Significant Test: Two Qs on ANOVA, Chi Square Tests

[honeysyd]

Significant test! please help

I have two questions.

1) if two sets of data have different sample numbers, then can we perform any significant test (e.g., anova, tukey had) to see whether there is a significant difference between the two sets of data?

2) This is a totally different question with (1). 80 blind people picked out one of 4 fruits from a basket. The results show that 64 blinds picked an apple, 16 blinds picked a strawberry, 13 blinds picked a banana, and 7 blinds picked an orange. Can we discover if the apples were picked by a chance or it has some statistical meaning? If so, can we use Chi square test?

Thanks in advance.

 

[haruspex]

1) It's not clear what you are asking. Is it this: the two sets of data are independent samples from two different populations, and you're trying to test whether the populations have the same distribution?

2) Passing over the fact that a blind person would have no more difficulty telling the fruit apart than would a sighted person:
- your stats add up to 100, yet each only took one?
- were there equal numbers of each fruit? Or is the point of the exercise to tell if there were different numbers?
- you can never use statistics to say whether something happened by chance; what you can do is estimate the probability of getting a specified result by chance from a specified distribution.

 

[Bacle2]

I have two questions.

1) if two sets of data have different sample numbers, then can we perform any significant test (e.g., anova, tukey had) to see whether there is a significant difference between the two sets of data?

2) This is a totally different question with (1). 80 blind people picked out one of 4 fruits from a basket. The results show that 64 blinds picked an apple, 16 blinds picked a strawberry, 13 blinds picked a banana, and 7 blinds picked an orange. Can we discover if the apples were picked by a chance or it has some statistical meaning? If so, can we use Chi square test?

Thanks in advance.

.There are significance tests for difference of means by using sample data. Is that what you're looking for?

Just a small nitpick that , in 2), 80≠ 64+16+13+7=100 . Did some people choose more

than one fruit?

If you want to test whether the frequency of a fruit chosen is statistically the same

as the frequency (distribution-wise ) of the fruits in the basket, you can do that, but

you need to have the distribution of the fruits in the basket and then run your test,

and decide if the choice of fruits was or not random.