Understanding how oral language skills affect the “Black-White Achievement Gap”

Many African American students are part of a “Black-White Achievement Gap,” in which they fail to perform academically at the levels of their White peers. For the past approximately 15 years, we have worked to improve understanding of the ways in which differences in oral language skills impact academic achievement of African American students.

Summary of Major Findings from the African American English (AAE) Research Project

• Child AAE reflects the operation of more than 30 features, which are different than the way the same words would be said by a speaker of Standard American English (SAE). For example, subject and verb agree in number for SAE: “ he doesn’t need to stand up,” but can vary in AAE: “he don’t need to stand up.”
  (Craig, Thompson, Washington, & Potter, 2003; Washington & Craig, 1994, 2002)

• Through at least 5th grade, AAE is used by many African American students living in large urban centers, with greater feature use by students in an urban-fringe community compared to a mid-size central city.
  (Craig & Washington, 2002, 2004)

• Feature production rates are highly variable. Kindergartners can range from as many as one feature per four words to only one feature per 91 words. First through fifth graders can range from one feature per five words to zero feature production.
  (Craig & Washington, 2004; Washington, Craig, & Kushmaul, 1998)

• At school entry, boys and children from low socioeconomic status homes produce more features than girls, or students from middle socioeconomic status homes, but these differences disappear in later grades in school contexts.
  (Craig & Washington, 2002, 2004; Thompson, Craig, & Washington, 2004; Washington & Craig, 1994, 1998; Washington, Craig, & Kushmaul, 1998)

• The context in which a child is speaking can influence AAE feature production rates. For example, children often produce more features in narrative tasks like picture descriptions than during spontaneous free play interactions.
  (Washington, Craig, & Kushmaul, 1998)

• Many students produce AAE features when reading SAE text aloud. Further, many students produce AAE features in their writing.
  (Thompson, Craig, & Washington, 2004)

• A dialect density measure (DDM) is a sensitive metric of rates of feature production. DDM varies systematically with grade, gender, SES, community, and discourse type.
  (Craig, Thompson, Washington, & Potter, 2003; Craig, Washington, & Thompson-Porter, 1998; Washington & Craig, 1998; Washington, Craig, & Kushmaul, 1998)

• For many (approximately 2/3) but not all students, feature production rates shift dramatically downward at 1st grade, for classroom tasks which involve spontaneous speech, such as describing pictures.
  (Craig & Washington, 2004)

• Another decrease in feature production rates occurs at 3rd grade, for reading.
  (Craig, Thompson, Washington, & Potter, 2003)

• Decreases in feature production rates between kindergarten and 5th grade are associated with higher reading achievement scores.
  (Craig & Washington, 2004, 2006)