Humans can produce a lot more sounds than any individual language ever uses. Each language just uses a selection - so that, inevitably, when you learn a new language you come across sounds that your brain is not used to. This first of all means that it has difficulty recognising them, and then of course hasn't the foggiest idea how to pronounce them. That's why so many learners have problems with the "th" sounds for instance, which occur in relatively few languages. The brain, trying to make sense of what, to it, is a "non-existent" sound, understands it as the nearest sound that it does recognise. That's why "th" will often be understood (as well as pronounced) as "s" or "t".
This also happens with sound combinations or the position of sounds in a word. Take the English "ng" sound as in
sing for instance. That only ever occurs at the end of a syllable in English - never in initial position. Try saying "ngis" - it tends to come out as "nyis", which is not the same. You really have to concentrate to say it properly. The same with sound combinations "str" "scr" and "spr" are frequent combinations in English and cause us no problems. But try saying "slr" or "sntr" -harder work.
I don't know the languages you mention, but I would hypothesise that it is this which is happening - that those particular sounds or sound combinations never occur in word final position (or possibly never occur at all) in those languages. It therefore becomes "hard work" for the learners. They may be able to do it when they're really concentrating on pronunciation, but as soon as they're thinking of other things too - what they want to say, the grammar and vocab necessary etc, - the brain "can't cope" and just simplifies life by reverting to the first language patterns. Which in this means dropping the final sounds.
For Thai students, you may find this site useful.
http://www.btinternet.com/~ted.power/l1thai.html