| Variable | Description |
| Id | Identification Number |
| County | County Name |
| State | State |
| Area | Land area in square miles |
| Pop | Estimated population in 1990 |
| 18-34 | Percent of the population aged 18-34 |
| 65+ | Percent of the population 65 years old or older |
| Physicians | Number of active physicians |
| Beds | Number of hospital beds |
| Crimes | Total number of serious crimes (murder, rape, robbery, aggravated assault, etc.) |
| Graduates | Percent of adults (25 years old or older) that are high school graduates |
| Bachelor | Percent of adults (25 years old or older) that have a bachelors degree |
| Poor | Percent of the population with income below the poverty level |
| Unemployed | Percent of the labor force that is unemployed |
| Income | Per capita income of the population |
| TotIncone | Total personal income in millions of dollars |
| Region | Geographic region: 1= Northeast; 2=North Central; 3=South; 4=West |
1/ On consirère ici seulement comme variables explicatives la population totale,
revenu total, et la région.
a/ Tester au niveau 0.05 si, pour les valeurs Northeastern et North Central,
le facteur région a un effet significativement différent sur le nombre de médecins.
b/ Tester s'il y a un effet géèographique significattif au niveau 0.05.
2/ Faire une analyse basée sur une procédure de sélection. Prendre garde aux normalisations
des variables lors de la construction du modèle. Plusieurs pistes sont envisageables.
On n'utilisera pas la variable Beds