Writing and editing tests

A writing test lets an employer assess an applicant’s response to a common brief, deadline and audience. An editing test examines the ability to identify and correct errors, including the style conventions expected in medical communication.

Purpose and applicants’ experience

In two polls of AMWA members reported in 2008, about 60% had been offered or taken writing tests, commonly an abstract or review based on supplied material; about 45% considered them reasonable, two-thirds were frustrated by absent feedback and a quarter suspected they supplied free work to employers. ¹ Employers used tests to assess thinking on one’s feet, meeting deadlines and using skills creatively, after encountering unrepresentative samples, overpromising, missed deadlines or inability to self-edit. ¹

Four recommendations for writers seeking to avoid misleading an employer are: ²

1. Assess abilities and preferences.
2. Improve skills through courses, such as AMWA’s.
3. Prompt the employer to assess its own needs before requiring a test.
4. Turn down work that does not fit or requires a test.

²

Formats and assessment criteria

Tests vary from a two-hour exercise emailed to the applicant to two-day or week-long exercises, and some must be completed at the employer’s premises; the test language follows the job’s language. ^{3, 4} One employer gives every applicant an identical three-hour test using clinical study report tables: a few sentences describing each table, followed by a conclusion interpreting the findings for the product. ^{5, 6} This permits comparison of English, medical interpretation and writing style. ⁶

That employer considers unpaid tests acceptable because her test is not client work, while acknowledging that a two- or three-day full-time exercise raises the question of whether it will be used for a client. ^{5, 6, 7} An experienced regulatory writer instead regards unpaid testing of experienced applicants as inappropriate, preferring samples, references and a knowledgeable hiring manager’s assessment; the writer took a ninety-minute employment test but considered its complex statistics irrelevant to the writing work. ⁸

Medical communications agencies usually test scientific analysis and aptitude for communicating unfamiliar scientific information to a specified audience, alongside the ability to follow a brief and use language well. ^{9, 10} A 2019 agency account described marking for accuracy, readability, inclusion of requested information and suitability for the objectives and audience. ¹¹ Tasks can include a manuscript introduction based on an abstract, an abstract based on a manuscript, or slides, allowing demonstration of publication structure and often requiring background research as well as accurate use of supplied material. ¹¹ Short on-site error-reading tests may ask for grammatical, spelling and consistency corrections, and experienced applicants may have opportunities to demonstrate project and publication development knowledge. ¹¹

Working from a brief

A medical communications adviser describes initial test briefs as often designed for one to six hours, with wide variation, and recommends allowing a few extra hours and honestly reporting the actual time at interview. ¹⁰ Her three-step approach is: ¹⁰

1. Read the brief and list the items to cover.
2. Match information from the supplied background materials to each point.
3. Read other articles for the intended audience and pitch the article similarly.

¹⁰

Her first test required a review of about fifteen abstracts on hypnotics for insomnia and took twice the specified time; despite spelling and punctuation errors that would normally have meant failure, organising the findings by subcategory with an introduction, discussion, conclusion and title secured a job with further training. ¹⁰ Nearly two years later, a second test required a 1,200-word article for nurses about parenteral antibiotics at home; she read nursing magazines and journals and matched their style using the supplied material. ¹⁰

She advises careful checking and warns against asking an industry friend to proofread a test because a gap between the test and subsequent work may become apparent; in her own second-test case, however, a writer friend proofread the article and found little to change. ¹⁰ Feedback is likely to be discussed at interview, but an agency is unlikely to comment after rejecting an applicant before interview. ¹⁰

Errors in medical editing tests

Editing tests deliberately contain errors and assess spotting typographic mistakes and correcting errors without introducing new ones. ¹²

A 2012 study of thirteen companies’ editing tests found 864 errors of sixty types, averaging 66.46 errors (SD 37.48) and 23.38 types (median 26.0, SD 9.90) per test. ^{13, 14} Ten tests specified AMA style, two APA and one Chicago. ¹³ Errors were grouped into grammar and mechanics, punctuation, spelling, style, content and design. ¹⁵

The six-category comparison used an expected frequency of 144 in each category under an even-distribution hypothesis. ¹⁶ The category counts, binomial p-values and 95% confidence intervals for their shares were: ¹⁷

Category	Errors	Binomial p	95% CI
Style	281	<.001	.294–.358
Grammar and mechanics	234	<.001	.241–.302
Content	136	.494	.134–.183
Punctuation	115	.008	.111–.158
Spelling	57	<.001	.050–.085
Design	41	<.001	.034–.064

¹⁷

Style and grammar/mechanics appeared more often than expected, content about as often, and punctuation, spelling and design less often. ¹⁸ The results suggest that these hiring managers placed particular emphasis on grammar, mechanics and style. ¹⁹

The ranked error list combined frequency (the error count divided by 864) and dispersion (the number of tests containing it divided by thirteen), giving each equal weight by averaging them. ^{16, 17} Twenty-one error types appeared in at least seven tests, and more than 60% of these related to grammar/mechanics or style. ¹⁷

Rank	Error	Category	Frequency index	Dispersion index	Weighted index
1	Unnecessary or missing capitalisation	Grammar	.063	.923	.493
2	Redundant, expendable or incomparable language	Style	.049	.846	.447
3	Vague or missing language	Content	.035	.846	.440
4	Missing or wrong article	Grammar	.028	.769	.398
4	Unnecessary shift in verb tense	Grammar	.028	.769	.398
5	Misplaced or dangling modifier	Grammar	.016	.769	.393
6	Misspelling	Spelling	.066	.692	.379
7	Incorrect number format	Style	.051	.692	.372
8	Incorrect singular/plural application	Grammar	.035	.692	.364
9	Hyphen, en-dash or em-dash error	Punctuation	.034	.692	.363
10	Faulty parallel structure	Style	.025	.692	.359
11	Missing comma with a nonrestrictive element	Punctuation	.020	.692	.356
12	Inconsistent terminology	Style	.017	.692	.355
13	Wrong word	Content	.036	.615	.326
14	Lack of subject–verb agreement	Grammar	.028	.615	.321
15	Equation error, such as an incorrect calculation or symbol	Content	.027	.615	.321
16	Incorrect or missing preposition	Grammar	.024	.615	.320
17	Informal or discriminatory language	Style	.049	.538	.293
18	Unnecessary passive construction	Style	.015	.538	.277
19	Space missing or needed	Design	.014	.538	.276
20	Missing period	Punctuation	.012	.538	.275

¹⁷

Capitalisation was second by frequency but first by the weighted index because it appeared in 92% of tests; misspelling was most frequent but ranked sixth because it appeared in 69%. ^{17, 20} Of spelling errors, 62% could not be detected by a spellchecker, including compounds, homonyms and proper nouns; 26% were general misspellings and 12% British or alternative spellings. ²⁰ This suggests an emphasis on attention to detail, although the study could not establish whether the lower ranking of wrong words than misspellings was anomalous or reflected differences between student writers and professional assessors. ²⁰

Twelve of the thirteen tests contained passages about people’s safety or treatment, including patients, research participants and healthcare workers, making respectful language relevant alongside consistency and concision. ¹⁹ The study suggests that medical editing tests may be especially intensive in style, but its small sample limits the conclusion; preparation needs to consider context as well as recognise error types. ^{12, 21}