Research
Examining Error in the Technical Communication Editing Test
Researcher’s note (June 2026). I wrote this early in my time at North Texas because the field’s sense of which errors matter was almost entirely anecdotal, and I wanted to see what real editing tests actually penalized. The core idea was to treat the editing test as the one authentic workplace document that deliberately contains errors, and let it tell us how technical communicators prioritize them. That same question — what counts as an error, and who decides — is now central to my work on AI in writing and assessment, where the tools generate and flag the very usage issues this study tried to catalog.
Overview
Usage error is a popular topic for technical communicators. However, its anecdotal discussions remain the best source of information on the errors that technical communicators might value over others.
In this paper, I report the types and frequencies of errors found in 41 editing tests administered to prospective technical writers and editors.
Results indicate that misspellings and faulty/missing capitalization were the most frequent and dispersed errors. Eight of the most frequent errors related to style; however, grammar and punctuation errors remain the most dispersed. A larger dataset will better determine how technical communicators prioritize specific errors.
BACKGROUND
Usage error remains a popular topic for technical communicators. A recent discussion thread on STCTESIG-L solicited input for a newsletter article entitled, “The Top Ten Errors That Technical Communicators Make.” The thread elicited a myriad of responses, including the correct use and punctuation of restrictive and nonrestrictive clauses and the ungrammatical use of “for example.” Unfortunately, these anecdotal discussions remain the best source of information on error for technical communication practitioners and academics.
The results from relevant studies have only reported how distracting specific errors are to practitioners instead of examining the actual errors found in workplace writing (e.g., Beason, 2001; Gilsdorf & Leonard, 2001; Hairston, 1987; Leonard & Gilsdorf, 1990). Two empirical studies identified the prominent errors in college writing (Connors & Lunsford, 1988; Lunsford & Lunsford, 2008), but these results do not necessarily reflect how practitioners, in general, or technical communicators, in particular, prioritize errors.
In this paper, I begin to address this gap by discussing the types and frequencies of errors found in 41 editing tests administered to prospective technical writers and editors. Editing tests evaluate an applicant’s ability to spot obvious typographical errors as well as fix rather than introduce new errors (Hart, 2003, p. 12). The editing test remains the only workplace document to purposely contain errors and therefore serves as a first step in identifying how technical communicators prioritize specific errors. Results will extend the earlier-noted research by providing the first list of errors that were derived from an authentic workplace document. These data will help prospective technical communicators prepare to take an editing test and help employers evaluate how well their own test assesses its applicants. Finally, these data can be compared to the existing taxonomies of errors in college writing. Identifying discrepancies in how college and workplace writers are trained can help instructors better prepare their students for the writing and editing professions.
LITERATURE REVIEW
Within the fields of business and technical communication, only a handful of studies have empirically addressed usage error. Hairston’s 1981 study was the first to determine how practitioners responded to specific usage errors. Hairston constructed a 65-sentence questionnaire in an attempt to understand the emphasis business professionals placed on various errors. The 84 respondents, who represented 63 occupations, were overwhelmingly bothered by errors that Hairston classified as status markers: “When Mitchell moved, he brung his secretary with him” (p. 796). Other status markers that overwhelmingly bothered participants included double negatives and beginning a sentence with an objective pronoun. The next tier of bothersome errors were grouped by mechanical mistakes—sentence fragments, fused sentences, and faulty parallelism. Errors that bothered participants the least included the comma splice and its/it’s error.
Leonard and Gilsdorf (1990) extended this study by comparing academics’ and executives’ reactions to specific errors. The researchers also used a questionnaire to measure these perceptions; their population was composed of randomly selected members from the Association for Business Communication. The questionnaire contained 58 sentences, each one including a single usage error. The list of most distracting errors often aligned with Hairston’s: fused sentences, faulty parallel structure, sentence fragments, and danglers. Participants appeared least bothered by sentences that included “feel” instead of “believe” or the clipping of “quotation” to “quote.” The results also indicated that academics were much more bothered by certain errors than executives were, particularly its/it’s confusion, missing commas, and missing apostrophes. Gilsdorf and Leonard followed up on their own study in 2001 and found strikingly similar results.
