Month: April 2019

The Rule of Three: Simpler Evaluation

Teachers are uhhhh obsessive, especially about marking. We write and rewrite assessment instruments, when we could be hitting a bachata class, ripping up the Grand Wall after work, or kicking back with our five-year-old.

^ wanna be overloaded like her? ^ 😞😞

We spend too much time thinking about grading. Luckily for us, I’m gonna make the rest of your teaching career waaaaaay simpler by showing you how to make marking simple.

Various assessment gurus will tell you something fairly similar regarding attaching Numberz to Performancez: there are only three, (or maybe four), real levels of skill that one can accurately describe.  These are basically,
1. not yet proficient
2. functionally proficient
3. fully proficient
.

Breaking things down further is complicated, and therefore makes marking slower (and rubrics more complex and therefore harder for students to understand). The more you refine descriptors and levels, the harder it is to distinguish between them. 

Yes, sometimes more complex rubrics are called for, but not in a language class. And why not? Because the only teacher action which makes a difference for language learners is the amount and quality of input. 

So…imagine if you got marked on partying. They give you a Number for how well you party.
Q: what would the rubric look like?
A: like this…

1 You are on your way to the party.
2 You are standing in the doorway, chatting with the host, eyeing a nice martini.
3 You are shaking it on the dancefloor with thirty others, with your second drink, and the sexiest person at the party is checking you out.

Works? Sure! It’s simple, quick and accurate. Your Party Mark will be 34%, 66% or 100%. Now, say we also wanted to grade outfits. So we add this:

1 Sweats and slides are kinda basic…but hey, you got out of bed!
2 Business casual? You look good and respectable but no eyeballs/mentions for you.
3 Oh yeah! What’s yr Insta, gorgeous? 😁

If we mark our partiers on both behaviour and dress, we could get from 1/6 to 6/6, or 16%, 33%, 50%, 66%, 83%, 100%. This is pretty good.  We could add another criterion– say, flirting skills– and then our marks would range from 3/9 to 9/09, or 33%, 44%, 55%, 66%, 77%, 88% and 100%.

So here is our Rule of Three for Evaluation:

1. We focus on three levels of skill (not yet, just got it, fully proficient).
2. There is a clear difference between each level.
3. We do not mark more than three criteria.

Now, I’ma show y’all how this works for a language class. Here’s our oral interaction rubric (end of year, zero prep, totally 100% spontaneous & unplanned Q&A with a student, Level 2 and up in any language).

Here is the rubric. We are evaluating comprehension, functional accuracy and quantity of speech. 

3. I understand everything said. My errors have minimal impact on how understandable I am. I ask and answer questions, and keep conversation going, appropriately.

2. I understand much of what is said with some obvious gaps. My errors occasionally make me impossible to understand. I try to keep conversation going but sometimes have problems adding to/elaborating on what has been said.

1. I don’t understand much of what is said. My errors often make me hard to understand. I have consistent problems keeping the convo going.

This rubric is a 3×3 and generates marks between 3 and 9 out of 9 (ie 33%-100% in 11% intervals). It’s a nice mix of detail, fast, and simple. You basically never want a rubric more complex than 3×3 cos it gets too texty for kids to read.

There you go. Use it if you want it.

Anyway, a few notes to go with this (and with marking writing, or anything else):

A. You can mark via selective sample. Eg, for writing, say your kids pump out 300-word stories (mid Level 1). I’ll bet you dinner and a movie that marking any three sentences will show you their proficiency as accurately as reading the entire thing. Same goes for answering questions about a reading, or listening. Pick a small sample and go.

B. You will generally see marks “clustering.” The kid who understands all the questions/comments in an oral interview will probably also be able to speak well. This is cos most “skills” develop in concert. With our partying rubric, it is likely that Mr Dressed To Kill is also quite sociable, a good and enthusiastic dancer, etc. Yes, there will be the odd kid who understands everything but can’t say much, but this is uncommon.

Now would somebody please make rubrics for spontaneous written output and reading comp also? Create & share.

Let’s be DONE with marking questions and focus on what matters: finding cool input for kids, and making our grading quick & simple, so that we can relax after work & show up energised. Remember, one of C.I.’s greatest innovators at one point said that their method was developed to boost their golf score. The logic? Well-rested, happy teacher = good teacher 😁😁.

 

 

Can Anyone Teach a Language?

Today somebody asked about a school in Mexico that teaches Spanish via C.I., as the student very much likes her TPRS (etc) class. There were a bunch of responses and suggestions, and then the following.

OK, two points.

First, I havn’t said– or implied– anything Karen Rowan here says. No, language teachers are *not* replaceable “by any native speaker.” The teacher (native speaker or not) needs skills. A language teacher needs to be able to go slowly, repeat the words a lot, be interesting and restrict the vocab load. These are important skills.

Sure, everyone is entitled to their opinion.  I’m not sure how that is relevant, and my opinion is not what Rowan here says. I generally prefer it when people don’t put words in my mouth. However, in one-on-one situations…

Second, today’s question:
Q: Can anyone teach a language?
A: With basic training, yes!

If you are a native speaker of ____, you need to do the following– in a one-on-one setting; classrooms are different– to teach a language. If you are a student of ____, all you have to do is tell (or get somebody to tell) your teacher to do these.

  1. Base instruction around whatever interests the student. If Johnny gets to China and wants to talk food, have at! When I was in Guatemala, I was interested in politics, food and the lives of migrant workers. Most of the vocab one needs to know– high-frequency– will come up naturally during conversation.
  2.  Limit vocabulary and recycle it. People need to hear the vocab over and over, in slightly varied contexts, to get to the automatic recognition stage of acquisition. Here is an example from a conversation I recall from Xela, Guatemala, in January of 1992:

Teacher: do you like Guatemalan food?
Me: Yes.
T: I like Guatemalan food. I also like Mexican food. I like Salvadorean food a bit. Do you like Mexican food?
M: Yes.  Tacos delicious.
T: Yes, tacos are delicious. In Guatemala, tostadas are like tacos.  They are also delicious.

The day we had this convo, we later went to the market and ordered tostadas. You can see here that the teacher used tacos, delicious, are and I/you like over and over.

My teachers were able to circle– repeat in Q&A form– whatever we both wanted to talk about. When we needed new vocab, they said it, I wrote it down, and we used it.  Daily, my understanding of my host family’s Spanish grew. I never in four weeks saw a vocab list, got grammar homework, or got asked to conjugate verbs.

Most of my teachers were Uni students, but not full-time professional teachers.

3. If I had a student going abroad, I would tell them to get the teacher to tell stories, talk with the student about whatever interests them, go slow,  and circle the vocab. I would also suggest they use any apps (eg LingQ, Duolingo) that don’t bore them, and read whatever they find interesting.

This is really all you need to be an effective  language teacher for one-on-one situationsIf it’s a class, you’ll need a method (TPRS, Movietalk, Picturetalk, Story Listening, reading) plus training in these plus novels etc, plus general knowledge about assessment, classroom management, etc.