I am collecting data to put into a db from application forms and I am not sure as how to design my tables. Example questions are:

Do you have a car? yes no.
IF yes, what model

If my understanding is correct I will have empty records for people who dont have a car. Does that mean I have a table called CARS that would have a 'model' attribute and some kind of primary key linking the person to the car.
What confuses me is that the table seems to be created dynamically. If a person has a car, create a new car instance etc.

Another which baffles me is:
What is the society's main source of income? Please tick relevant and estimate its percentage of your total income.
Options:
Government funding (TICK) Percentage :
Public donation (TICK) Percentage :
Self--funded (TICK) Percentage:

Does that mean in my schema I create a table for each option to store whether its ticked, and the percentage, and a society_num(PK) for the relationship to the society?
That would mean that a society_num would be duplicated as many times as options selected are selected? Am I thinking in the right direction or am I completely wrong?

Also what is the best approach for having multiple addresses or telephone numbers in a contact table for a single person when you dont know how many addresses or numbers that person may have.

Any help would greatly be appreciated!

Most of the items you mention sound like they could be handled in a flat table and, yes, there will be null fields ( or percentages set to 0 if the associated true/false item is "false). The addresses and phone numbers, however, might work best in a separate table, with a M:1 link to the primary table. This is because the number of entries is uncertain. If it were in the primary table, you wouldn't know how many fields to allot.

Hope that helps some.