I'm not a student and the answer to this might actually save lives

[Njeri]

Does the following argument make statistical sense given the data? – if not how can it be corrected/improved? It is all about the huge number of babies dying in their first year in poor countries and whether we should invest in preschools to support the hygiene resources and practices in the home.

Recent US Aid statistics covering 72 countries show an average 32.1 month interval between births. The average child in a preschool may be taken as 4.5 years of age and the average number of children per family in the high child mortality countries may be taken as 5 years of age or above. So we can reasonably assume that every average child attending the preschool (with 50% of his siblings older, and 50% younger) will have a baby brother or sister under the age of 1.

- therefore engage with parents through the preschool and also benefit from the preschool child as an agent of change in the home...

[Dason]

I don't see how you're connecting anything here.

[Njeri]

Maybe that was due to a typo - it should read:
Recent US Aid statistics covering 72 countries show an average 32.1 month interval between births. The average child in a preschool may be taken as 4.5 years of age and the average number of children per family in the high child mortality countries may be taken as 5 or above. So we can reasonably assume that every average child attending the preschool (with 50% of his siblings older, and 50% younger) will have a baby brother or sister under the age of 1.

- therefore engage with parents through the preschool and also benefit from the preschool child as an agent of change in the home...

My question is do random distributions warrant the statement we can "reasonably assume"...I have no statistical background only general knowledge on the subject

[lorb]

average child attending the preschool (with 50% of his siblings older, and 50% younger)

Why? As there are more children aged 0 to 4.5 than there are children aged 4.5 to 9 wouldn't it be reasonable to assume the average 4.5 years old kid has more younger than older siblings? (Also, why do we need to take assumptions about that? Are there no demographic infos available about that?)

[emt22]

I will let the statistic pros to deal with the details, but there is a disconnect between the question you are asking and the data offered.

In particular, the notion of "average" here is problematic: the stimulus takes data from 72 countries (potentially with different levels of variation) and then goes to list the "average" child of that data set is 4.5 years old with a family size of 5 children or above in high child mortality countries (is the sample all high child mortality countries? no idea- but even if it is you are likely in trouble due to substantial variation).

So we can we reasonably assume anything? Probably not- the average child in this example probably doesn't exist, and even if he/she does, it is not likely to exist in all of the countries.

It's never prudent to form policy based on an aggregated sample of 72 countries because the notion of average is different in case- it has nothing do with random distributions.

Maybe that was due to a typo - it should read:
Recent US Aid statistics covering 72 countries show an average 32.1 month interval between births. The average child in a preschool may be taken as 4.5 years of age and the average number of children per family in the high child mortality countries may be taken as 5 or above. So we can reasonably assume that every average child attending the preschool (with 50% of his siblings older, and 50% younger) will have a baby brother or sister under the age of 1.

- therefore engage with parents through the preschool and also benefit from the preschool child as an agent of change in the home...

My question is do random distributions warrant the statement we can "reasonably assume"...I have no statistical background only general knowledge on the subject