1
$\begingroup$

Sorry if the title is a bit vague however, i'm not sure exactly how to make my sentence concise. I have two times series: Amount invested into Iraq across time (in months) Price of a stock across time (averaged each month)

There area few things i want to find out however, i'm only at an intro level stats and econometrics in Uni.

Firstly how would i go about correlating these two time series i've read a few posts on here i'd like some clarification on the best approach to finding a correlation between them. so far i believe i have to use cross correlation which i hope to do using minitab. I want to do this to see if there is an over correlation between the two sets of data.

Secondly is there anyway i can find out when the data correlates the most and when it does not i have over 150 data points and i really like to see during which months the correlation grew or was at its strongest. I want to do this so i can analyse why correlation might have decreased and so on (perhaps attributing it to other macro factors).

I hope this makes sense and if you need any more clarification i'll answer straight away. Thanks for any points and i look forward to learning a little more today.

Improve this question
$\endgroup$
5
  • $\begingroup$ Are you using "correlation" in its conventional sense as a measure of linear association between two paired sets of data? If so, then in what sense do you mean that correlation "grew" or was "strongest"? Do you perhaps conceive of your sets of data in terms of a time-based moving window across narrow periods of time and you wish to study the association (perhaps after some fixed lag of time) between the data as that window progresses? $\endgroup$
    – whuber ♦
    Feb 25, 2013 at 16:21
  • $\begingroup$ Hello, currently i've only been looking at it in the linear association between the two sets of data. But i believe what your describing is what i want. I want to study the association across narrow periods so i can analyse what macro events other then an increase in war funding might have affected them. How would i go about doing what you have described with the monthly data i currently have for average price and monthly investment? Sorry for sounding dense i'm new to this area of stats. $\endgroup$
    – Joshnathan
    Feb 25, 2013 at 16:30
  • 1
    $\begingroup$ Cross correlation functions will allow you to do what you outline. Plotring te cros correlation of two time series will show peak at times of highest mutual correlation $\endgroup$
    – BGreene
    Feb 25, 2013 at 17:37
  • $\begingroup$ @Joshnathan - it's unclear what is the stumbling block in this or in the related question you asked a few hours later. Are you asking how to obtain correlations in Minitab (or --hint--in Excel)? Are you asking how to go about getting correlations for various time windows? Maybe your questions are not about statistics so much as about how to use your software, in which case Youtube or Help files might be the way to go. $\endgroup$
    – rolando2
    Feb 26, 2013 at 4:09
  • $\begingroup$ @Rolando2 i know how to use minitab i guess now its about how i choose the right lag for the data i have. I've been playing around with it but due to the fact that its relating to the financial markets and i only have monthly data i'm finding it hard to pick a an appropriate lag for the data i have. $\endgroup$
    – Joshnathan
    Feb 26, 2013 at 13:53

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.