Patterns of Word Recognition Errors Among Adult Basic Education Native and Nonnative Speakers of English

R. Davidson
J. Strucker
Author(s) Organizational Affiliation
National Center for the Study of Adult Learning and Literacy (NCSALL)
Publication Year
Resource Type
Number of Pages

This two-page research brief addresses teaching higher level readers and how to support continued progress to higher levels of reading; their primary focus was on readers’ error patterns.  The authors carried out a secondary analysis of descriptive data from the Adult Reading Components Study (ARCS).  They investigated word analysis and word recognition of both native speakers of English (NSE) and nonnative speakers of English (NNSE) who scored between GE 4 and GE 6. Four research questions guided the study:

  • Would the pattern of relatively stronger print vs. meaning skills in NNSE emerge in the GE 4-6 data?
  • Would the pattern of relatively stronger print vs. meaning skills in NSE emerge in the GE 4-6 data?
  • When matched for word recognition and pseudoword decoding, would the patterns of word recognition errors made by NSE differ from those made by NNSE in the GE 4-6 data?
  • Do the patterns of word recognition errors of NNSE differ depending on whether their exposure to English took place before or after age 12?

Error analysis on the highest level of the DAR Word recognition list resulted in four classifications of errors: correct, phonetically plausible substitutions, phonetically implausible substitutions, and substitutions of real English words (e.g. collaborate for celebrate).


  • Highly similar scores on decoding tests do not necessarily mean people use similar decoding skills.
  • Error patterns of NSE and NNSE differ.
  • Error patterns among NNSE differ depending on whether exposure to English occurred before or after age 12.
  • NNSE more closely resemble normally developing younger readers while NSE more closely resemble children with learning disabilities.

This short paper provides implications for practitioners:

  • Offer different approaches to accommodate the needs of different kinds of intermediate readers.
  • Teachers need to know the phonics principles adult learners appear to have mastered AND those they use with automaticity when reading.

Implications for researchers include:

  • Future studies should analyze both phonetically regular and nonphonetic words.
  • A more fine-grained error classification system could be beneficial.
  • Explore whether the word recognition error patterns also emerge in oral reading of connected text.

NOTE:  A complementary resource from the Workforce Competitiveness Collection (English Language Acquisition) is How should adult ESL reading instruction differ from ABE reading instruction?

What the experts say

Summary of Reading Review Comments: 

This Research Brief could be useful as part of a professional development seminar. During the seminar, participants could be directed to the box of key findings and implications for practitioners. The leaders of the seminar could summarize the Brief (participants should be encouraged to read it after the seminar), and a discussion can ensue on how practitioners can work with non-native speakers of English and native speakers of English learners to increase their decoding abilities.

Summary of English Language Acquisition Review Comments: 

This brief report will be of interest to practitioners, especially those without a background in teaching non-native speakers.  A more rigorous study, of course, is the longer original. This resource is valuable to the field of adult education for the following reasons:

  • The succinct synopsis of a larger study makes it an easily accessible resource for teachers, highlighting findings and implications that teachers would find helpful.  The original article is listed and accessible online for those wishing to read the details of the study or the research background.
  • This resource models the value of diagnostic assessments in understanding reading challenges for adult learners and what these diagnostics can tell us about needed reading instruction.  This is timely, given the increasing emphasis on evidence-based reading practices (such as Student Achievement in Reading or STAR).
  • There is very little research that so clearly compares native and non-native English speakers who read at the same level, yet have different ways of processing and very different instructional needs.  Teachers with little background in teaching language, yet who have non-native speakers in their classrooms, will find this enlightening.
Methods the resource used to collect and analyze the data for the research: Secondary analyses of the descriptive data from the ARCS study. Included 212 of the 676 learners who scored between grade equivalency 4 and 6. Researchers conducted error analysis on the last and highest level GE of the DAR Word Recognition lists.