 of 20". One writer recommends that every page of any document carry its number and the total number of pages, so that anyone can be sure that all the pages are there.
Page numbering depends on two things, linked headers and footers and section breaks, so a numbering problem is very likely to be one of the two. A section break is not a page break, although many people take it for one. It tells Word that a new section begins, in the sense of a chapter or a different part of the document, and a section can have its own paper size, margins, orientation and page numbering. Because Word reads the break as the start of something new, it often resets the page number to 1, so that after a section break on page 28 the next page is numbered 1. The break that caused it is found by turning on show/hide.
A linked header or footer tells Word to make it look like the header or footer of the section before. A linked header inserted in Appendix B is linked to that of Appendix A, so whatever is typed into the header of Appendix A, such as "Appendix A", fills the header of Appendix B as well. Some people like the automation, and it is useful in places in complicated documents, but in the editor's view it brings more problems than it solves. She removed every link as a matter of routine, without ever regretting it, and advised writers to add none and, until they are confident with page numbering, to remove them all.
Unless a document is very complicated, section breaks are generally needed in only four kinds of place: after the title page or the front matter, such as the table of contents or the acknowledgements, where the main text starts; before a change from portrait to landscape orientation, and again to change back; at the end of the main text, before the appendices; and after each appendix, because the appendices will probably be numbered differently. No other place needs one. Whether a place needs a break can depend on the organisation. The editor's cover page needed one because it had a larger margin, and an organisation that does not count the table of contents and the lists of tables and figures as part of the main text would want a section break among them.
The repair is made in three sweeps from the top of the document, which can be combined once the writer is proficient, and the same approach and commands fix page headers. The first sweep looks for section breaks and asks of each whether it is needed. A break can hide behind a paragraph mark and is found by pulling the paragraph mark down. Breaks that are not needed are deleted, and a missing one, such as one between two appendices, is inserted from the breaks menu on the page layout tab as a next-page break. The style list is closed first, because page numbering does not need it and the screen space is needed later.
The second sweep removes the linked headers and footers. Links do not show in the main part of the document, so the header and footer area has to be opened. With field shading on, text that somebody typed there can be told from what Word generates and can be deleted without changing the page numbering. Each link is removed as it is met, including one that does no harm at present, because it adds confusion when pages are added.
The third sweep sets the numbers. A section told to start at 2 still starts at 2 when the title page before it is deleted, and does not fall back to 1, so the editor set the section to continue from the previous section. Where Word keeps renumbering pages in an endless loop as they are corrected, a link has been left in somewhere. The editor suggested practising page numbering a lot.
Figure placement
Figure placement was the second most common thing to go wrong in the editor's office: the figure is placed in the document and then disappears, or floats on top of the text somewhere else. The editor's method is the simplest placement, because most biomedical papers are not fancy and do not need figures placed behind the text.
Two things decide whether placement works: the paragraph the figure is put into, and the figure's layout, which is its text wrapping. The first step is always to turn on show/hide. The paragraph is then formatted, before the figure goes in, by clearing its formatting, by applying the normal style, or by applying a style set up for inserting figures. Such a style has its line spacing set to "at least" and never to "exactly". The number does not matter, but with exact spacing the figure disappears: a figure in a paragraph set to exactly 12 points shows only 12 points of itself. The figure is then pasted in, right-clicked and set, under Wrap Text, to In Line with Text, which is the step that some people forget. In one case a figure went wrong although its paragraph's line spacing was right, because its wrapping was set to tight, and setting it in line with text corrected it.
The style the figure sits in matters as well as its appearance. A figure placed in a paragraph that has the figure-title style can look right with show/hide off, and is then likely to be picked up when the table of contents is generated. Clearing the paragraph's formatting and checking that the layout is in line with the text corrects it.
The table of contents and its styles
Heading-level styles used consistently throughout a document let the table of contents be generated automatically, as of 2012 from Insert, then Reference, then Index and Tables, then Table of Contents. Word's tables of contents update automatically to reflect changes in the document.
As of March 2014, formatting applied by hand to a table of contents vanishes because the table is regenerated. Word generates the table from the headings in the document and controls its appearance with styles named TOC 1 for the top-level entries, TOC 2 for the second level, and so on. These styles do not appear in the style list, because only tables of contents use them, but they control font, alignment, indentation and justification as other paragraph styles do. When the table is updated, Word deletes the old one and generates a new one from the TOC styles, and the manual formatting is replaced.
The formatting must therefore be changed in the TOC styles. With the cursor anywhere in the active table, the writer clicks the Table of Contents command in the Table of Contents group on the References tab, chooses Insert Table of Contents near the bottom of the menu, and clicks Modify in the dialog that follows, which lists the TOC styles. Each style is selected and modified there as any other paragraph style is, and the changes are applied whenever Word updates the table.
Files shared across versions of Word
As of December 2012, everyone in the industry seems to use a different version of Word. A regulatory medical writer working on a Mac is advised to save work as both docx and doc files, circulate both, and remind team members to work with the version compatible with their own system.
Mark-ups, including comments and tracked changes, are then visible to everyone, but information on who made them can be lost between versions of Word. To keep it, the writer should add initials to every comment and tracked change and have the other stakeholders do the same.