Posted: 4/14/2010 10:19:44 AM EDT
|
Let me start by saying that this is not a homework assignment, not for a class I'm taking, or for someone else's school work.
I'm curious about the chi squared test for survey data. A friend is doing a survey where she wants people to identify a voice that they have heard. They first hear the sample voice, wait a while, then try to pick that voice from 7 possible choices. How do chi squared statistics apply here? My guess so far is that you use the statistics to say how likely the observed distribution is to be random chance or something else. The problem is I've never used this for something that wasn't already described by a mathematical model, so I'm unsure what the expected distribution is. Do you basically assume that by chance each choice would have an equal probability (1/7th)? So if you had 10 people surveyed, the expected frequency would be 10/7 for each choice, while the observed is the actual number that picked it? Of course, in her survey the sample size (10) is low, so its hard to tell whether the observed distribution is likely to be pure chance or an actual trend that is observed. Am I on the right track here? :head scratch: |