While these studies generated salient findings, the usefulness of the data is somewhat limited by the methodological design. The questionnaires included errors that the researchers believed could be the most bothersome to practitioners. These results may not accurately reflect the errors that practitioners value. Similarly, data collected from questionnaires depend on self-reporting, which can motivate participants to respond in ways they think are appropriate to the research. While these researchers all stressed to participants that they were not being tested, this potential threat to validity is difficult to control. Hairston conceded: “I feel certain that many respondents self-consciously looked for the error in each sentence, and there was a certain amount of ‘trying to do well on the test’” (p. 798). She added that it was easier for participants to find one error isolated in a sentence rather than multiple errors scattered throughout a longer text.
Error taxonomies of college writing compensated for these methodological limitations by assessing the errors in authentic writing samples, but the results cannot necessarily be generalized to the workplace or the field of technical communication. Connors and Lunsford’s study (1988) produced a list of over 50 formal and mechanical errors that college students made in their writing (p. 396). Misspellings outnumbered the other errors by 300% and were removed from the formal study for independent analysis. Connors and Lunsford ranked the remaining errors by occurrence, selecting the top 20 for further inquiry. The list began with missing comma after an introductory element (occurring 11.5% of the time) and ended with its/it’s error (occurring 1.0% of the time).
In 2008, Lunsford and Lunsford extended the original study. The results reflected how a broader use of academic genres and the expansion of technology changed the error patterns in college students. Due to an increase in argument papers, the new list included errors related to using sources, quotations, and attributions. Technology also played a role in the rank of specific errors. Misspellings now ranked fifth, and wrong word errors emerged as the top error. Lunsford and Lunsford attributed these shifts to electronic spellcheckers. The technology helps students remedy misspellings, but a reliance on the automated spelling suggestions likely correlates to the increase in wrong words.
Both of these studies comprise the most thorough and current error taxonomies, but they do not specifically relate to workplace writing where stylistic approaches — tone, word choice, consistency — could play a more prominent role. Therefore, I examine the errors found in 41 technical writing and editing tests. Editing tests remain the only workplace document to purposely include errors, so this study operates under the assumption that these are the errors that technical communicators prioritize over others. While technical communicators arguably value error more than other practitioners, the results from this study begin to suggest how practitioners, in general, prioritize error compared with academics.
Further, this examination contributes empirical data to technical editing, which remains the most underdeveloped scholarly subfield of technical communication. Researchers have only recently moved beyond anecdotal approaches of inquiry to explore issues on larger, more empirical scales, including the alleged adversarial relationship between editors and their clients, and the presumed differences in communicating edits to native and nonnative clients (Eaton, Brewer, Craft Portewig, & Davidson, 2008a, 2008b). Eaton (2010) notes that anecdotal explorations only produce “cup of coffee articles” — the resulting research is useful, but only as useful as having a cup of coffee with someone and chatting about an experience (p. 9).
METHODOLOGY
Because of the privatization of editing tests, the study’s sample proved difficult to collect. I obtained 41 editing tests through my requests on various email lists, including STC’s Technical Editing and Management SIGs, Copyediting-L, and Edprof. I signed a nondisclosure agreement with a majority of the companies to further protect the integrity of the tests.
Participating organizations represented a variety of industries, including computer programming services, healthcare and social assistance, and education services. Organizations were from 15 U.S. states with the most representation in California (n=8) and Texas (n=6). This majority correlates with the findings from STC’s 2008 Salary Database, which identified California and Texas as the two states with the most technical writers (STC’s 2008 Salary Database, 2009).
Every error in the editing tests was coded and then quantified. When noted, tests were coded with the appropriate style manual, primarily The Chicago Manual of Style (15th edition). Two raters independently coded the tests on the basis of the supplied answer keys (when available). This approach ensured that the errors were classified from the organization’s perspective. Whenever possible, errors were classified by the error/error patterns used in the Connors and Lunsford and (or) the Lunsford and Lunsford studies; however, multiple new errors related to style were identified. The Cohen kappa test identified an overall inter-rater agreement of 81.3%, indicating a high level of consistency.
RESULTS & DISCUSSION
Coders identified 72 errors within the 41 editing tests. Each test contained an average of 54.7 errors (SD =14.1) and an average of 20.2 different errors (SD =5.66). The standard deviation of both means indicate disparity, which could relate to the privatized nature of the editing test and the lack of publicly available examples.
