[R] package stats / prop.trend.test function

[Paulo Barata]

Professor Dalgaard,

Thank you for your reply. My question: shouldn't every statistical function in R, every statistical procedure, have some reference in the respective Help file?

Basic functions like sqrt, log, sort, sample, these are part of the base system in R, they generally have a clear purpose, a clear calculation to perform, and they are supposed to work well, no matter the exact details of their mathematical methods.

But higher level statistical functions, like prop.trend.test, or glm, they each perform some calculations according to a given method, but the method could possibly be somewhat different from the one implemented, with possible variations in the results.

Shouldn't the methods of statistical functions be always documented in the Help files, with references? Why there is just no information about the method of prop.trend.test?

Thank you once again for your attention.

Paulo Barata

Rio de Janeiro - Brazil

-------------------------------------------------------

On 12-Nov-25 9:49, Peter Dalgaard wrote:
> There's a worked example of the procedure in Altman: Practical Statistics for Medical Research, also featured in help(ISwR::caesarean).
> 
> (I have forgotten the origin of example(prop.trend.test), though. If I ever knew it. Apparently it is the same as in help(prop.test): Fleiss(1981), p.139, but I don't have that edition to hand and I don't see it in Fleiss(1973).)
> 
> -pd
> 
> 
>> On 12 Nov 2025, at 11.55, Michael Dewey <lists using dewey.myzen.co.uk> wrote:
>>
>> Dear Paulo
>>
>> You did realise that if you type prop.trend.test you get
>>
>> function (x, n, score = seq_along(x))
>> {
>>     method <- "Chi-squared Test for Trend in Proportions"
>>     dname <- paste(deparse1(substitute(x)), "out of", deparse1(substitute(n)),
>>         ",\n using scores:", paste(score, collapse = " "))
>>     x <- as.vector(x)
>>     n <- as.vector(n)
>>     p <- sum(x)/sum(n)
>>     w <- n/p/(1 - p)
>>     a <- anova(lm(freq ~ score, data = list(freq = x/n, score = as.vector(score)),
>>         weights = w))
>>     chisq <- c(`X-squared` = a["score", "Sum Sq"])
>>     structure(list(statistic = chisq, parameter = c(df = 1),
>>         p.value = pchisq(as.numeric(chisq), 1, lower.tail = FALSE),
>>         method = method, data.name = dname), class = "htest")
>> }
>> <bytecode: 0x0000021d81d4c850>
>> <environment: namespace:stats>
>>
>> Does that help?
>>
>> Michael
>>
>> On 11/11/2025 18:27, Paulo Barata wrote:
>>> To the R-Help list,
>>> About the prop.trend.test function, in the package stats.
>>> What calculations exactly does that function perform? What test is actually carried out? May I have some reference for the calculations performed by that function?
>>> I would like to suggest that some references (one or more) be included in the Help file for the function.
>>> Thank you very much.
>>> Paulo Barata
>>> Rio de Janeiro - Brazil
>>> ______________________________________________
>>> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>> PLEASE do read the posting guide https://www.R-project.org/posting- guide.html
>>> and provide commented, minimal, self-contained, reproducible code.
>>
>> -- 
>> Michael Dewey
>>
>> ______________________________________________
>> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide https://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>