Novice Picture Writing Proficiency Assessment w/ Rubric

I am trying out something new with writing in my Latin 1 classes. In previous years when I taught Latin 1, I had students doing timed writes after the first week based on readings/clipchats, but big picture, I never felt comfortable with that, because I felt like all students were doing was spitting back to me memorized sentences from the story, not necessarily any true acquired language. Since they have only had 8.5 weeks of Latin so far this year, they are "babies" when it comes to writing in the target language.

As a result, this year in my Latin 1 classes, I have been holding back on writing in Latin per se (outside of copying sentences from readings in lots of various activities or doing very guided writing. NOTE - these types of writing in the target language still have MUCH value!) so that I am not forcing output too quickly and have chosen instead rather to focusing on bathing them with sheltered vocabulary input before any true output.

After 8.5 weeks of Latin, I gave students their first written proficiency assessment based on a series of pictures which created a story (AI created the pictures). Although students had never seen the pictures before, they did mirror many of the sentences which had been in readings and also in bellringers. The difference, however, was that they had never written this many sentences before in Latin at once in one sitting and never in a sequence for the creation of a story. 

I gave them six pictures, of which there were at least 3-4 sentences which they could write for each of them. HOWEVER, I only asked them for 1-2 sentences for each picture. I debated about this:

• I knew that there would be some students who could only write one short, choppy sentence per picture.
• I also knew that there were many students who were already writing compound sentences using "and" or "because," so in writing one compound sentence, they were already demonstrating proficiency beyond the expectation.
• I was grading this on a rubric, so the minimum number of sentences was not the sole criteria but rather were they also adding details, combining details, etc in the picture?

Below is an example of one of the pictures used (I made two others):

Observations

1. Wow, I was absolutely amazed at what students wrote and how much they wrote! It was like after bathing them in so much input for the past 8.5 weeks that it overflowed similar to turning on a faucet - it all finally came out!
2. Student examples from picture 1 (in Latin and English):
1. Puer Marcus. Puer Julius. (The boy Marcus. The boy Julius).
2. Puer est Marcus. Marcus est parvus. (The boy is Marcus. Marcus is small).
3. Marcus est parvus. Julius est magnus (Marcus is small. Julius is big).
4. Marcus est parvus puer, et Julius est magnus puer. (Marcus is a small boy, and Julius is a big boy).
5. Hic est puer. Ei nomen est Marcus. Marcus est parvus et laetus puer. Hic est magnus puer. Ei nomen est Julius, et is est laetus. (This is a boy. His name is Marcus. Marcus is a small and happy boy. This is a big boy. His name is Julius, and he is happy).
3. It was so weird to see the Theory of Ordered Development occurring in their output right before my eyes. A number of students left out the Latin word "is" when trying to write "the boy is big" and instead wrote "the boy big." Before, I would have been frustrated over this and would have blamed the students, but knowing now the Theory of Ordered Development, I was like "Yay, you're proceeding along EXACTLY as you should be!"
4. Grading on a novice-low writing rubric informed by ACTFL Proficiency Guidelines made it so much easier! Because I was not focusing on errors and performance but rather on "show me what you can do and where you fall on the proficiency continuum/spectrum," this freed me greatly!
5. Yes, there were spelling and grammar errors - LOTS of them! However, when grammar errors are the norm and expectation since these students are novice learners, I only marked those errors if they impeded my understanding of their messages (even me as the "sympathetic receptor") which lowered their placement on the rubric.
6. I was surprised by how many students used vocabulary which I had not been truly targeting yet or for which I had not yet held them accountable. However, I had been using these words but usually they were glossed for bellringers, reading passages, and listening. Somehow, these students had already acquired those words, and those words were inside of them ready to come out!

I plan to do more of these kinds of writings!

New Page on my Blog - Proficiency-Based Grading

There is a new page on my blog - Proficiency-Based Grading. It is now listed at the top in the navigation bar. I hope that you will find the information useful.

Proficiency-Based Grading 

Interpretive Reading Comprehension Rubric

Recently in my Latin 1 classes, I gave them a sight reading comprehension assessment and graded it using a proficiency-based rubric created by Martina Bex and the Comprehensible Classroom. Because Martina's interpretive rubrics are on TPT(they are free!), I will only post a link here to her blog page which discusses how she uses them - her blog post has the actual link to that TPT page. 