Table 1 summarizes the top 20 errors by frequency, beginning with misspellings (13.6% of total errors) and ending with repetition (2% of total errors). Spelling and capitalization emerged as the top two most frequent errors. Results indicated that the presence of both correlated to style conventions. Seventy-one percent of the misspellings were classified as general, and many of these errors included the misspelling of proper nouns like company or solution names. Twenty-five percent of the misspellings were homonyms (e.g., affect instead of effect), and 4% were compound words (e.g., bi-polar instead of bipolar).
This distribution correlates more with the original Connors and Lunsford study rather than the updated study. Seventy percent of the student papers that Connors and Lunsford assessed in 1988 were handwritten (p. 398) compared with “almost every” paper being word processed in Lunsford and Lunsford’s 2008 follow-up study (p. 79). As mentioned earlier, the latter researchers noted that the medium of the student-written texts likely attributed to the decreased rank of misspellings — the electronic spellcheckers lessened the frequency of this type of error but simultaneously produced an increase in wrong words.
However, a sizable majority of misspellings identified in this study were not proper nouns and therefore were detectable by a spellchecker. Further analysis revealed this discrepancy might be linked to how companies administered the editing test. Sixty-one percent of companies administered the test on-site, and 74% of those tests were paper-based.
The anxiety of the testing situation coupled with the paper-based administration likely rationalizes the inclusion of these general misspellings. The introduction of these more noticeable errors also supports Hart’s description of the editing test’s purpose: to measure applicants’ ability to spot obvious typographical errors.
The misspellings and capitalizations that related to proper nouns often referenced the company or a product. While 56% of companies required applicants to demonstrate knowledge of a style guide on the test, most other companies integrated these style-related errors into the tests differently. One test provided applicants with the correct spellings of four proper nouns in bold print, and its instructions noted that “everything in boldface is correct.” The test’s creator said that the instructions assessed how consistently applicants applied spelling and capitalization and followed instructions.
Another test included the correct spellings and capitalizations of proper nouns on the first page but then misspelled them throughout the subsequent pages. The test’s creator said this approach gauged how perceptive applicants were in catching errors and then determining which spelling or capitalization forms looked consistent with the rest of the document. Finally, dozens of tests included a misspelling of the company name. Many of these tests’ creators and hiring managers cited branding as a motivation behind introducing this error as well as measuring applicants’ familiarity with the company.
Table 1: Most Frequent Errors
| # | Error or Error Pattern | # found in 41 tests | % of total errors |
|---|---|---|---|
| 1 | Spelling (including homonyms) | 305 | 13.6% |
| 2 | Unnecessary or missing capitalization | 152 | 6.8% |
| 3 | Hyphen, em- or en-dash error | 93 | 4.1% |
| 4 | Language or nomenclature consistency | 86 | 3.8% |
| 5 | Wrong word | 83 | 3.7% |
| 6 | Faulty predication | 74 | 3.3% |
| 7 | Format/consistency text | 69 | 3.1% |
| 8 | Unnecessary shift in verb tense | 68 | 3.0% |
| 9 | Unnecessary or missing apostrophe (including its/it’s) | 62 | 2.8% |
| 10 | Number, date, percentage, time format | 54 | 2.4% |
| 11 | Wordiness, rewrite for concision | 53 | 2.4% |
| 12 | Misplaced or dangling modifier | 51 | 2.3% |
| 13 | Missing or incorrect article | 51 | 2% |
| 14 | Lack of subject-verb agreement | 50 | 2% |
| 15 | Organization | 49 | 2% |
| 16 | Unnecessary passive construction | 48 | 2% |
| 17 | Faulty parallel structure | 48 | 2% |
| 18 | Missing comma in a series | 47 | 2% |
| 19 | Missing comma with a nonrestrictive element | 46 | 2% |
| 20 | Repetition | 44 | 2% |
Eight of the top errors were stylistic and not represented in the earlier referenced taxonomies. Language or nomenclature consistency (#4) related to how well test takers maintained consistency throughout the document (e.g., a specific solution name would change throughout the test). Format/consistency text (#7) related to the consistency of heading and term formats, including the use of boldface and italics as well as font choice and size. Number formats (#10) related to the spelling of numerical information (e.g., 100 versus one hundred, % versus percent). Wordiness (#11) was found on a word level that often involved adverbs (e.g., it was a very hard puzzle) and on the sentence level as in the following:
Original: The Risk Control Organization will establish a relationship of meeting with the Research department on a regular basis for the purpose of discussing model development activities.
