The basic 2 x 2 tables for exposure and disease is:

	Exposed	Not exposed	total
Case	693	307
Control	320	680
total			N

OR =

Serum porcelain.

Joke. Please laugh politely. Since there is no real variable that performs the way this one does, we have made it up.Suspend disbelief, and answer the questions as if this were a real variable. We have divided the group into those with a high serum porcelain and those with a low serum porcelain. The following table gives the proportion of persons in the main table who have a high serum porcelain (table 6)

Proportion with high serum porcelain
	Exposed	Not exposed	total
Case	0.71	0.51
Control	0.31	0.74
total

For example, this means that 31% of exposed controls had a high serum porcelain...etc.

From these two tables, construct the tables for potential confounding, calculated the stratum-specific ORs, the OR_mh and answer the questions.

High serum porcelain		Low serum porcelain
	Exposed	Not exposed	total			Exposed	Not exposed	total
Case					Case
Control					Control
total					total

OR₁ = OR₀ =

OR_mh =

1. Is confounding present? If so, is it NEGATIVE or POSTIVE?

2. Is effect modification present?

General Rules:

Does OR₁= OR₀? If yes, there is no effect modification

If no, effect modification is present

Does OR_mh = OR? If yes, there is no confounding.

Is OR_mh < OR? If yes, there is POSITIVE confounding (OR is falsely elevated).

Is OR_mh > OR? If yes, there is NEGATIVE confounding (OR is falsely lowered).