As part of a large study looking at factors associated with pre-term birth, the investigators were interested in the hypothesis that maternal smoking during pregnancy is associated with pre-term birth.

This is a case-control study. It included 204 women whose first child was born pre-term (at gestational age 37 weeks or earlier), and 3,766 women whose first child was born at or near full term (greater than 37 weeks of gestation).

1.1 State the null hypothesis and alternative hypothesis for a test of association

in a way that explicitly reflects the case-control design.

1.2 Explain why these data satisfy the assumptions to use the ? 2 test to assess association and the ? 2 distribution to find the P value.

1.3 Using the P value from this chisq.test function to explain the conclusion.

with a detailed summary sentence, taking account of the context and study design.

(Incorporate the null and alternative hypotheses in your summary.)

1.4 What proportion of mothers who had a pre-term birth smoked all through pregnancy?

What proportion of mothers who had a full term birth smoked all through pregnancy?

1.5 Calculate an estimate of the population odds ratio (OR) using these data

1.6 interpret the odds ratio in words.

1.7 Explain why these data, by themselves, do not provide evidence that smoking throughout pregnancy causes pre-term birth.

smoking during pregnancy all throughout never preterm birth 36 168 204 full term birth 370 3396 3766 > preg_smoke full term preterm never smoke 3396 168 smoke all thru 370 36 > > preg_smoke_test > preg_smoke_test Pearson's Chi-squared test data: preg_smoke X-squared = 12.898, df = 1, p-value = 0. 000329> epitab(preg_smoke, method = "oddsratio", conf. level = .95, oddsratio = "wald", pvalue = "chi2", correction = F) key part of output, reformatted full term preterm p1 never smoke 3396 0. 90175252 168 0. 8235294 smoke all thru 370 0. 09824748 36 0. 1764706 oddsratio lower upper p. value never smoke 1. 000000 NA NA NA smoke all thru 1. 966795 1. 350927 2. 86343 0. 0003290114