I already have a blog post from a few years ago written about my experience doing a sight reading comprehension assessment, so I will share that here. However, let me sum up/recap what I did and the whys:

I gave students a short level-appropriate passage which they had never seen before but primarily used vocabulary which we had been targeting. This way then the reading was at least already 95% comprehensible (that is key!). Also, the sentence structures/patterns needed to reflect what they had seen - now was not the time to spring on them a different writing style since that was not what I was assessing! I only used a few words which needed to be glossed and did not overdo it (if a passage has too many glossed words, then it is not really readable/comprehensible since that disrupts reading flow). Why did I create a sight passage?

• It needed to be a sight passage of primarily known vocabulary, because this way I could tell if they had acquired these target words and could understand them in a new context.
• It also needed to be a sight passage of primarily known vocabulary in order to assess their reading comprehension skills. If I were to use an already-seen/known passage, that would not give me a true understanding of their proficiency, i.e., many students could fake their proficiency skills, because they already knew the passage ahead of time and could just answer from memory without ever reading the passage on the assessment.

For the record, I had ChatGPT create a framework of a story using the prompt, "Create a short story in 4th grade Latin using the words ___________________________. Repeat these words as much as possible. Do you understand?" Once ChatGPT created the story, I heavily edited it to fit my needs.

The rubric does address various depth of knowledge (DOK) reading comprehension questions, such as basic information, finer details about the plot, and inferring about the conclusion. As a result, the sight passage needed to have a basic plot but with underlying details to address, and a conclusion where an inference could be made about characters' motivation, choices, etc. All of the questions and answers were in English (see Martina Bex's blog post why reading comprehension should be addressed in English). 

I hope you will consider creating a sight-reading assessment and using the proficiency-based interpretive rubrics by Martina Bex!

Translation Rubric

Last week, I posted about using a proficiency-based rubric to grade student writing. Following that, a Latin teacher contacted me, asking if I had a proficiency-based rubric to grade student translation into L1 (English). The answer is, "Yes, I do, but..."

The question which was asked is a rather, complex one because of the terms "proficiency-based" and "translation." In many ways, the two terms are polar opposite. When we Latin teachers use the term "translation," we are grading for "accuracy" - did the student translate this word correctly and its form correctly, translate the verb tense correctly, indicate singular vs. plural, translate the participle/ablative absolute correctly, translate the sequence of tenses properly in light of the sentence, etc. From there, we mark off/deduct points - this is "performance" grading. Performance grading is quantifiable and reflects student performance in terms what they did not do correctly.

As stated in my previous post, proficiency-based grading is the opposite: 

This type of grading is not based on performance, which is the traditional "marking errors and deducting points," but rather on proficiency, i.e., what is it that students can do and what skills are they exhibiting in their language learning process? Language learning proficiency-based grading is holistic in nature and implements rubrics which are aligned with the ACTFL proficiency guidelines.

This now bring me to the other issue at hand: translation of L2 into L1 is not considered an ACTFL proficiency skill, and I fully agree with this. If you look at Bloom's Taxonomy, translation is listed as a very low-level skill, because in translating something from L2 into L1, all which a person is doing is establishing meaning. While translating may involve some degree of higher level thinking, no new meaning is created other than now the artifact exists in that person's L1. In a traditional Latin class, once the text has been translated from Latin into English, further higher-level discussion takes place in English.

So returning back to my original point - I have created a rubric for grading student translation. Translation is A skill which I teach, but it is not THE skill which I teach in my Latin classes, i.e., I also am focusing on listening, speaking, reading, and writing skills of the Latin language. Also, it is not based on ACTFL proficiency levels (Novice Low, Mid, etc), because these proficiency levels do not exist for translation. Rather, I have created a general "proficiency-inspired" translation rubric which can be used regardless of the level. 

Observations:

1. I love the holistic nature of the grading, because again, it shows me what students can do and not what they cannot.
2. Because I am already expecting there to be some grammar/translation errors due to their language level, that is not my true focus.

Caveats:

1. The key words are "instructional level" - I will simply say here that I believe that Latin needs to seriously realign what is considered to be "at instructional level" with what is actually realistic in terms of language acquisition. In upper level Latin courses, we are asking students to translate Superior-level Latin just after 2.5/3 years of the language.
2. This rubric could be applied for upper level classical readings or AP Latin, but I also would express caution. In my previous days of teaching AP Latin, I have found that most students will just memorize the English translation of a classical text and "spit it back" for a translation assessment. Essentially, the assessment becomes more about their memorization skills than translation abilities.

So if you are wanting to assess translation according to a rubric, consider using this but also understand that proficiency-based instruction (and not performance-based) is informing it.