Readability for Developing and Pretesting Concepts, Messages, Materials, and Activities
Readability levels. Readability is an important measurement tool, but it is only a rough indicator of some aspects of comprehension. Therefore, it should not be mistakenly considered the prime indicator of easy-to-use information. There is no single best readability level, given the diversity of American’s health literacy skills. Readability should not exceed 7th to 8th grade (average), the lower end of the estimated average reading level of the U.S. population. In addition, because many people read below that level, a 4th to 6th grade level (easy) is a better goal for information that is critical for people to understand (such as medication instructions) or that is delivered to groups with lower health literacy (such as older adults). Most health risk information can be written well at about a 6th grade level without sacrificing content or style. Most current health communications are written above the 9th grade (difficult) level. For more information, refer to helpful readability discussion in the Centers for Medicaid and Medicare Toolkit.
Readability testing. There are more than 40 readability tests, and their reliability at various grade-reading levels differs, as do their underlying readability formulas. The following validated tools are widely used.
- SMOG (the Simplified Measure of Gobbledygook test)
- The Fry Readability Test
- The Flesch Reading Ease Test
- The Lexile® Framework for Reading
- The Flesch-Kincaid Readability Test is incorporated into Microsoft Word’s (Microsoft, Redmond, WA) readability software and is easily accessible. However, the formula is truncated at a 12th-grade level in Microsoft Word and frequently presents falsely low evaluations.
Because the readability tests vary in their reliability at various grade levels (especially below 6th grade and above 12th grade) and because draft health communication materials often contain text at many readability levels, it is recommended to assess materials using multiple tests (excluding the Flesch-Kinkaid test). Content should be sampled, prepared, and tested according to the test instructions.