+ Reply to Thread
Results 1 to 2 of 2

Thread: How to calculate significant differences in correlation using stata/spss

  1. #1
    Points: 4, Level: 1
    Level completed: 7%, Points required for next Level: 46

    Posts
    1
    Thanks
    0
    Thanked 0 Times in 0 Posts

    How to calculate significant differences in correlation using stata/spss




    Hello,

    This is probably a fairly simple question but I don't know to handle it properly.

    I have monthly data of 2 categories. Each category contains 8 assets, and of each asset I have the monthly returns for the period 1995-2014. I calculated the correlation matrix for each of the two categories and I have also computed the correlation matrix of the the categories combined using STATA.

    In my hypothesis, i expect that the 8 assets of category 1 are more correlated with each other in the 2008-2010 period than they are in the other 17 years. How can I calculate this and draw statistically significant conclusions using STATA?

    In another hypothesis, i expect that the average returns of category 2 are higher in the 2008-2010 period compared to the average returns of category 1.

    Finally, i expect the variance of returns in category 2 are lower in the 2008-2010 period compared to category 1.

    So how to calculate this? I can calculate the returns and variance of each individual category or asset and then compare the results with another category/asset but i Don't know how to draw significant conclusions from that using stata/spss.

    Sorry for my english, as I am not a native english speaker.

    Any help is much appreciated!




    Also, I expect that the average

  2. #2
    Points: 3,631, Level: 37
    Level completed: 88%, Points required for next Level: 19
    staassis's Avatar
    Location
    New York
    Posts
    226
    Thanks
    2
    Thanked 41 Times in 39 Posts

    Re: How to calculate significant differences in correlation using stata/spss


    To compare correlations between disjoint time periods or asset groups you can use an independent samples t-test (look up "t-test for correlation" and extend it). Alternatively, in almost any situation, you can use bootstrap.

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats