MacArthur CDI for American Sign Language

The MacArthur Communicative Development Inventory for American Sign Language (ASL-CDI; Anderson & Reilly, 2002) is a tool for measure early vocabulary development of deaf children acquiring ASL.

The goal for developing the ASL-CDI was to learn more about the normal language development of deaf children, i.e. early lexical and grammatical development of language in deaf children of deaf parents. Since most deaf children of hearing parents do not have early access to ASL, those children are ”clearly at risk for early language delay; the need for early language assessment and normative data is critical” (Anderson & Reilly, 2002, 83). Additionally, one main objective for the development of the ASL-CDI was to provide a larger sample (as compared to the small number of subjects in previous studies on ASL acquisition) for documenting children’s development of ASL.

The ASL-CDI primarily focuses on deaf children between the ages of 8-36 months. It was developed to measure the production of early lexical development of ASL.

The ASL-CDI is based on the  English CDI (Fenson et al., 1993), a parental report form that allows parents to check the first words and early grammatical structures their child produces.

The English CDI is based on a recognition format and exists in two versions (1) for infants from 8-16 months and (2) for toddlers 16-30 months old. It has been adapted into various spoken languages. The ASL-CDI exists only in one format, i.e. for productive vocabulary for children 8-36 months of age.

In the process of developing the ASL-CDI, a number of changes were made in order to meet the needs of cultural and linguistic differences. Based on research studies on early sign language vocabulary development and in consultation with Deaf colleagues, the authors decided to omit inappropriate words for ASL from the item pool, e.g. animal sounds because deaf children do not have that audiological experience. They added, however, signs that reflected Deaf culture and -experience, e.g. TTY. Words that were considered as having equivalent forms in ASL, e.g. horse/pony were modified, only the formal word/sign HORSE was included in the ASL-CDI. Also signs with (almost) identical forms, e.g. EAT-FOOD or SIT-CHAIR and different grammatical function (verb vs. noun) were changed to include only one sign. This process affected five signs. These signs were entered as verbs in the action sign category.

In the English CDI, the parents are also able to check early grammatical structures. Because of the lack of a written form of ASL to represent grammatical aspects, the grammar section was omitted. Instead, questions about language use at home, the ability of fingerspelling, and the longest sign produced by the child were included.

Thirty deaf children of deaf parents from 10-36 months were included in the pilot study. From these 30 families 97% reported that they used ASL at home. After deaf parents had filled out the ASL-CDI, the data was analyzed and certain changes were made to the original ASL-CDI: (1) signs that children used infrequently (10% or less) were deleted, (2) signs were added when at least 10% of the parents spontaneously noted their child using them, and (3) the semantic category for body parts was eliminated since they required only pointing to the particular body part to identify this part. Based on the changes after the pilot study, the ASL-CDI consisted of 537 signs in 20 semantic categories.

Data from 110 ASL-CDIs (69 children cross-sectional, 34 children longitudinal) could be compiled. All 69 deaf children from the cross-sectional study came from deaf families. They were between 8-35 months old.

The researchers analyzed the results with regard to reliability and validity. Test-retest reliability was established with 25% of the normative sample (n=16) by re-administering the ASL-CDI five to seven months later. The results revealed a high test-retest reliability exceeding an r of .91. This suggests that parents were very reliable in reporting their child’s sign language production.

Investigating the external validity ”refers to the extent that parents are actually checking signs that their children produce” (Anderson & Reilly, 2002, 87). In order to assess the external validity, the researchers videotaped 10 children (age range: 17-34 months) on the same day parents filled out the ASL-CDI form. The video-recorded sessions captured both free play as well as structured tasks to elicit language production. Only signs that the parents checked on the ASL-CDI and that the authors found on the video (parents did not report all signs their child produced) were used for the analysis. ”A ratio between these two scores (number of signs produced by the child during videotaping and endorsed by the parents on the ASL-CDI compared to the total number of signs produced by the child that the parents could have either endorsed or not on the ASL-CDI) yielded an external validity score of .87, ranging from .71 to 1.00. This suggests that parents accurately reported the signs that their children produced as measured by the ASL-CDI” (Anderson & Reilly, 2002, 88).

The researchers report that they will keep working on the project: with a bigger sample size, it will be possible at some point to provide normative data on normal vocabulary development in ASL. In addition, it also will allow to use the ASL-CDI as a screening tool for early sign vocabulary development or for monitoring deaf students’ progress and evaluate the results of intervention measures.

The preliminary normative data is available for vocabulary size, first signs, emergence of linguistic categories such as wh-forms, emotion signs, and negation. With this normative data, it is possible to see the developmental level a deaf child is at with respect to others (in his/her early ASL development).

The development of the assessment of early grammatical structures within the ASL-CDI is currently being developed. It will be in a video format and assess deaf children’s skills across a range of grammatical areas (D. Anderson, personal communication, June 2, 2004).

The preliminary findings of the ASL-CDI provide information on the early lexical development in ASL. These findings are presented in the table below.

Table 1: Approximate Age and Vocabulary Ranges for the Emergence of Specific Lexical Items*






Emotion Signs

Cognitive Verbs


below 18 months

150 signs






18-21 months

150-250 signs






general points to others

21-24 months

250-350 signs






30-35 months

>350 signs






*This table was provided by Diane Anderson

Strengths: easy to administer by parents, young age group, reported reliability and validity, provides information about the lexical development in deaf children acquiring ASL as their first language

Weakness: only production, not comprehension

For the BSL version of the CDI, please visit the website of  City University London.


Summarized by Tobias Haug (2004, in cooperation with Diane Anderson).

For further questions regarding the ASL-CDI, please contact  Diane Anderson at the University of California, Berkeley or  Judy Reilly at San Diego State University.