A quantitative study is proposed to assess statistcally whether there is a causal relationship between having a smoking history and having pancreatic cancer.
By using statiscal tools, medical test results and personal data figures of the people included in the study can be compared. Some of the patients have and others do not have this type of cancer. Therefore patterns are investigated within the sample, and this will allow to conclude whether there statitics provide an evidence or not on the effect of smoking in generating pancreatic cancers.
The main weaknesses are, first of all, that the correct aplication of such methods requires someone who is both expert in medicine and statistics and sometimes is difficult to find this profile. Moreover, depending on the type of information that is possible to collect from the patients the resulting figures can be biased due to statiscal errors, and we can be attributing an causal effect of smoking on cancer that is not actually true.