Website Spelling Checker

What does it do?

This test checks for spelling errors.

Why is it important?

Poor spelling leaves a poor impression of your website, it can also make your content harder to understand.

How is it measured?

The words on each page are spell checked individually. Where possible, WebScan attempts to guess the language(s) used and selects an appropriate dictionary where possible (for example, American English / British English).

The score is based on the number of pages which have any spelling errors on them, as a proportion (e.g. no pages contain errors = 10 out of 10). If the site cannot be tested for spelling (e.g. the language is not supported) then no score will be awarded, and this will be reported. The test ignores words which have been previously excluded.

Technical explanation

The entire plain text for the page is tested, which includes meta keywords, alternative text, the drop down options in forms and any other potentially human readable text. The dictionary makes reasonable attempts to eliminate words which are casually acceptable. For instance, words written entirely in capitals are assumed to be acronyms, and are not tested (e.g. ABCD).

Any words you add to your dictionary are excluded for all of your websites. If for example, you added 'iPod' to your dictionary for one website, testing another would automatically exclude that word.

To determine language, WebScan looks for:

  • The Content-Language (in a HTTP header, or meta tag), if specified
  • Any lang or xml:lang attributes, if specified
  • If a high percentage of words appear unrecognised (over 10%, and at least 3 words)
  • To determine whether American, British, Australian or Canadian English should be used, the domain name is considered (e.g. .ca = Canadian).

Potential problems

  • Proper names and places. The spell check is case sensitive for some words, e.g. “david” must be spelt “David”.
  • My language is not supported. We're adding new language support over time. Only English is supported at present.
  • My language is not detected correctly. The webpage may not be specifying what language(s) it is using correctly, and WebScan may not have detected this automatically. See Testing non-English websites.
  • Misspelt text does not appear in the webpage. WebScan tests inside areas of the webpage that may not be obvious, for example inside any drop down lists, or meta tags. In some rare cases, the website may be writing out text in a way that appears legitimate onscreen, but to agents like WebScan or Google, appears as something else.

How to improve this score

Exclude any words WebScan does not recognise that are correctly spelt, and correct any remaining spelling mistakes in the website.

How to use this test effectively

This test nearly always needs to be run twice. After testing the first time, someone should review the list of errors reported and click Exclude next to each one which is not a spelling mistake. It is almost guaranteed that the website makes use of one or two words which are not real spelling errors (typically the name of the website, and other brand terms). Running the test again will update the score accordingly. This test can be immensely useful if the dictionary is 'taught' in this manner.

This test can also export spelling mistakes into a to-do list, which is an excellent way of working through the list of problems and correcting them in one go. Simply click the Export link next to the list of spelling errors.

Impact on Summary Scores

Summary Score Percentage Penalty
Content Score 6.5% None
Overall Score 0.8% None