Dear colleagues,
I have some data on Poling Survey where we ask people about which party
they would vote for if they were to hold elections the next day. We listed
about 15 political parties and people ticked one as they prefer/support.
However, a group of people refused to answer the question or stated they
would not vote, etc. But we have another question where we list each of
these parties again and ask everyone to state on the scale from 1 to 5 how
close they are to each of these parties (1 being closest and 5 being most
distant). Thus we have each person stating his/her closeness status to
each of these parties. There are people who stated closeness to more than
one party at the same level. My intention is to work on classifying each of
these people who did not state the party they would vote for but stated
their closeness status to each of the political parties.
Does anyone have suggestions or literature on how to go about classifying
them into the possible parties they are likely to vote for using the fact
that they stated parties that they are close to, maybe using discriminant
analysis techniques?
Thanks in advance for your help.
Khangelani
|