A case study enrolled case individuals with glioma (a type of brain cancer) and similar control individuals without glioma. Participants were asked whether they had been regular cell phone users for at least a year. We use logistic regression to model glioma status (1 = glioma, 0 = no glioma) as a function of regular cell phone use (1 = yes, 0 = no) . Software gives the following information:
For log odds of glioma/no glioma
$\begin{array} { l c c } \text { Predictor } & \text { Coef } & \text { SE Coef } \\\text { Constant } & - 0.08074 & 0.04987 \\\text { Regular use } & - 0.02565 & 0.07275\end{array}$ What is a 95% confidence interval for the odds ratio e^β₁?

Question

A case study enrolled case individuals with glioma (a type of brain cancer) and similar control individuals without glioma. Participants were asked whether they had been regular cell phone users for at least a year. We use logistic regression to model glioma status (1 = glioma, 0 = no glioma) as a function of regular cell phone use (1 = yes, 0 = no) . Software gives the following information:
For log odds of glioma/no glioma

\begin{array} { l c c } \text { Predictor } & \text { Coef } & \text { SE Coef } \\\text { Constant } & - 0.08074 & 0.04987 \\\text { Regular use } & - 0.02565 & 0.07275\end{array}

What is a 95% confidence interval for the odds ratio e^β₁?

Quizplus · Accepted Answer

The 95% confidence interval for the odds ratio is calculated using the coefficient and its standard error. The interval for β1 is calculated as -0.02565 ± 1.96 * 0.07275, which gives the range for the log odds. Exponentiating these limits gives the confidence interval for the odds ratio, approximately (0.85, 1.12).

A Case Study Enrolled Case Individuals with Glioma (A Type