Computer-Assisted Vocabulary Intervention for Spanish-English Bilingual Children With Developmental Language Disorder (CABVI)

Spanish-English bilingual children with developmental language disorder are randomized to bilingual CABVI. The computer-assisted behavioral vocabulary program teaches 36 concepts in 24 small-group sessions, ordinarily over about six weeks. Each session includes approximately 45 minutes of instruction within a 60-minute school block. Sessions 1 and 3 of each unit are in Spanish; Sessions 2 and 4 are in English. School-calendar disruptions may extend the schedule while preserving the planned sequence.

Behavioral vocabulary intervention using the CABVI app with trained adult oversight in fixed groups of three to five children. Twenty-four planned sessions teach 36 concepts across six units. Each session includes approximately 45 minutes of instruction within a 60-minute school block. Spanish is used in Sessions 1 and 3 of each unit and English in Sessions 2 and 4. Activities include shared reading, receptive and expressive vocabulary practice, songs and movement, and supported conversations. Four sessions ordinarily occur within seven days, for about six weeks total; school-calendar disruptions may extend the schedule while preserving session sequence.

Spanish-English bilingual children with developmental language disorder are randomized to English-only CABVI. This condition uses the same computer-assisted behavioral vocabulary content, 36 concepts, 24-session sequence, small-group format, and planned dose as bilingual CABVI, with all instruction delivered in English. Sessions include approximately 45 minutes of instruction within a 60-minute school block, ordinarily over about six weeks, with school-calendar extensions as needed.

Computer-assisted behavioral vocabulary intervention using the same 36 concepts, instructional content, 24-session sequence, fixed groups of three to five children, and planned dose as bilingual CABVI. All instruction is delivered in English. Sessions include approximately 45 minutes of instruction within a 60-minute school block, ordinarily four sessions within seven days over about six weeks. School-calendar disruptions may extend the schedule while preserving session sequence. Trained adults oversee the shared reading, receptive and expressive vocabulary activities, songs and movement, and supported conversations.

Spanish-English bilingual children with developmental language disorder are randomized to their usual English-language school services. A CABVI-matched research dose is not imposed. Weekly logs document actual service frequency, duration, content, and attendance. Required educational and individualized education program services continue independently of study participation.

Usual English-language school services without a CABVI-matched research dose. Weekly logs document actual service frequency, duration, content, and attendance. Required educational and individualized education program services continue independently of study participation.

Forty Spanish-English bilingual children with typical language development receive bilingual CABVI without randomization, providing a developmental comparison. They receive the same 36-concept, 24-session program and planned dose as the bilingual CABVI group with developmental language disorder. Sessions 1 and 3 of each unit are in Spanish; Sessions 2 and 4 are in English. Delivery is ordinarily completed over about six weeks, with school-calendar extensions as needed.

Behavioral vocabulary intervention using the CABVI app with trained adult oversight in fixed groups of three to five children. Twenty-four planned sessions teach 36 concepts across six units. Each session includes approximately 45 minutes of instruction within a 60-minute school block. Spanish is used in Sessions 1 and 3 of each unit and English in Sessions 2 and 4. Activities include shared reading, receptive and expressive vocabulary practice, songs and movement, and supported conversations. Four sessions ordinarily occur within seven days, for about six weeks total; school-calendar disruptions may extend the schedule while preserving session sequence.

Researcher-developed picture Naming task for 36 directly taught concepts in Spanish. Images differ from intervention pictures and are reused across assessment timepoints. After any clearly other-language-only initial response, regardless of accuracy, give one fixed reminder to answer in Spanish, without translation or answer cues. An accepted requested-language label initially or after that reminder earns 1; incorrect, absent or other-language-only responses earn 0. Mixed responses containing an accepted requested-language label need no reminder. Record initial response, reminder and subsequent response separately; retain other-language conceptual knowledge separately. Complete-inventory total: 0-36; higher scores indicate greater requested-language Naming knowledge. Record actual denominators; refused, unadministered and uninterpretable items remain missing.

Researcher-developed Definition task for 36 directly taught concepts in Spanish, without a picture. Supply the label and allow up to THREE TOTAL fixed neutral prompts: the initial meaning question and up to two neutral rephrasings. Continue only when no clearly accurate requested-language attribute, function or description has emerged and the child remains willing. Once binary credit is clear, stop extra prompting but let the child finish all volunteered information. No answer cues or correctness feedback. Accurate attribute, function or description: 1; incorrect, irrelevant, label-only or absent response after valid administration: 0. Preserve all prompts and responses for scoring and Elaboration exclusions. Complete-inventory total: 0-36; higher scores indicate greater semantic knowledge. Record actual denominators; refused, unadministered and uninterpretable items remain missing.

