Statistics Question - Please help

Hi all. I am stuck on this question, and I really was needing some help/wanted to get your opinions. Let me know. Thank you!

Researchers conducted a case-control study to identify risk factors for kidney cancer. They asked 50 cases and 50 controls about 100 different exposures and personal characteristics, and calculated the odds ratio for kidney cancer for each risk factor. They found statistically significant odds ratios with 2 factors: coffee intake (p=.03) and cell phone usage (p=.04). The authors should conclude that:

A. Coffee and cell-phone usage increase the risk of kidney cancer.

B. The risks of coffee and cell-phone usage have been exaggerated due to the use of odds ratios.

C. These associations are likely chance findings.

D. The study had insufficient statistical power.

E. Coffee and cell-phone are statistically significant but not clinically significant risk factors for kidney cancer.

I know that E is NOT the answer. I was thinking that C might be the answer, but then I read this somewhere online: "In general, p-values of either 0.05 or 0.01 are used as a cutoff value, although this value is arbitrary. Results larger than the cutoff are considered likely to attribute the event to chance, while results smaller than the cutoff value are likely to have occurred because of a real explanation."

So now I am unsure. I just know that E is not the answer and I thought C, but I wasn't sure.

Thank you!!

You think that "C. These associations are likely chance findings." is correct? What do you think "statistically significant" means?

Well I wasn't sure about C being correct anymore because at first, I wasn't sure what the p value threshold was set relative to. I see now that statistically significant level was set to .05. I guess I have ruled out C and E. Can you give me some guidance on how best to approach this problem now.

Thank you.

statshelp123 said:

Hi all. I am stuck on this question, and I really was needing some help/wanted to get your opinions. Let me know. Thank you!

Researchers conducted a case-control study to identify risk factors for kidney cancer. They asked 50 cases and 50 controls about 100 different exposures and personal characteristics, and calculated the odds ratio for kidney cancer for each risk factor. They found statistically significant odds ratios with 2 factors: coffee intake (p=.03) and cell phone usage (p=.04). The authors should conclude that:

A. Coffee and cell-phone usage increase the risk of kidney cancer.
....NO, can't make an absolutely positive statement without qualifying the confidence level.

B. The risks of coffee and cell-phone usage have been exaggerated due to the use of odds ratios.

C. These associations are likely chance findings.
....NO, the problem states "statistically significant."

D. The study had insufficient statistical power.

E. Coffee and cell-phone are statistically significant but not clinically significant risk factors for kidney cancer.
....NO, can't refute "clinically significant" (whatever that means) if it wasn't measured.


I know that E is NOT the answer. I was thinking that C might be the answer, but then I read this somewhere online: "In general, p-values of either 0.05 or 0.01 are used as a cutoff value, although this value is arbitrary. Results larger than the cutoff are considered likely to attribute the event to chance, while results smaller than the cutoff value are likely to have occurred because of a real explanation."

So now I am unsure. I just know that E is not the answer and I thought C, but I wasn't sure.

Thank you!!

Click to expand...

I am left with B and D as possible answers. It would appear they used p=0.05 as the critical value, and the two specified factors net that criterion. mow we are stuck with trying to understand just how the data were handled. What is an "odds ratio"? did they use a t-test to find the "power function" corresponding to the sample size? Never having taken a course in statistics, I am not familiar with a lot of the jargon, and I seldom use any distributions beyond \(\displaystyle \chi^2\). But I often can follow your arguments - based on what you have studied in your course, what do YOU think?

So what did the data look like? There are two groups (size \(\displaystyle N=50\)) which either do or do not have kidney cancer, and they recorded coffee consumption for everybody. If I were doing it, I would calculate mean and standard deviation (\(\displaystyle \mu\pm\sigma\)) for coffee within each group. That would set me up to test for the difference of the two means. When treating the measured \(\displaystyle \mu\) as an estimator of the population, don't forget the standard deviation of the estimate is \(\displaystyle \sigma / \sqrt{N}\). There are tables that say how large the sample size has to be to be able to reject the null hypothesis (which would be that the two means are equal) at a specified confidence level.

On the other hand, they may have taken the ratio of the two expectation values, and compare to the null hypothesis that \(\displaystyle R= \mu_1/\mu_0=1\). IF they handle the standard deviation properly when taking the ratio, then the statistical result (probability that the null hypothesis be rejected) should be the same. [If I had no more information about what they did than what is given in the problem, I would junk the whole thing.]

Ah I know it's confusing me so much. I understand your reasoning for B and D but I don't know which one to choose. A is definitely incorrect?

Unfortunately, they didn't give me any more information. Anyone else on the forum want to weigh in??

Thanks!

Anyone have any updates for me? Any advice would be very helpful! Thank you!

You have been given a number of suggestions and asked some questions to clarify your initial question. Why have you not answered any of the questions? Have you applied any of the suggestions?