The Cloze Test

Go Back
A Cloze test is a way to measure how difficult a text is to read. Unlike the readability formulas included on this website which measure reading difficulty using only the text itself, a Cloze test requires that someone actually reads the text.
A Cloze test is performed by removing words from a text and replacing them with blank spaces. A reader who is not familiar with the text then tries to fill in the missing words. For example, here is a text with the words removed:
Lexical density is defined as _______________ [1] number of lexical words divided ______________ [2] the total number of words.

___________________ [3] words are the kind which ________________ [4] more information or meaning. More _____________________, [5] lexical words are simply nouns, ______________________, [6] verbs, and adverbs. Nouns tell ______________ [7] the subject, adjectives tell us ________________ [8] about the subject, verbs tell ______________ [9] what they do, and adverbs ________________ [10] us how they do it.
the [1]
by [2]
Lexical [3]
give [4]
precisely [5]
adjectives [6]
us [7]
more [8]
us [9]
tell [10]
The percentage of answers the reader gets correct is the measure of how readable the text is. The higher the percentage, the more readable the text.

Complexity vs. Comprehensibility

The readability formulas used on this website (Gunning fog, Flesch-Kincaid, SMOG, Coleman-Liau, Automated, Fry, and Raygor) are determined only by the text itself without anyone having to read it. This is because these formulas measure only the complexity of a text which can be calculated by considering such things as word lengths, sentence lengths, syllable, counts, and so forth.
On the other hand, a Cloze test measures the comprehensibility of a text. That is, readability is defined in terms of how understandable, or comprehensible, a text is to an actual reader. Thus, this measure can change according to who reads the text.
To understand the difference between complexity and comprehensibility, let us consider an example. Consider the following two sentences:

1) He wore a suit.
2) He filed a suit.

In terms of text complexity, both sentences are very simple. And not surprisingly, their readability scores indicate they require only the most basic reading level in order to read them.
If we shift the focus to comprehensibility, however, the story changes. Certainly the first sentence is comprehensible to a young reader, but the second sentence requires a more mature reader in order to comprehend the full meaning of the text.
The Cloze test is a way to measure the comprehensibility of a text and, as such, measures a much different aspect of readability.

Target Audience

Aside from having a reader actually read the text, the Cloze test requires that we have a target audience in mind from the outset. This is because the result of a Cloze test will not only vary from person to person, but also from one group to the next. For example, a Cloze test for a newspaper article will see much lower scores in a school-age population than an adult population. Thus, we must know our audience in advance.

Performing a Cloze Test

The methodology of a Cloze test depends on the purpose. Some Cloze tests are created by removing certain kinds of words (for example, only nouns). Other tests are created by removing words chosen by the person conducting the test. Moreover, some Cloze tests are created for assessing the difficulty level of a text while others are created with the purpose of student assessment in mind.
As this website is most concerned with giving its users tools to analyze their writing, the most common way to determine the readability of a text for a target audience is to remove every 5th word or every 6th word.
If more than 60% of the blanks are filled in correctly by your sample of test takers from your target audience, then the text is considered to be easy reading for your target audience. If 40% to 60% of the blanks are filled in correctly, then the text is considered "instructional" reading for your target audience. In other words, the text is right for an audience which is learning the material contained in the text. If less than 40% of the blanks are filled in correctly, then the text is considered difficult for your target audience.
The table below summarizes the above paragraph.
Percent Correct Readability
60% or aboveEasy Reading
40%-60%Instructional Reading
40% or belowDifficult Reading
For a Cloze test to determine readability, we emphasize that the test should not be timed. Furthermore, an incorrectly spelled response still counts as a correct reponse for the cloze test. Test takers should be also encouraged to fill in every blank space even if that means guessing.
A Cloze test can be made easier by removing fewer words (by removing every 10th word, for example). Conversely, removing more words makes the test more difficult (by removing every 3rd word, for example). A common standard for the above methodology to apply is to remove every 5th word or every 6th word.

Generating a Cloze Test

The reader may use this website to generate their own Cloze test on our homepage. Our Cloze test generator allows you either to remove every nth word, to remove words of your choice, or to randomly remove a fixed percentage of words.
Go Back

Links and References

DuBay, W. H. 2006. Smart language: Readers, Readability, and the Grading of Text. Costa Mesa:Impact Information.
Fanguy R., Kleen B., Soule L. (2004). Privacy policies: cloze test reveals readability concerns. Issues in Information Systems, iacis.org
Hunt, Irene (2010) The Cloze Procedure, Across Five Aprils (New York: Berkley JAM, 2002)
Nielsen J. (2011), Cloze Test for Reading Comprehension, from The Nielsen Norman Group, Evidence-Based User Experience Research, Training, and Consulting
Taylor, W. L. (1953). Cloze procedure: A new tool for measuring readability. Journalism Quarterly 30: 415–433.
© Holt.Blue