Researcher-developed picture Naming task for 36 directly taught concepts in English. Images differ from intervention pictures and are reused across assessment timepoints. After any clearly other-language-only initial response, regardless of accuracy, give one fixed reminder to answer in English, without translation or answer cues. An accepted requested-language label initially or after that reminder earns 1; incorrect, absent or other-language-only responses earn 0. Mixed responses containing an accepted requested-language label need no reminder. Record initial response, reminder and subsequent response separately; retain other-language conceptual knowledge separately. Complete-inventory total: 0-36; higher scores indicate greater requested-language Naming knowledge. Record actual denominators; refused, unadministered and uninterpretable items remain missing.

Researcher-developed Definition task for 36 directly taught concepts in English, without a picture. Supply the label and allow up to THREE TOTAL fixed neutral prompts: the initial meaning question and up to two neutral rephrasings. Continue only when no clearly accurate requested-language attribute, function or description has emerged and the child remains willing. Once binary credit is clear, stop extra prompting but let the child finish all volunteered information. No answer cues or correctness feedback. Accurate attribute, function or description: 1; incorrect, irrelevant, label-only or absent response after valid administration: 0. Preserve all prompts and responses for scoring and Elaboration exclusions. Complete-inventory total: 0-36; higher scores indicate greater semantic knowledge. Record actual denominators; refused, unadministered and uninterpretable items remain missing.

Researcher-developed picture-supported Elaboration task for 36 directly taught concepts in Spanish. After the Definition block, show the fixed concept picture and ask for something ELSE about the word, with up to two neutral continuations. Score 0-3 distinct accurate target-relevant ideas not expressed anywhere in that item's preceding Definition exchange, including volunteered details. Exclude every earlier Definition fact, repeated idea, unrelated scene detail and target-label repetition alone. Accurate knowledge beyond the picture may count. Preserve Definition evidence and each newly credited idea; record other-language contributions separately. Images are reused across timepoints and differ from intervention images. Complete-inventory total: 0-108; higher scores indicate more additional information with picture support. Record actual denominators; refused, unadministered and uninterpretable items remain missing.

Researcher-developed picture-supported Elaboration task for 36 directly taught concepts in English. After the Definition block, show the fixed concept picture and ask for something ELSE about the word, with up to two neutral continuations. Score 0-3 distinct accurate target-relevant ideas not expressed anywhere in that item's preceding Definition exchange, including volunteered details. Exclude every earlier Definition fact, repeated idea, unrelated scene detail and target-label repetition alone. Accurate knowledge beyond the picture may count. Preserve Definition evidence and each newly credited idea; record other-language contributions separately. Images are reused across timepoints and differ from intervention images. Complete-inventory total: 0-108; higher scores indicate more additional information with picture support. Record actual denominators; refused, unadministered and uninterpretable items remain missing.

Researcher-developed task in Spanish using six scenes, one per unit, each depicting six target concepts in contexts different from intervention scenes. The SAME scenes are reused at pretest, posttest and follow-up; they are not previously unseen at every assessment. Administer first within each language, before Naming, target-label supply or receptive tasks. Per concept: 2 for spontaneous accurate requested-language target use; 1 for accurate use only after a logged permitted nonlexical continuation; 0 for absent or inaccurate use during valid administration. No target labels, pointing cues or teaching. Preserve initial and prompted responses separately. Complete-inventory total: 0-72 across 36 concepts; higher scores indicate more contextualized target use. Report languages separately, actual denominators and source limitations. Refused, unadministered and uninterpretable items remain missing.

Researcher-developed receptive task in Spanish assessing 36 directly taught concepts. The examiner supplies the target label and presents a fixed four-image array. The first independent choice receives 1 point if correct and 0 if incorrect or no choice; no retries or corrective feedback are given. The complete-inventory raw total ranges from 0 to 36, with higher scores indicating greater receptive recognition. Actual item denominators and reasons for nonadministration are recorded; nonadministered items are missing, not zero.

Exploratory researcher-developed expressive task in Spanish for the 15 selected incidental concepts encountered in instructional stories, elaborations, routines or directions outside the 36 directly taught targets. Fixed image/scene prompts do not supply or define the tested word. An appropriate requested-language selected word, phrase or accepted inflection earns 1; an incorrect or absent selected form after valid administration earns 0. Accurate paraphrases and other-language knowledge are retained separately as supplemental conceptual evidence, without replacing the word-production score. The same inventory, pictures, prompts and response keys are used across timepoints. Complete-inventory total: 0-15; higher scores indicate greater selected-word expressive knowledge. Report actual denominators; refused, unadministered and uninterpretable items remain missing.

