Chi squared (goodness of fit) test. How to find expected values?

I have a chi-sq qu, I'm having trouble with.

Bees visit different parts of a flower spike as follows:



Top = 8

Middle = 12

Bottom = 22



Do the bees show preference for the different parts of the flower?



How do I find the expected values??



Do I then use:

chi-sq = SUM(((O-E)^2)E);

so will I have 3 terms to add together??



Then I guess I need to look at the chi sq table and see if there is a difference.

Chi squared (goodness of fit) test. How to find expected values?
you would need to consider your basic assumption tht all bees will visit each part equally to begin with to formulate your EXPECTED numbers visiting each part (this need not be a whole number in theory)



so total bees - 42



42/3 = 14 so this is our EXPECTED no bees visiting each part



OBSERVED - EXPECTED gives



(8- 14) ^2 /14+ (12 - 14) ^2/14 + ( 22- 14)^2/14 = [6^2 + 2^2 + 8^2 ]/14

=[ 36 + 4 + 64] /14= 104/14 = 8 1/7



now look up on your chi squared table with degrees of freedom 3 - 1 = 2



then look along and you will find at what percentage this result become significant



usually greater than 5% is significant and greater than 1% is considered HIGHLY significant
Reply:If bees really have no preference for part of the flower visited then the expected frequencies would all be the same and total 42 so they would be 14 each.



Yes work out (O - E)^2/E for each of the three and then add them. You look up the Chi-squared table for two degrees of freedom. (A full explanation of this would be rather lengthy, but basically the third number in the expected table is fixed by the total.)