Bilingual SLP Tools: PGU + MLU Guide
In a previous post, we walked through how to calculate Percent Grammatical Utterances (PGU) using a narrative language sample. While PGU is a powerful tool in monolingual evaluations, it doesn’t always tell the full story for bilingual children.
Today, we’re taking it a step further. We’ll review both PGU and Mean Length of Utterance in words (MLUw), explain why these measures can be misleading in isolation for bilingual students, and share how to apply them using narrative tasks like the Frog series. Plus, I’ve created a free Excel spreadsheet to help you score both PGU and MLU with ease.
Why PGU and MLU Aren’t Enough on Their Own
PGU measures the proportion of grammatically correct utterances in a sample. MLU captures sentence complexity based on average utterance length. Both are commonly used in identifying language impairments—but research shows that when used in isolation, they can fall short for bilingual learners.
According to Hernandez et al. (2024), bilingual children with Developmental Language Disorder (DLD) often show variability across languages, and error patterns may differ between English and Spanish. For example, children might struggle with tense in English but with articles in Spanish. That makes single-language measures insufficient.
The study explored two solutions:
Best Language Approach: Using the higher PGU and MLUw from the Spanish and English samples.
Total Language Approach: Combining measures across both languages to reflect distributed skills.
Both approaches outperformed monolingual-only metrics with the best results from combining PGU and MLUw under the “best language” model.
How to Use Narrative Language Sampling with Bilingual Students
Narrative tasks are ideal for eliciting connected, complex language in both English and Spanish. That’s why we love using the Frog wordless picture books by Mercer Mayer. They provide enough visual structure to guide students, while still allowing for open-ended storytelling.
Use two different books to gather samples in each language. English and Spanish scripts for each of the 4 books are available in written and audio format on SALT’s website for free.
Step-by-Step: How to Calculate PGU and MLU
We recommend transcribing narrative samples in Word or Excel. We have provided a simple tutorial of how to calculate in Word, but we recently created a free bilingual PGU + MLU spreadsheet makes this process even easier and helps you track data across languages.
Divide the transcript into utterances (C-units).
Mark grammatical vs. ungrammatical utterances.
Skip any that are unintelligible or code-switched.Count the total utterances (minus excluded ones).
Calculate PGU = Grammatical ÷ (Grammatical + Ungrammatical) × 100
Calculate MLUw = Total words ÷ Total utterances
Repeat for both languages.
Interpreting the Data: Compare to Cutoff Scores
Use your child’s best language sample (highest PGU) to calculate PGU and MLUw, and compare against age-based cutoffs from Hernandez et al. (2024). For example:
A 5-year-old’s best language would be from where their PGU was the highest (example above is Spanish with 68.42%). Then using the MLU from the best language (Spanish), we would compare the MLU against the table below. The cut-off PGU would be 34.8 at an MLU of 5.0 (I rounded down since he was below 5.5 MLU in Spanish). This would indicate this child does not fall in the range of impairment.
From: Hernandez, M., Ronderos, J., & Castilla-Earls, A. P. (2024). Diagnostic Accuracy of Grammaticality and Utterance Length in Bilingual Children. Language, Speech, and Hearing Services in Schools, 55(2), 577–597. https://doi.org/10.1044/2024_lshss-23-00100
You can also combine data from both languages using the Total Language Approach, if a child’s skills are more evenly distributed. In this scenario you take the MLUw from your samples to find the PGU cut-off (I have continued to round down 5.27->5.0 Spanish, 5.72>5.5 English) to obtain a PGU cut-off of 31.7. This child’s PGU was above the cut-off for both language indicating they do not fall in the range of impairment.
From: Hernandez, M., Ronderos, J., & Castilla-Earls, A. P. (2024). Diagnostic Accuracy of Grammaticality and Utterance Length in Bilingual Children. Language, Speech, and Hearing Services in Schools, 55(2), 577–597. https://doi.org/10.1044/2024_lshss-23-00100
Final Thoughts
When evaluating bilingual students, PGU and MLU are most powerful when used together—and when evaluated across both languages. These narrative measures help us go beyond surface-level assessments and offer a more accurate, equitable picture of a child’s language ability.
Want to try it yourself?
📥 Download our free bilingual PGU + MLU calculator and cut-off reference chart
Tag us on instagram @hablacadabraslp with your feedback or questions—we’d love to hear how it’s working for you!
