Essay on An Area of Phonology which my Learners have struggled With

Essay
An area of phonology which my learners have struggled with is pronouncing the /s/ and /z/ sound at the end of words.
This is something I first noticed when I started teaching in Cambodia.  It surprised me at first, because it had not been a problem in the other countries I had taught in.
At times, it was not clear if this was a grammar problem or a pronunciation problem.  When the omitted /s/ and /z/ sounds were at the end of plural nouns, or 3rd person singular present tense verbs, it appeared to be a grammar problem.  However, when it was part of the lexical item, it appeared to be a pronunciation problem.  (Some of the most common occurrences were “rice” pronounced as “rai”, “ice” pronounced as “ai”, “sometimes” and “always” pronounced as “sometime” and “always” and “is” pronounced as “i”).
What I also found strange is that it did not appear to be an articulation problem.  The students could pronounce /s/ and /z/ sounds at the beginning of words.  And they could pronounce the /s/ sound at the end of words in controlled practice if I prompted them (i.e. if I drew their attention to the error, and asked them to repeat it correctly.)  However during free production, they frequently dropped the final /s/ or /z/ sound.  I suspect this was because it sounded unnatural to them, and so they were reluctant to say it 
It also appeared that for years this error had gone uncorrected, and, since often their primary source of English was from other Cambodian speakers, it was also constantly being reinforced.  For most of my students, it appeared to be a fossilized error. 
Over the years, I have tried many different strategies to address this.
When I was fresh off of my CELTA, I used delayed error correction.  After every speaking activity, I would write a list of pronunciation errors up on the board, and ask the students to help me correct the errors.  Most of these pronunciation errors would involve /s/ or /z/ at the end of words.
As the class went on, the students and I began to realize that the same errors would appear again and again.  The students began to get embarrassed at their inability to correct these pronunciation mistakes, and some of them began to tell me that they believed it was counterproductive for me to constantly be writing these errors up on the board after every speaking activity.  “We know it’s not right,” said one of them.  “But when we are speaking, we don’t have time to remember everything.”
This coincided with literature I had been reading on second language acquisition.  Namely, human attention is limited, and so in free production, the learner does not have the cognitive resources to attend to all the features of language accuracy while also  focusing on the meaning of their message.
As a result, I stopped correcting errors with /s/ and /z/ during free production activities.  Instead, I decided to only address these errors during times when the students were reading aloud.
I reasoned that because reading aloud activities did not require the learners to form new messages, they should therefore have the cognitive resources free to attend to pronunciation errors.  So when the students were reading-aloud, I stopped them and corrected them every time they missed an /s/ or /z/ sound.
I also increased the amount of “read-aloud” activities in the classroom.  I put several children’s picture books on PowerPoint, and had the students take turns reading out the text of the book during class.
In retrospect, I may have underestimated just how demanding “read-aloud” activities actually are.  Although they appear simple, Martin Bygate points out that reading aloud actually requires “considerable attention” because you have to decide on the correct intonation as you read, and many times you have to predict the correct intonation before you know what the end of the sentence is going to be.  (Martin Bygate, Speaking p.10).  But this is something I would not realize until years later.
When I moved to Vietnam, I discovered a very similar problem existed with Vietnamese students.
However, unlike Cambodian students, I discovered that Vietnamese students often insisted on more overt correction from the teacher.  I had a couple classes complain to me that I was not correcting their pronunciation often enough.
Although it went against my CELTA training, I started experimenting with doing more hot correction during speaking activities. 
I did this despite the fact that my manager cautioned me against it.  My manager warned me that it was possible that not all the students wanted immediate correction, but some of them may have been too shy to voice their opinions.  My manager suggested I do learner training instead, and teach the students the difference between accuracy focused activities, and fluency focused activities.
Nevertheless, I decided that the easiest way to keep my students happy was to give them what they wanted.  (I was working in an English school that was also a business based on customer satisfaction, and part of my job evaluation was based on positive feedback from my students.)
I found that immediate correction often had very little effect on grammar errors, particularly complex grammar errors.  So the problems with the 3rd person singular “s” persisted.
However, immediate correction did have a very noticeable effect on pronunciation errors.  Particularly errors were there were no problems with the articulatory features, and correction was just a matter of drawing the learner’s attention to the error.  Whenever a learner would forget the final /s/ sound, I would draw an “s” in the air with my fingers, and they would correct themselves.  Eventually we got to the point where drawing the “s” was unnecessary.  I just had to look at them, and they knew they had to correct themselves.
Eventually, new habits began to form,
Many students that I have been teaching for a long time no longer have problems with pronouncing the /s/ at the end of words when it is part of base form of the word (e.g. “rice”, “is”, “always”).  However they still have trouble with the /s/ or /z/ when it is a grammatical morpheme (e.g. plurals and subject-verb agreement markers).