Co-Principal Investigator: Petscher, Yaacov

Purpose: The purpose of this project was to explore a nationally representative sample of U.S. adults with low reading skills to identify factors that may predict literacy performance. According to the 2017 PIAAC report, roughly 1 in 5 (19 percent) of U.S. adults read at or below basic levels. This group has inherent heterogeneity, but there may also be commonalities that practitioners or policymakers could leveraged to help improve instruction or other interventions. In this study, the researchers explored data from adults who score low in literacy to understand their component reading skills (e.g., vocabulary, comprehension), broad literacy skills, and possible correlations with other malleable factors (e.g., reading, writing, numeracy, and information and communication technology (ICT) behaviors at home and at work), numeracy skills, and background factors (e.g., demographics, education).

Project Activities: The researchers used U.S. data from an international survey of adult skills, the Program for the International Assessment of Adult Competencies (PIAAC). They examined the reliability and validity of the reading components scales (e.g., vocabulary, sentence processing) and dimensions of the literacy scale for adult struggling readers. The researchers also created profiles of U.S. adult struggling readers' engagement in reading, numeracy, writing, and ICT skills-use at home and at work and examine literacy performance and demographic risk factors (e.g., learning disability, age). Finally, the researchers examined predictors of the literacy and numeracy performance of adult struggling readers. This project also provided training opportunities (advanced analyses, working with large educational datasets, and dissemination efforts) for postdoctoral scholars and graduate students.

Structured Abstract

Setting: This study used the nationally representative U.S. sample from the PIAAC dataset.

Sample: The U.S. PIAAC data include nationally representative cohorts (collected in 2012, 2014, and 2017) of adults aged 16 to 74 (total N = 12,330). The research team focused on a subset of adults who scored at or below Levels 1 and 2 (the lowest levels) on the PIAAC literacy domain (N = 7,359 2012/2014/2017 data) and all low-skilled adults who completed a paper-based reading components subsection (N = 1,279 2012/2014 data).

Factors: In this study, the researchers explored whether the component reading skills measured in the PIAAC (i.e., vocabulary knowledge, sentence processing, and passage comprehension subsections) are valid and reliable for low-skilled adults (at or below Level 2 in literacy). They also explored relations among literacy item dimensions, adults' performance on reading skills, literacy scores, and numeracy scores, demographic factors (e.g., gender, age, race/ethnicity, English language proficiency, disability status, and educational background), and other more malleable factors (e.g., readiness to learn, reading, writing, numeracy, and ICT behaviors).

Research Design and Methods: Using secondary data analysis and various psychometric and predictive modeling approaches, the researchers conducted three lines of study. Combined, these studies verified the validity and reliability of the PIAAC literacy and component reading skills for low-skilled adults. In addition, these studies examined subgroups of adult struggling readers and skill and demographic factors that predict performance. In study 1, they assessed the psychometrics of the item-level data from the component skills portion and the paper-based literacy assessment to see how they function for adult struggling readers. In study 2, they used PIAAC reading, writing, numeracy, and ICT skills-use data to create profiles and identify risk factors of adult struggling readers. In study 3, they examined predictors of literacy and possible demographic and reading component skill moderators of these skills. A final, and more exploratory subcomponent of the third study was to explore the relations among literacy, component reading skills, demographic characteristics, and numeracy skills for low-skilled adults.

Control Condition: Due to the exploratory nature of the research design, there was no control condition.

Key Measures: The measures included literacy, reading component skills, and numeracy scores from the 2012, 2014, and/or 2017 U.S. cohorts of PIAAC. The researchers also used self-reported engagement variables (reading, writing, numeracy, and ICT skills-use at home and at work) as well as demographic variables.

Data Analytic Strategy: For study 1, the researchers conducted classic item response modeling, confirmatory factor analyses, bifactor analyses, and explanatory item response modeling. For study 2, they used latent mixture modeling and structural equation modeling. For study 3, they used regression analyses with moderators and structural equation modeling.

Related IES Projects: Using Process Data to Characterize Response Profiles and Test-Taking Behaviors of Low-Skilled Adult Responders on PIAAC Literacy and Numeracy Items (R305A210344)

Products and Publications

ERIC Citations:  Find available citations in ERIC for this award here.

Publicly Available Data: The PIAAC dataset used is publicly available through NCES and OECD. If interested in learning more about specific analyses, interested individuals can contact the research team.

Select Publications:

Inside IES Research Blog

Tighe, E. L. and Larson, M. (2018, September 24). New IES Grantee Focuses on Improving Adult Literacy. Inside IES Research. /blogs/research/post/new-ies-grantee-focuses-on-improving-adult-literacy