Calculate MLU in a Language Sample
In ChatGPT or your favorite AI chatbot, cut and paste the following prompt to help you analyze the mean length of utterance for your speech therapy language sample. To get started, just replace each bracket with the information for each section.
Calculating MLU Prompt
You are an expert speech language pathologist, highly skilled in analyzing language samples for mean length of utterance to determine how well a student’s language skills are developing. Your task is to analyze the following language sample and calculate the mean length of utterance. Here is the language sample: [INSERT LANGUAGE SAMPLE].
Example Prompt
You are an expert speech language pathologist, highly skilled in analyzing language samples for mean length of utterance to determine how well a student’s language skills are developing. Your task is to analyze the following language sample and calculate the mean length of utterance. Here is the language sample: [INSERT LANGUAGE SAMPLE].
Additional Prompting Strategies
Ensure that the language sample does not include personally identifiable information (PII).
Enlist the chatbot to calculate percentage grammatical utterances (PGU). If your student is putting words together but they are not as long or as complicated as you want, the PGU measurement can help you hone in on grammar goals and make a plan for treatment.
In calculating MLU, type in just the utterance that was stated, not the articulation corrections. For example, if your student says, “dat is wight.” And you type it as “dat/that is wight/right.” The AI chatbot will calculate it as 5 morphemes, not 3.
The chatbot cannot completely replace your knowledge as an SLP. While it can quickly calculate and analyze grammar and syntax, please also review all of your information before submitting your final report or reporting to parents.
Developed in collaboration with Kristen Ponce, M.S., CCC-SLP, Speech-Language Pathologist @aac_to_the_core.
