First of all, this sounds like a wonderful opportunity for you! Second, although you are quite right to exercise caution in test selection and interpretation of results, you'd be doing yourself and the study a disservice to rely solely on subjective reports and informal observations.

One thing working for you is that the official national language is English. That means that anyone who's been to school for a while or lives in a major city will be fluent enough in the language. But you still ought to include a self-reported fluency level in your protocol. How long people have attended school is also quite important for some measures. This may include ones that involve numbers or require a person to use a pencil, for example.

Also: Even though they may be fluent in English, it won't necessarily be the "same" English as is spoken in either the USA or the UK. So when you are administering a language measure (such as a receptive vocabulary test), you may find that some "easier" words are really "harder" and vice versa. Probably not a big issue, but worth examining (i.e., look at your data on the item level). More important: do not trust any measure of general knowledge that was developed for a different culture. And do not use proverb interpretation tests unless they were developed locally.

All that being said, there is no reason not to include measures of rote memory (verbal and visual), visuospatial cognition (e.g., constructional tasks like design copying or block design), working memory, or executive functions in your battery. Check out the literature; I think there are some learning/memory tests that have been developed specifically for West African populations.

Yes, thank you for your response but I have not been able to find adults test normed for West Africa and in particular Sierra Leone, therefore unless I have some appropriate norms to use I feel it would be better not to use standardised tests because of the norms-issue.

The two people I need to assess do not speak English but Criol and tribal languages, which would impact on verbal learning due to translation.

Is there evidence that executive tests assessed through pen and paper tasks have face validity with non-western culture? because to my knowledge there is actual evidence that even the formal assessment process can be difficult to engage with by some cultures, for instance the ones where formal education is very limited. 

Face validity isn't validity at all. Its only importance is insofar as it affects the person's willingness to take the task seriously.

If you don't have appropriate norms, you don't place full trust in any standardized scores or percentiles. But you need to actually assess what a person can do in order to assess what they can do. You can't just wing it based on self-report or intuition. You face a double challenge, then: 

First, finding or inventing tasks that basically work for the people in question. (For verbal learning, as an example, a list of words that are appropriate in difficulty, or a story that makes the right amount of sense, or both).

Second, gauging how well or poorly a person has done in the absence of formal norms. In a way, almost every neuropsychological assessment faces this problem, because we usually don't know what the person's premorbid level was. (Unless we have results from an earlier assessment - but that's rare.) So you would need to rely on demographic data (e.g., how far they went in school, what they did for a living) and maybe the results of any neuropsychological studies conducted there.

You can conduct a formal assessment without norms. It's what neurologists and psychiatrists often do when they run a formal mental status examination. Also think of Luria's approach. But you must put the mind through its paces with some set of tasks.

Thanks for your comment but I beg to disagree.

One can take into account behavioural observations and self-reports and third party reports. I agree that a list or story in the original language may be useful, but it is known that Black American tend to score lower on test standardised in the USA. A low score for a Black African from West Africa, what would it really mean? Would it indicate a decline or not? Sure one usually takes premorbid variable into account, and I would add usually one can give a test measuring premorbid functioning like the ToPF (not sure why you did not mention it?)

I think that rather than relying on no norms (psychiatrists usually use screening tests, like the MMSE or MoCA, which have not very good norms but a lot better than no norms at all) real life observations and data may be better than no valid data.

If anyone here can point me in the direction of cognitive tests that have been validated and normed for West Africans, that would be great.

Well, we're going to disagree. But there has been work done on neurocognitive functioning in sub-Saharan Africa. Here are a few leads, should you care to pursue them:

Here are a couple of articles reporting on neuropsychological investigations conducted in Cameroon a few years ago.

There was a big study of cognitive functioning after stroke conducted in Nigeria not long ago. First author Rufus Akenyemi is on ResearchGate. See link or article in the Journal of the Neurological Sciences. I think Raj Kalaria of Newcastle University was PI on this.

A team headed by Christopher Akolo conducted a similar study of cognitive impairment in early-stage HIV in Nigeria. He, too, is on ResearchGate; see link or article in Journal of NeuroVirology.

Norma Kabuba has done neuropsychological work in Zambia using western-developed tests and probably has useful input. She is on ResearchGate: see link or her article in Aids & Behavior last year.

Article Use of Western Neuropsychological Test Battery in Detecting ...

Article Neurocognitive Impairment Associated with Predominantly Earl...

Article Profile and determinants of vascular cognitive impairment in...

 Thanks for the suggestions and I will try contacting the authors, however the 1st and 3rd abstract clearly show than in the 1st case they had a normative sample of more than 300 people, the 3rd they used matched control group. 

The 2nd abstract is not clear on where they took the T-score demographically adjusted so I have asked for the article.

It would be of interest to me obviously to try get hold of the normative database, but norms are needed...

It may be you assume I will have endless resources and I will assess many people, but I am only going to assess one or two.