Question 3 (3 points): Assuming there is no random variability or bias, indicate the type of confounding (each worth 0.5 point).

Unadjusted Relative Risk	Adjusted Relative Risk	Confounding Type
4.1	3.8
0.9	1.7
0.6	0.2
2.0	2.6
0.2	0.5
1.9	0.6

Question 4 (9 points): Researchers are studying the association between alcohol consumption and the development of pancreatic cancer. They are concerned there may be confounding based upon sex at birth and have decided to see if there is an association with alcohol consumption and pancreatic cancer development. Please calculate the odds ratio for the following 2x2 tables and interpret the results to indicate if you think there may be confounding.

Alcohol consumption vs pancreatic cancer development

	Cases	Controls	Total
High Consumption	124	98	222
Low Consumption	367	282	649
	491	380	871

Odds ratio =

Interpretation:

Sex at birth vs alcohol consumption

	Males	Females	Total
High Consumption	173	49	222
Low Consumption	151	498	649
	324	547	871

Odds ratio =

Interpretation:

Sex at birth vs pancreatic cancer development

	Cases	Controls	Total
Males	165	159	324
Females	326	221	547
	491	380	871

Odds ratio =

Interpretation:

Question 5 (4 points): The researchers studying the association between alcohol consumption and the development of pancreatic cancer now want to see if sex at birth explains their observed association between alcohol consumption and pancreatic cancer by using stratification-based methods. Please calculate the odds ratio for the following 2x2 tables and interpret the results to indicate if you think sex at birth explains the association between alcohol consumption and pancreatic cancer.

Males

	Cases	Controls	Total
High Consumption	43	130	173
Low Consumption	68	83	151
	111	213	324

Odds ratio =

Females

	Cases	Controls	Total
High Consumption	36	13	49
Low Consumption	198	300	498
	234	313	547

Odds ratio =

Interpretation:

Question 6 (3 points): Please calculate the Mantel-Haenszel weighted odds ratio for the association between alcohol consumption and the development of pancreatic cancer after stratification on sex at birth. Do you think it is appropriate to calculate this OR_MH?

Question 7 (3 points):Define residual confounding and discuss if there is a possibility for residual confounding in question 5 after controlling for sex at birth.