statistical difference

#1
can anyone help me. Trying to determine if there is a significant difference between the death rates of sweden (2194) population 10,380,000 and Norway (201) population 5380000. I used an online calc, just not sure I did it right. It said 100% chance swedens is higher. and sweden would need a population of 60,000,000 before that went down to 69%, any help would be appreciated.
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
Do you understand the definition of a pvalue, if that is what you want to deal with? I don't think off hand you can get a pvalue of 100%, probably rounded off, but it should be small not large. Can you provide a link to the calculator you used.

I would recommend trying to calculate a rate difference with 99% confidence intervals. If you are feeling a little more adventurous you should check out the Bayesian version of this approach in order to get an actual probability related to a non-null hypothesis.

However, if you are truly dealing with populations, you don't need to conduct any tests at all, the difference in rates is all you need to make a conclusion - since there is not sampling variability - you have the true values!
 
#3
I used measuringu.com A/B Test Calculator. It reported a two tailed p value of 0, 100% the proportions are different and 100% chance sweden has a higher porportion. When I increased swedens population 50,000,000 it gave a p value .0373528 and norway became the higher group. Its been 30 years since I used real stat theory or applied statistics. Really trying to compare their death rates from covid, as they have different approach to handle it. Thanks for your reply.
 

hlsmith

Less is more. Stay pure. Stay poor.
#4
Well i clicked on the webpage, it is using a chi-sq test and is rounding the pvalue to 0 since it likely only reports so many decimal places. Its "report" should be interpreted as, there is a ( insert pvalue here) probability that you would get a difference in proportions this large or lager given that the two proportions are not different.

However, i would go back to the idea that no statistics are needed if this is the true confirmed case fatality rate between these two POPULATIONS, since there is no sampling going on. Given this, your solution would be a difference in case fatalities of 0.00016.