Exploratory researcher-developed receptive task in Spanish for the same 15 selected incidental concepts encountered in instructional stories, elaborations, routines or directions outside the 36 directly taught targets. After expressive tasks, supply the target label and show a fixed four-image array. Score the first deliberate independent choice: 1 correct; 0 incorrect or no choice after valid administration. No teaching, corrective feedback or child retries. Correct documented recording/motor errors with an audit trail. Use the same inventory, pictures, positions and response keys across timepoints. Complete-inventory total: 0-15; higher scores indicate greater receptive knowledge. Expressive and receptive scores remain separate. Report actual denominators; refused, unadministered and uninterpretable items remain missing.

Researcher-developed task in English using six scenes, one per unit, each depicting six target concepts in contexts different from intervention scenes. The SAME scenes are reused at pretest, posttest and follow-up; they are not previously unseen at every assessment. Administer first within each language, before Naming, target-label supply or receptive tasks. Per concept: 2 for spontaneous accurate requested-language target use; 1 for accurate use only after a logged permitted nonlexical continuation; 0 for absent or inaccurate use during valid administration. No target labels, pointing cues or teaching. Preserve initial and prompted responses separately. Complete-inventory total: 0-72 across 36 concepts; higher scores indicate more contextualized target use. Report languages separately, actual denominators and source limitations. Refused, unadministered and uninterpretable items remain missing.

Researcher-developed receptive task in English assessing 36 directly taught concepts. The examiner supplies the target label and presents a fixed four-image array. The first independent choice receives 1 point if correct and 0 if incorrect or no choice; no retries or corrective feedback are given. The complete-inventory raw total ranges from 0 to 36, with higher scores indicating greater receptive recognition. Actual item denominators and reasons for nonadministration are recorded; nonadministered items are missing, not zero.

Exploratory researcher-developed expressive task in English for the 15 selected incidental concepts encountered in instructional stories, elaborations, routines or directions outside the 36 directly taught targets. Fixed image/scene prompts do not supply or define the tested word. An appropriate requested-language selected word, phrase or accepted inflection earns 1; an incorrect or absent selected form after valid administration earns 0. Accurate paraphrases and other-language knowledge are retained separately as supplemental conceptual evidence, without replacing the word-production score. The same inventory, pictures, prompts and response keys are used across timepoints. Complete-inventory total: 0-15; higher scores indicate greater selected-word expressive knowledge. Report actual denominators; refused, unadministered and uninterpretable items remain missing.

Exploratory researcher-developed receptive task in English for the same 15 selected incidental concepts encountered in instructional stories, elaborations, routines or directions outside the 36 directly taught targets. After expressive tasks, supply the target label and show a fixed four-image array. Score the first deliberate independent choice: 1 correct; 0 incorrect or no choice after valid administration. No teaching, corrective feedback or child retries. Correct documented recording/motor errors with an audit trail. Use the same inventory, pictures, positions and response keys across timepoints. Complete-inventory total: 0-15; higher scores indicate greater receptive knowledge. Expressive and receptive scores remain separate. Report actual denominators; refused, unadministered and uninterpretable items remain missing.

Observed session-level fidelity on the 10-criterion CABVI Fidelity Checklist, completed by a trained reviewer who did not deliver the session. Criteria are rated Y (delivered as planned), N (not delivered as planned), NO (not observable), or NA (not applicable). Fidelity is 100 x Y/(Y + N), ranging from 0% to 100%; higher percentages indicate greater delivery as planned. NO and NA are excluded from the denominator and reported separately; a zero denominator is not estimable. Observation method and denominator are reported. Independent review covers at least 20% of completed sessions per therapy group, rounded up (at least 5 of 24). Provider self-checks are separate and excluded from independent fidelity scores. No passing cutoff is assumed.

Six researcher-developed items (C1-C6) assess confidence in supporting bilingual vocabulary within professional role; assigned activities when child and clinician languages differ; use of assigned materials and technology within the instructional plan; accurate clinical and research records; responding to breaks or unwillingness; and recognizing and obtaining needed help. Each item is rated separately: 1 Not at all confident, 2 Slightly, 3 Moderately, 4 Very, 5 Extremely confident. Higher ratings indicate greater confidence. The same items are given before and after implementation. For usual-care clinicians, assigned activities mean usual care and study documentation. Not applicable and Declined are recorded separately from numeric ratings. Item-level responses are retained; no summed or mean composite is specified.