Revision: Risk Control will regularly meet with Research to discuss model development.
Organization (#15) included the need for directional language (e.g., first, next, finally) as well as the need for topic, transitional, and forecasting statements. Since most company representatives provided an answer key with their test, it was easy to determine when passive construction was considered unnecessary (#16). Faulty parallel structure related primarily to a lack of parallelism within a list. Repetition (#20) appeared at the word level (e.g., Laura was vaccinated for the HPV virus) and the sentence level.
However, Table 1 only demonstrates how frequent (or popular) a specific error was within the 41 editing tests. This information proves valuable in better understanding how technical communication practitioners prioritize error, but it does not offer applicants a sense of how the errors were dispersed throughout the tests. The data in Table 2 summarize how the errors were dispersed throughout the tests. These data indicate which errors clustered in a small number of tests and which errors appeared more consistently throughout the 41 tests. Errors related to spelling and capitalization remained the top dispersed errors, solidifying their prominence in this dataset. Fourteen additional errors appeared in both the most frequent and most dispersed lists. Notable jumps in rank included missing comma with a nonrestrictive element (#19 in frequency, #4 in dispersion) and faulty parallel structure (#17 in frequency, #9 in dispersion). In other words, these errors might not have appeared multiple times within a test, but they had a presence in 65% and 55% of the 41 tests, respectively.
Table 2 also indicates that some of the most frequent errors did not appear in a majority of the editing tests. Interestingly, the four errors that were frequent but weakly dispersed were all stylistic conventions: number format (#22 in dispersion), format/consistency of text (#25), wordiness (#28), and organization (#33). These conventions were replaced with grammar and punctuation errors: unnecessary comma, wrong or missing preposition, missing comma in a parenthetical or transitional expression, and incorrect singular/plural use.
Table 2: Most Dispersed Errors
| # | Error or Error Pattern | % found across 41 tests | # by frequency in 41 tests |
|---|---|---|---|
| 1 | Spelling (including homonyms) | 75.6% | 1 |
| 2 | Unnecessary or missing capitalization | 65.9% | 2 |
| 3 | Faulty predication | 65.9% | 6 |
| 4 | Missing comma with a nonrestrictive element | 65.9% | 19 |
| 5 | Misplaced or dangling modifier | 63.4% | 12 |
| 6 | Wrong word | 61.0% | 5 |
| 7 | Unnecessary shift in verb tense | 61.0% | 8 |
| 8 | Hyphen, em- or en-dash error | 58.5% | 3 |
| 9 | Faulty parallel structure | 56.1% | 17 |
| 10 | Missing or incorrect article | 53.7% | 13 |
| 11 | Lack of subject-verb agreement | 51.2% | 14 |
| 12 | Missing comma in a series | 51.2% | 18 |
| 13 | Unnecessary comma | 51.2% | 21 |
| 14 | Unnecessary or missing apostrophe (including its/it’s) | 48.8% | 9 |
| 15 | Unnecessary passive construction | 46.3% | 16 |
| 16 | Wrong or missing preposition | 43.9% | 23 |
| 17 | Missing comma in a parenthetical or transitional expression | 43.9% | 24 |
| 18 | Incorrect singular/plural application | 43.9% | 30 |
| 19 | Language or nomenclature consistency | 41.5% | 4 |
| 20 | Repetition | 41.5% | 20 |
In other words, while these stylistic errors appeared frequently in specific tests, their presence was not representative of the study’s dataset. For example, errors related to number formatting appeared in only 39% of the editing tests, and errors related to organization appeared in only 0.3% of the tests. Likewise, the grammatical and punctuation errors that rose in rank were dispersed throughout 44% to 51% of the tests.
The findings in Table 2 suggest that style-intensive tests exist but are infrequent. However, stylistic conventions had a strong presence in all of the editing tests, including faulty parallel structure, active voice preferred, language or nomenclature consistency, and repetition. Similarly, test takers should always be aware of their own justification for fixing an error in an editing test. Hiring managers indicated that they examine the decision-making process behind fixing specific errors as well as observe how well the candidate follows instructions and how informed they are about the company (thus, the rationale for the frequent misspellings of company names).
