STAT 5303 Assignment 9
Predicting Flu Shot Incidence
An important issue among public health officials when flu epidemics are imminent,
is the encouraging of the taking of flu shots, especially among older people
who are at higher risk of developing complications. A study was conducted to
determine whether age and health awareness were important predictors of the
incidence of flu shots in a certain population. Toward the end of the flu
season, 50 people were randomly
selected and asked whether they had actually received a flu shot that
season. Their age was also recorded, along with information relating to their
awareness on general health issues which was combined into a health awareness index. A description of the variables appearing in the dataset along with abbreviated names is as follows.
- Flu Shot: Whether the person had a flu shot (yes=1, no=0).
- Age: Age (years).
- Awareness: Health awareness index (higher values indicate
greater awareness).
Using these data, conduct a chi-square test of independence and build a
logistic regression model to predict flu shot incidence, and summarize your
findings. Specifically, carry out/answer the following questions:
- Coding cases whose age is 50 or greater as "old" (the remainder being
"young"), carry out a chi-square test of independence between the factors Flu
Shot (yes or no) and Age (young or old).
- Build a multiple logistic regression model to predict flu shot incidence
using Age (in years) and Awareness (and possibly their interaction) as covariates.
- Test the fitted logistic regression model for goodness of fit. Are there
influential cases?
- Report the odds ratio corresponding to age from your fitted
model, and interpret its value. Does this odds ratio depend on Awareness? If so,
report and interpret it when Awareness=60.
- Using your fitted model, predict the probability that a person aged 50 with a
health awareness index of 60 gets a flu shot.
Instructions:
- Typeset your results in the same format as Assignment 7.