Independent sample t test!

#1
Hi all! One basic question I hope you can answer. I would really appreciate it :D:D

I have to columns of data. Both are daily returns (scale variable). The second column does not containt as many observations as column 1.

e.g

column 1 column 2
0,2%
0,3% 0,3%
0,5% 0,5%
0,2%

I do not understand how i can run the test. I am supposed to group the variables. But I dont know how to do that? Can somebody help me?
 
Last edited:

Dason

Ambassador to the humans
#2
Well we don't know what you're trying to test so we really can't help you. What is your hypothesis?
 
#3
Im trying to test the hypothesis that the return on the two series are different from each other. Ho: u1=u2, Ha u1 different from u2.

The first column contains 4016 observation. The numbers are returns on a stock. The second column has less than 4016 observations and is return data on the same stock. ((It only contains observations when the price of the asset is larger than the 50 day moving average (its a technical trading system)).

Both are scale variables (numbers only). I dont know how to grooup them. The only examples i see on this test online is when you have eg. heights for males and females and want to see if they have different heightes. Then you use the grouping variable sex to see if it is a different in the means of the two series. However i only have numeric data. Hence, I dont know what to do.

Was this clear, or do you have any further questions?

Iam online now and througout the evening!
 

Dason

Ambassador to the humans
#4
Are the column 2 values always the same as their corresponding column 1 values if there is a column 2 value? You example implies that.
 
#5
Yes, that is correct. But not always. Sometimes they dont take on any value at all (blank cell). And if certain conditions is fullfilled they take on the same value. I want to test if there is a difference between the two columns.

The thing is that column number two should be expected to have a higher mean than column 1 (which contains the full series) according to a financial theory.

Maybe it is possible two arrange them all in one column (under each other). and label them 1 if it is the original column 1 and 2 if it is original column 2?