Early Literacy Research Library (ELRL) - Article

An English-Language Clinic-Based Literacy Program is Effective for a Multilingual Population

Silverstein, M., Iverson, L., Lozano, P. (2002) An English-Language Clinic-Based Literacy Program is Effective for a Multilingual Population. Pediatrics, 109(5), 1-6.,

Access: FREE/Open Access

Publication year


study description

Before/after comparative study of 2 cross-sectional groups.

core topic(s)

Reach Out and Read (ROR)

Population Characteristics

Lingually Diverse , Toddler/Preschool , Urban

Exposures, Outcomes, Other

Child Behaviors and Skills , Home Language/Literacy/Learning Environment , Home Routines , Parent Behaviors and Skills , Play


To assess the effectiveness of a clinic-based pediatric literacy intervention on a multilingual population.


Reach Out and Read (ROR)

outcomes evaluated

Reading and home literacy behaviors.


Pediatric clinic in an urban county hospital.


Reading practices of 2 cross-sectional groups were assessed by standardized interview before and after the intervention. The intervention consisted of waiting-room volunteers reading to children, literacy counseling, and gift of a children's book at each well-child visit from 6 months to 5 years. Outcomes were assessed separately for primary English-speaking and primary non-English-speaking families.

sample size

n=85 (baseline); n=95 (post intervention)


Measure of Reading and Literacy Behaviors: 30 item survey with demographic and literacy-related questions about play, home activities, bedtime rituals, and television viewing to measure the following outcomes:

    • Coded responses for presence or absence of reading as one of the child’s favorite activities, or as one of the caregiver’s favorite activities to do with the child.
    • Number of days per week that the caregiver reported reading to the child.
    • Number of days per week that reading was incorporated into the child’s bedtime routine.
    • Number of children’s books in the home.
    • Ownership and use of a public library card.


The baseline (N=85) and postintervention (N=95) groups were similar with respect to child age and sex, parental education, and length of time in the United States. Fourteen languages were represented in total, the most common being English (41%), Somali (28%), Spanish (9%), Vietnamese (7%), Oromo (3%), and Tigrinyan (3%). Compared with baseline, postintervention respondents were more likely to report reading as a favorite activity for the child (10% vs 25%) and parent (18% vs 40%), to read to their child before bed at least weekly (45% vs 71%), and to possess over 10 children's books at home (49% vs 63%). Among English-speaking families (N=30 baseline, N=40 postintervention), weekly bedtime reading increased (63% to 93%), reading as child's favorite activity increased (7% vs 30%), and reading as the parent's favorite activity to do with child increased (33% vs 58%). The proportion of English-speaking families possessing over 10 books at home and those reading with their children at least weekly showed no difference between the baseline and postintervention groups. Among non-English-speaking families (N=55 baseline, N=55 postintervention), weekly bedtime reading increased (36% vs 56%), reading as the parent's favorite activity increased (11% vs 27%), and the number of families to possess >10 children's books in the home increased (31% vs 49%). Reading as child's favorite activity (13% vs 24%) and weekly book sharing (60% vs 76%) showed nonsignificant trends between the non-English-speaking baseline and postintervention groups.


This clinic-based literacy intervention influences home literacy behavior in this multiethnic setting, in both English-speaking and non-English-speaking families. Although efforts should be made to make such programs more appropriate for linguistic minorities, non-English-speaking families do stand to benefit from English-language-oriented programs.


First, our comparative cross-sectional design is prone to inaccuracy in that it compares 2 distinct populations separated in time by an intervention, and therefore does not definitively establish causality. Second, because all study endpoints depend on parental report, our results—particularly the postintervention data—are subject to both recall and social desirability biases. This would presumably bias our results in a direction that would exaggerate the effects of the intervention. In addition, although the survey was standardized, variables uniformly coded, and questions asked in a uniform manner in both the baseline and postintervention waves of data collection, a different investigator performed each collection. Bias could, therefore, have been introduced in this way. Similarly, the use of convenience samples also could have introduced a sampling bias.