Word Frequency Analysis As A Means To Improve Writing Quality

In the old times of Windows 1.0 back in the 1980’s there was a tool called Word Frequency that came with the MS Word distribution package. As someone who uses English as a second language I used it heavily, because it helped me to improve my vocabulary and to correct misspellings beyond the capacity of the available spelling checkers.

That MS Word add-on created a list of all the words in a document, ordered by frequency. It made it easy to detect overuse and/or abuse of a certain word or expression. The little used words were also of help, because sometimes I wrote Thomson instead of Thompson, car instead of cart, or similar errors that the spelling checker does not detect.

Frequency analysis can also be used as a means to establish the “signature” of a certain author, the cultural level of the writer, its use of slang or technical jargon, and other writing features. It is possible to extrapolate the number of words used in a certain text to the total vocabulary of a person. Frequency analysis can accuse some writers to have the vocabulary of a 10-year-old. Or the word-richness of a Chinese-born 2nd year English student.

Frequency analysis combined with a synonym dictionary, as provided in currently available “synonymizer” software, can help writers to enrich its lexicon and avoid abuse of certain expressions.

It is also a means to avoid producing identical text for those who need to make its text different from a source. For instance, a web content writer that needs to fill many similar but not identical pages, and students who want to avoid plagiarism detection and accusation. Rightly or wrongly.

Plagiarism detection also makes use of frequency analysis, because comparison of a given text with the whole Web contents is a major task, and the detection system does not know where to look and where to start. Thus, analysing the word frequency can give some clue on the writing style and the authorship of a given text, without indexing the whole thing.

Search engines use word frequency to establish the subject of web pages. They developed complex linguistic analysis in order to classify pages by subject without human intervention. In turn, webmasters do the same, to try to fool search engines into assigning high keyword relevance to the pages they create. For instance, using a word with a 3% frequency gives a text good relevance on that word (or keyword, in a search engine context). A 10% frequency is still OK, but it is close to “keyword stuffing”, a technique used by webmasters who try to force their websites into the top places of the search engines. Keyword stuffing is penalized by the search engines, and needs to be prevented by smart use of synonyms. Either with synonymizer software or good writing skills.

This article, for instance, has the following Word Frequency :

word : 9, frequency : 7, used : 6, not : 6, search : 6, text: 6, engines: 6, analysis: 5, can: 5, use: 5 ......

I could have edited the text after the analysis, to avoid intensive use of “word” and “frequency” for linguistic purposes. However, it is OK for Search Engine Optimization purposes (attempting to make this article more findable by Google and Yahoo).

Are there any serious writers that still avoid the use of a wired computer? Probably not many can avoid using the Web and the search engines to find the correct word, the most used expression, to perform spelling or grammar checking. Checking word usage in Google is faster and more efficient than using a dictionary, either in paper, disc or the Web. The search engines list every word ever written, not only the well-written words as dictionaries do.

Be prepared to have your texts analysed for word frequency, educational level, plagiarism, technicality, jargon usage and other parameters, in addition to old-fashioned spelling.

According to these tendencies, the ultimate challenge for a job candidate would be to write an essay with paper and pen. Most of us are not prepared to pass such a test.

I expect not to see synonymized versions of this article...