My approach to editing began with creating a stylesheet and cleaning extraneous and unwanted typing mistakes from the document (see The Business of Editing: The AAE Copyediting Roadmap II), moved through tagging the manuscript by typecoding or applying styles (see The Business of Editing: The AAE Copyediting Roadmap III) and inserting bookmarks for callouts and other things I noticed while tagging the manuscript (see The Business of Editing: The AAE Copyediting Roadmap IV), to creating the project- or client-specific Never Spell Word dataset and running the Never Spell Word macro (see The Business of Editing: The AAE Copyediting Roadmap V). The last stop was using wildcards to fix reference formatting problems, running the Journals macro to correct incorrect journal names, and editing the reference list (see The Business of Editing: The AAE Copyediting Roadmap VI). Now it’s time to tackle duplicate references using the Find Duplicate References macro.
Until recently, finding duplicate references was difficult and very time-consuming. I often deal with reference lists of 300+ references, with many lists running between 500 and 800 references and some running close to 2000 references. Before I created the Find Duplicate References macro, the only way I had to check for duplicate references in a numbered reference list was to use Word’s Find and do two or three different searches based on the same reference. One search might be on author names, another on article or book title, and a third on the cite information (i.e., journal name, year, volume, and pages). Unfortunately, many authors are sloppy with how they cite references so that the same reference is cited slightly differently each time it is cited. Sometimes a reference is cited completely, other times a reference is missing material.
Careful editing of references solved part of the problem, but duplicates of each reference still had to be searched for individually. Time — and profit — flew away.
Today’s Approach: Find Duplicate References
The process was taking too much time and costing me too much profit. I needed a better solution, which led to EditTools’ Find Duplicate References (FDR) macro, a much quicker and better solution to the problem of finding duplicate references.
As good as FDR is at finding duplicates (and from the heavy use I gave it in my last project, which project had more than 21,000 references in total, I know it is very good), it is important to remember that FDR, like other macros, is dumb — it will find only exactly what it is told to find, not something close. If two entries in the reference list are identical except that one has an extra space, FDR might not tag them as possible duplicates because of that extra space. Similarly, if the references are identical except that one uses a colon to separate portions of the article title and the other uses a dash, they will not be tagged as duplicates. Even the dashes have to be identical. For example, a page range that is identical except that one uses a hyphen as a separator and the other an en-dash will result in the cites not being tagged as identical. Again, close only counts in horseshoes.
Tip: Because the macro looks for matches within set number of characters, it is occasionally worthwhile to run the macro more the once using a different number of characters as the search parameter. The macro lets you choose 24, 36, 48, 60, 72, 84, 96, 108, or 120 as the number of characters in the search string. Consequently, if you choose 96 characters and two references are identical except that one uses an em-dash and the other uses an en-dash in the article title, the macro will not find them as duplicates. (This is why the macro does two passes — one from the beginning of the reference forward and one from the end backward — in case there is a match from one direction even if not from both directions.) Changing the search string length to 72 might find the duplicates if the dashes appear as the 73rd character or later. Of course, it may still not find the duplicates if the opposing characters still appear within the search string length. The macro is dumb; the characters within the search string must be identical.
The dialog that appears when FDR is run, which is shown below (you can make the image, as well as other images in this essay, larger by clicking on the image), provides detailed information about the macro. As I noted previously, it is my habit to move the reference list to its own file. I do this for several reasons: First, the Journals macro runs more quickly because there is less material it needs to check.
Second, if the manuscript requires my using Superscript Me (see The Business of Editing: The AAE Copyediting Roadmap VI), it eliminates the possibility that Superscript Me will make unwanted changes to reference cites (e.g., changing 1986;52(14):122 to 1986;5214:122 or 1986;5214:122).
Third, it makes it easier to renumber and/or add or delete references during editing of the manuscript (I use three monitors and have found it is easier and quicker to access and edit the references when the text is open on one monitor and the reference list is open on a second monitor).
Fourth, the Find Duplicate References macro does several things to the document on which it will run before running, namely, save the current document, create a copy of the current document, remove any highlighting and queries/comments, and accept all changes (see #1 in below image). The idea is that the duplicates will be found in the copy document but the editor will note them in the original document, which is the document that the client will see.
The longer the reference list, the more important I think it is that the reference list is moved to its own document. (When I am done editing the manuscript, I reincorporate the reference list in the edited document. I turn tracking off in the manuscript and use Word’s Insert Text from File feature to reinsert the reference list with all its tracked changes. I then turn tracking back on in the manuscript and save the file.) But if you do not want to move the references, you can leave them where they are and use the Bookmarks buttons (see #4 below) to insert the required dupBegin and dupEnd bookmarks at the beginning and end, respectively, of the reference list. (These bookmarks are not needed if the reference list is in its own file.)
The key to the Find Duplicate References macro is remembering that the macro only identifies information that is identical (see #2 in the FDR dialog image above). Consequently, after running the Journals macro and before running FDR, I edit the reference list, making the references consistent. All page ranges, for example, use an en-dash; every time the CDC is named as the author, the name is conformed to “Centers for Disease Control and Prevention (CDC).” In addition, I check URLs and add any missing information.
Finally, I tell the macro the number of characters I want it to match (see #3 in the FDR dialog image above; also see the Tip above). Because the macro is a two-pass macro, it will check that number of characters (including spaces) from the beginning of the reference forward (it ignores the reference number and begins the count from the first alphanumeric character that is part of the cite itself) and then from the end of the reference backward.
When done, FDR produces a “report” that it places at the beginning of the reference list that looks like this:
When the report is generated I use Word’s Find Pane to check each entry. (For a more in-depth discussion of the process, see Find Duplicate References at the wordsnSync website.)
As I go through the list of possible duplicates, I mark those I find that are duplicates. However, I do not want to make changes to the reference list at this point; I just want to mark the duplicates. To mark them, I do two things: First, I insert a standard comment using the Insert Query macro, replacing the underscore with the numbers of the references that are duplicates of the current reference:
I also insert a bookmark at each location using the Bookmarks macro. I use this format (see the highlighted text):
The bookmarks act as a check, as well as make it easier to deal with the duplicate references. When, for example, during editing of the text I come to the callout for reference 18, I can see — from the comment and the bookmarks — that three other references are identical to 18, namely references 72, 91, and 102. Should the author have numbered references out of order and called out reference 91 before 18, I can see at a glance which references are duplicates. The bookmarks let me easily navigate to each of the duplicate references; once I have deleted a duplicate reference, I can delete its bookmark. The bookmarks provide an easy way to track which duplicates remain.
Recording & Reporting Duplicates
I also mark the information in the Reference Number Order Check macro (which is the subject of The Business of Editing: The AAE Copyediting Roadmap VIII). The Reference Number Order Check macro can provide my client with a report showing which references were deleted as duplicates and what those references were renumbered as. A sample report is shown here:
As the sample report shows, references 78 (#5) and 201 (#6) were deleted and all callouts numbered 78 were renumbered as 19 and all callouts numbered 201 were renumbered as 85.
Find Duplicate References works very well. In one chapter I edited in a recent project, the macro found 23 duplicate references in a 700-entry reference list (one reference was duplicated five times!). It took the macro seconds to find those duplicates; had I looked for them without using FDR, it would have added hours to the project and turned the project into a profit-loser.
Richard Adin, An American Editor