Heritage Language Retention: A Quantitative Study

This study examines and presents environmental and personal factors that contribute to a successful retention of a heritage language (HL).

The study is based on a newly-designed questionnaire which included questions on selfrated oral, reading, and writing proficiency, as well as HL input, formal language learning in HL, visiting the heritage country, membership in heritage culture related organization(s), voluntary HL use, and exposure to HL culture.

226 HL speakers of five languages (Armenian, Arabic, Korean, Vietnamese, and Tagalog) participated in the study. Participants were divided into three groups depending on age of arrival (AOA): Group 1: born in the U.S. (n=106); Group 2: early arrivers (before 10 yrs old, n=65, mean AOA: 5); Group 3: late arrivers (10 yrs and older, n=55, mean AOA: 14).

Statistical analysis was conducted with self-rated proficiency for each skill as a dependent variable. Independent variables were: formal HL learning, membership in heritage culture related organization(s), and visits to the heritage country.

Formal instruction in HL showed a significant effect on all the language skills for Group 1, only on writing skill for Group 2, and on none for Group 3. Membership in heritage culture related organization(s) and visiting heritage country were found to be associated with significantly higher speaking proficiency for Group 1 and with higher reading and writing
proficiencies of Group 2.

Simple and multiple correlation statistical analyses were conducted with variables of age,
age of arrival, total input in HL (received from parents, siblings, etc.), voluntary use of HL, voluntary exposure to HL culture (reading novels, watching movies, listening to pop songs in HL), and self-rated proficiency for speaking, reading and writing.

For Group 1, speaking proficiency was best predicted by voluntary use of HL, and reading and writing proficiencies by voluntary exposure to heritage culture. For Group 2, speaking proficiency was best predicted by total input, and reading and writing proficiencies by voluntary exposure to heritage culture. For Group 3, oral and writing proficiencies were best predicted by voluntary use of HL and reading proficiency by voluntary exposure to heritage culture.

The multivariate analysis presented here supports the division of HL speakers into separate groups. Individual variables can be used as strong predictors of HL proficiency, which allows HL instructors to apply them in testing, placement, and tracking of HL speakers.