Hello Khaled,

Forgive me if I'm misrepresenting the situation, but it appears as though this might be a class assignment or quiz.

Good luck with your work!

It looks like some referees comments on a paper.

is there a typo? Should it not be Chi sq = something, P=0.059

then the authors could report their power calcs and with only 20 used when 122 needed for presumed 80% power to detect required association at 0.05 alpha they could make some statemebt that more cases are required,

Thanks professor sir. The cases were rare regarding the title of the review. That is why I didn`t know how to answer Quest. no. 1 as there no chance to increase the sample size. The thesis was accepted in the faculty of medicine and statistical results were be treated. I will try to see any chance to increase the sample size with my team.

Thanks again .

Adding : SPSS didn`t show Fisher Exact test in the result , that is why I depend on chi square value.

Is there an alternative methods instead of increase sample size ( they are rare cases in the study ) ?

You can just make a statement that the difference in proportions was x (P=0.059)

while this is "not significant" at the conventional P=0.05 level you can still say that the probability of this difference ocuuring under the null hypothesis was 0.059.

Perhaps you could pool your results with those from another study, if you can find one. or get some more cases from another area/ hospital ..

Thanks a lot professor sir.

Finite Population Correction Formula (Fpc) :

Used to decrease ↓ the sample size (if the sample size more than 10 % of total population of interest),,, to save cost , time , data collectors` effort (Logistic support)…

The population of interest : 467

the sample was 22 child (patient) undergo pediatric surgery

22/467 = 4.7 % . That is why I couldn`t decrease sample size.

** So I will study the answer of dear professor sir James. Many thanks.

Dear discussants! Certain things strike mere here. (1) UDT means undescended testicle(s), but what is the question asked? What are we comparing (possibly left vs. right, or side-concordance vs. side-discordance)? (2) Whatever the research question is, it must be a general one and not one pertaining to the 467 boys. [Asymmetry in a concrete group of 467 would be interesting only if side-specific, very costly surgical equipment had to be purchased (!).] (2) The reviewer appears to have little statistical sense and seems to make a missing power calculation responsible for an 'unsuccessful' significance test. However, Leigh's typo guess sound sound in my ears. (3) The plus (+) in the denominator of the FPC formula should be a minus ( - ). Or else the formula is meant for a different purpose. If n = N, all uncertainty is gone since everyone has been examined, and the effective sample size becomes infinite (and not halved). Anyhow, where does this enter into your question, Khaled? Many question, sorry!!

Thanks professor sir,

- Regarding ques. no. 3 , this is my reference in the attached file.

- My colleague will answer the rest of the question.

Dear Jorgen

Total number of UDT operated at our unit was 467

From this cases we discovered 22 cases of cong anomalies of vas deferens and epididymis

Some UDT from this 22 cases were high(intraabdominal)

and some UDT from these 22 cases. were inguinal (law)

P value for location of UDT and associated anomalies in vas deferens and epididymis was 0.059

It is nearly equal to significant p value 0.05

Unfortunately the sample size is very small only 22case from467 case

but congenital anomalies of vas deferens and epididymis are law

I collected this 22 cases over 3 years

What can i do to improve my research?

Best regards

Sarah

Sarah , please check this Q&A thread:-

Sample size in rare diseases.

https://www.researchgate.net/post/Sample_size_in_rare_diseases

In question 1, it seems that the p = 0.059 and not the Chi square value.

In SPSS, you can use the Fisher's exact test: Analyse, descriptive statistics, crosstabs, exact, and then choose exact. The p value by Fisher's exact test will appear.

Yes professor Namir ,, the fisher exact test didn`t appear in the ( chi square value table) result like that in the attached file. I will try the steps that you listed to our team. Really thanks and I will check it as soon as possible.

Dear colleagues! The 467 is an irrelevant number; the pattern within the 22 (which you do not reveal to us) is the only important thing, and the conclusion you aim at is a general one (there is no finite ensemble that is being characterized). The exam example (Pass-vs.-Fail) in the screen copy above DOES show Fisher's exact P-value (0.048, and one-sided 0.024). Put in your data on the 22 boys, and report the result. Don't care about the P-value being 0.06 or 0.04.

Thanks for all dear professors for their scientific valuable answers . And I`m sorry to be late for acknowledgment. With best wishes doctor Sarah.

In the meantime I have looked at the finite population problem. The PLUS in the quoted text above should indeed be a MINUS.

The fpc = n' is meant to replace n in StdError formulas of the familiar form: SE = SDinterindivid/sqrt(n). So 'fpc' is not a correction factor but the RESULT ( also called n' ) of applying the appropriate correction to n.

The web has several explanations of this. Some contain a "-1" term that appears unwarranted. And they all forget a crucial point: if the individual observations are beset with measurement error, the correction applies to the contribution from inter-individual variation only, not to the measurement error part: SE2 = SDinterind2/( n' ) + SDmeasurem2/n .

Thanks a lot professor sir.

the results show that there are no significant differences

I think that the conditions of use of this test have not been checked

X² = (O-C) ² / C (o: observe, C: calculate) np> or = 5 for the chi square of Pearson and N> or = to 30

the chi square is test that allows compliance and independence

the use of this test makes it possible to show a relation or laison between two qualitative variables

it may seem that the chi square = 0.058 is an error, which must be checked

If p value is significant

Thanks doctor Atika for your answer , I appreciate your answer.