The Student Room Group

Chi-squared

When performing chi squared statistical test, how do you know what the results you expect are?


Posted from TSR Mobile
Reply 1
You have to assume independence and you multiply the probabilities together and then multiply by the sample size:

P(going to zoo) = 1/10
P(going to park) = 4/10

P(zoo & park) = 0.1 x 0.4 = 0.04 <- Probability that a randomly selected person goes to the park and the zoo, multiply by 10 to get expected value.
Reply 2
Original post by Albino
You have to assume independence and you multiply the probabilities together and then multiply by the sample size:

P(going to zoo) = 1/10
P(going to park) = 4/10

P(zoo & park) = 0.1 x 0.4 = 0.04 <- Probability that a randomly selected person goes to the park and the zoo, multiply by 10 to get expected value.


So on my example it's birds feeding from different feeders

Feeder A 53 birds
B 43 birds
C 30 birds

How do I approach that?


Posted from TSR Mobile
Reply 3
Original post by wrnicholls
So on my example it's birds feeding from different feeders

Feeder A 53 birds
B 43 birds
C 30 birds

How do I approach that?


Posted from TSR Mobile


Well assuming that the birds are all of the same type, there would only be 1 row/column. which means your expected would be the same as your observed, meaning that chi-squared is 0.
Reply 4
Your expected is number of observations/number of categories. So it would be (53+43+30)/3 because there are 3 feeders. It will be the same for a three feeders.
(edited 11 years ago)
Reply 5
Original post by Tabers
Your expected is number of observations/number of categories. So it would be (53+43+30)/3 because there are 3 feeders. It will be the same for a three feeders.


Ok so you expect the mean of the values you measure?


Posted from TSR Mobile

Quick Reply

Latest

Trending

Trending