CONCLUSIONS
I conducted this study to offer technical communication researchers, practitioners, teachers, and students some empirical insight into error usage in their field. Further empirical exploration will advance our field’s research beyond the anecdotal and simultaneously strengthen the pedagogy and practice in technical communication.
The results of this research, however, begin to suggest commonalities and differences among existing error taxonomies. For example, missing comma after an introductory element appeared frequently in the Connors and Lunsford study (#1) as well as in the Lunsford and Lunsford follow-up study (#2). However, these results also revealed that teachers marked this error in their students’ papers only 30% and 28% of the time, respectively. These percentage levels suggest that teachers may not consider missing commas in introductory elements as distracting as other errors.
This inference seems to be reflected in this study’s results as well; the error ranked #27 in frequency (appearing only 29 times) and #23 in dispersion (appearing in 39% of the tests). Since the editing tests purposely contained errors, as opposed to the errors introduced by students, a larger dataset might indicate if this error, while frequently made, is of lesser concern to technical communicators than other error patterns.
On the other hand, Gilsdorf and Leonard’s study found that errors that troubled readers the most related to basic sentence structure, such as fused sentences, fragments, faulty parallel structure, and dangling modifiers. Faulty parallel structure and misplaced and dangling modifiers appeared toward the top half of this study’s frequency and dispersion list, but fused sentences and fragments appeared infrequently (#47 and #28, respectively) and throughout a small amount of the tests (29.3% and 0.12%, respectively). Since Gilsdorf and Leonard’s results closely mirrored their 1990 study, conclusions related to this area cannot be made at this point.
The results of this research by no means represent a final word on any question involving the formal error in technical communication. Within the existing dataset, further examination to reconcile some of the differences between the frequencies and dispersion lists is needed. An index that factors the weights of both variables would offer a clearer indication of which errors technical communicators value as well as which errors applicants can expect to see in most editing tests.
Finally, additional tests will strengthen the current sample and demonstrate a more common distribution of errors between tests. However, perhaps the privatization of editing tests has produced this disparity and more research on the conventions common to this assessment genre is needed.
My interest in exploring these issues is ongoing. If you would like to contribute an editing test to my dataset, please email me at ryan.boettger@unt.edu. My research will not disclose company names or any sections of the tests themselves—my results will just present an anonymous comparison of the types of errors. I will also sign a nondisclosure agreement to further protect the integrity of your company’s test.
REFERENCES
- Beason, L. (2001). Ethos and error: How business people react to errors. College Composition and Communication, 53(1), 33-64.
- Connors, R. J., & Lunsford, A. A. (1988). Frequency of formal errors in current college writing, or Ma and Pa Kettle do research. College Composition and Communication 34(4), 395-409.
- Eaton, A. (2010). Conducting research in technical editing. In A. J. Murphy (Ed.), New Perspectives on Technical Editing (pp. 7-28). Amityville, NY: Baywood.
- Eaton, A., Brewer, P. E., Craft Portewig, T., & Davidson, C. R. (2008a). Comparing cultural perceptions of editing from the author’s point of view. Technical Communication, 55(2), 140-166.
- Eaton, A., Brewer, P. E., Craft Portewig, T., & Davidson, C. R. (2008b). Examining editing in the workplace from the author’s point of view: Results of an online survey. Technical Communication 54(2), 111-139.
- Gilsdorf, J. W., & Leonard, D. J. (2001). Big stuff, little stuff: A decennial measurement of executives’ and academics’ reactions to questionable usage errors. Journal of Business Communication, 38(4), 439-475.
- Hairston, M. (1987). Not all errors are created equal: Nonacademic readers in the professions respond to lapses in usage error. College English, 43(8), 794-806.
- Hart, G. J. S. (2003). Editing tests for writers. Intercom, 12-15.
- Leonard, D. J., & Gilsdorf, J. W. (1990). Language in change: Academics’ and executives’ perceptions of usage errors. Journal of Business Communication, 27(2), 137-158.
- Lunsford, A. A., & Lunsford, K. J. (2008). “Mistakes are a fact of life”: A national comparative study. College Composition and Communication, 59(4), 781-806.