Question

[The problem was slightly modified. Please read a problem carefully.]

A 1970s case-control study on cerebrovascular disease (thrombotic stroke) and oral contraceptive use in young women matched cases to controls according to neighborhood, age, sex, and race. The table below displays data from this study for thrombotic stroke. Perform an appropriate test at alpha level 1%. Please use McNemar’s test to test the association between cerebrovascular disease (thrombotic stroke) and oral contraceptive use

	Case exposed	Case Not exposed
Control exposed	3	8
Control Not exposed	43	54

1. State null and alternative hypotheses.

2. Compute a test statistics.

3. Compute a p-value using an appropriate table.

4. Can you reject the null hypothesis at alpha level 1%?

5. State your conclusion.

Answer 1

McNemar’s test

Claim : to test the association between cerebrovascular disease (thrombotic stroke) and oral contraceptive

## State null and alternative hypotheses.

Ho : there is not association between cerebrovascular disease (thrombotic stroke) and oral contraceptive

H1 : there is association between cerebrovascular disease (thrombotic stroke) and oral contraceptive

## Compute a test statistics.

chi square = (b - c)^2 / (b + c)

here a = 3 , b = 8 , c= 43 and d = 54

chi square = ( 8 -43)^2 / ( 8 + 43)

= 1225 / 51

= 24.0196

## Compute a p-value using an appropriate table.

The two-tailed P value is less than 0.0001 (use statistical table)

## Can you reject the null hypothesis at alpha level 1%?

yes we reject Ho because p value is greater than alpha value .

## State your conclusion.

there is sufficient evidence to conclude that there is   association between cerebrovascular disease (thrombotic stroke) and oral contraceptive .