Patterns of stress assignment are difficult to detect in CollSgE because of its syllable-timed rhythm. Since all syllables are given equal time, it is not always easy to detect relative differences in prominence among the syllables. This is unlike a stress-timed variety, where stressed syllables are typically realized with higher pitch, loudness and length.

Tay (1993: 27-28) suggests a number of ways in which CollSgE stress patterns are distinctive. One is the use of equal stress in words which otherwise receive primary and secondary stress. Thus, in RP, a word like celebration receives primary stress on the syllable bra. In CollSgE, however, all four syllables receive equal stress.

Another source of distinctiveness arises from the absence of differential stress patterns to mark changes in parts of speech. Thus, whether as a verb or a noun, the word increase is stressed in the same way; there is no difference in stress pattern corresponding to the change in grammatical category. This contrasts with RP, where in the case of increase, for example, stress is mainly on the second syllable (if meant as a verb) and on the first syllable (if meant as a noun).

Similarly, in RP, stress placement systematically distinguishes compounds from phrases. Thus if white house is a phrase, stress falls on house, while if it is a compound, stress falls on white. In CollSgE, however, regardless of whether it is a phrase or a compound, stress is consistently placed on the second word house. Thus stress in CollSgE does not distinguish nouns from verbs, nor compounds from phrases.

And finally, there is also the fact that in a number of words, the placement of stress simply occurs on a different syllable.

Trying to formulate a set of general rules that would predict how stress assignment works in CollSgE is not easy. However, there is a general opinion that stress in CollSgE tends to be oriented towards the end of a word. More specific attempts to describe the rules of CollSgE stress assignment run into difficulties. For example, Bao (1998: 169) suggests three possible rules: heavy syllables are stressed, stress occurs on alternative syllables, and if a word has more than one stressed syllable, the last stressed syllable carries the main stress. The distinction between heavy and non-heavy (light) syllables is based on the length of the vowel, which is assumed to be phonemically distinctive even though there is no phonetic evidence for this assumption. Bao thus acknowledges that for the rules to work, he has to assume that vowel length is phonemic in CollSgE. But this is a highly controversial assumption since there is no real evidence internal to CollSgE for treating vowel length as phonemic; the only justification is to argue, as Bao himself does, that RP (where vowel length is indeed phonemic) acts as the input to CollSgE. This is a position that other researchers may find untenable since it undermines claims that CollSgE can or should be analyzed as an autonomous variety without reference to more established varieties (e.g. Hung, 1995: 30).