Ten researcher-developed items (I1-I10) from the post-implementation clinician questionnaire assess preparation by training, clarity of instructions/materials, fit with the agreed school schedule, manageable documentation time, protection of required IEP services, access to help, child engagement, comfort with privacy/assent procedures, acceptability for children served, and overall satisfaction with study participation. Each item is rated separately: 1 Strongly disagree, 2 Disagree, 3 Neither agree nor disagree, 4 Agree, 5 Strongly agree. Higher ratings indicate greater agreement with the favorable statements. All participating clinicians may respond. Not applicable and Declined are recorded separately from numeric ratings. Item-level responses are retained; no summed or mean composite is specified. CABVI app items A1-A4 are reported separately.

Four researcher-developed items (A1-A4) from the CABVI app block of the same post-implementation clinician questionnaire, completed only by CABVI providers, assess whether app audio models helped support assigned Spanish-language activities; usefulness of clinician controls and captions; manageability of technical interruptions with approved procedures and support; and willingness to consider future use if available and institutionally approved. Each item is rated separately: 1 Strongly disagree, 2 Disagree, 3 Neither agree nor disagree, 4 Agree, 5 Strongly agree. Higher ratings indicate greater agreement. Usual-care clinicians skip this block and are not asked to use or review CABVI. Not applicable and Declined are recorded separately. Item-level responses are retained; no summed or mean composite is specified.

Researcher-developed item A5 asks CABVI providers whether they learned Spanish words or phrases while using CABVI. Responses are Yes, No, Unsure, Not applicable, or Declined and are retained as separate categorical responses; there is no numerical scale or total score. An optional nonidentifying general example may accompany the response. Usual-care clinicians skip this CABVI-only block. This is clinician self-report, not an objective Spanish proficiency test.

Researcher-developed item W1 asks all adult clinician participants to report the approximate extra time per week spent on research activities during implementation, excluding ordinary school therapy. Responses are recorded in minutes per week. Higher numeric values indicate greater reported research time. Unsure and Declined responses are recorded separately from numeric time and are not scored as zero. The form specifies no upper numerical bound. This is a retrospective report of weekly research time, not a sum of ordinary therapy minutes or a daily time log.

Researcher-developed item W2 asks all adult clinician participants how the extra research workload compared with what they expected. Ordered response categories are Much less, Somewhat less, About the same, Somewhat more, and Much more. Declined is recorded separately. Responses are retained as categories rather than an invented numerical total. Categories toward Much more indicate greater workload relative to expectations. Optional nonidentifying comments about the time required are collected separately.

Brief researcher-developed parent and teacher questionnaires assess perceived changes in the child's vocabulary use, communication confidence and engagement. The planned questionnaires use five-point item ratings and open-ended comments, with no CABVI-specific wording to reduce response bias. Ratings are summarized descriptively by item and respondent group; comments are thematically coded. No combined total is specified. This feedback is distinct from baseline language-background questionnaires.

MLUw is calculated from the whole available Spanish-elicited Frog story retell using SALT bilingual modified C-unit segmentation and fixed transcription rules. SALT-counted non-maze words in complete, intelligible verbal units are divided by the number of eligible units. Scores are expressed as words per unit, with no fixed maximum; higher values indicate longer average utterances. Intelligible code switches are marked and retained. Eligible-unit counts, sample length and transcription limitations are recorded. English-elicited and Spanish-elicited results remain separate. When no adequately transcribed eligible units are available, the score is missing with a documented reason, not zero.

NDW counts distinct word roots among non-maze words in complete, intelligible verbal units from the whole available Spanish-elicited Frog story retell, using fixed SALT bilingual transcription and root-coding rules. Intelligible code switches are marked and retained. Higher counts indicate more distinct roots observed; counts depend on sample length and are not length-independent diversity scores. NTW and eligible-unit counts are retained for interpretation. The maximum is the sample's included total-word count. Language results remain separate. An unscorable or absent sample is missing with its reason, not zero.

MLUw is calculated from the whole available English-elicited Frog story retell using SALT bilingual modified C-unit segmentation and fixed transcription rules. SALT-counted non-maze words in complete, intelligible verbal units are divided by the number of eligible units. Scores are expressed as words per unit, with no fixed maximum; higher values indicate longer average utterances. Intelligible code switches are marked and retained. Eligible-unit counts, sample length and transcription limitations are recorded. English-elicited and Spanish-elicited results remain separate. When no adequately transcribed eligible units are available, the score is missing with a documented reason, not zero.

NDW counts distinct word roots among non-maze words in complete, intelligible verbal units from the whole available English-elicited Frog story retell, using fixed SALT bilingual transcription and root-coding rules. Intelligible code switches are marked and retained. Higher counts indicate more distinct roots observed; counts depend on sample length and are not length-independent diversity scores. NTW and eligible-unit counts are retained for interpretation. The maximum is the sample's included total-word count. Language results remain separate. An unscorable or absent sample is missing with its reason, not zero.