Fwd: st: Question on a quasi-time series

[rachel grant <rachelannegrant@gmail.com>]

Hi Ariel
I have been using regression approaches and only including the data
that I have for each year. Ideally I would like to keep the regression
based approach as I have now gone so far along that path...I am not
able to start a new statitistical technique in the time I have
available so is there an easy way fo correcting for the serial
dependency within the negative binomial regression. I have seen in
some papers they set the confidence limit at p=0.01 to cover this. Is
that a valid approach?

many thans
Rachel Grant

On 23 March 2011 15:43, Ariel Linden, DrPH <ariel.linden@gmail.com> wrote:
>
> It seems to me that time series analysis is not the appropriate approach
> here. If there are only a couple of monthly observations available for each
> year of data, I cannot imagine how the model will fit the missing 10 months
> in any accurate way.
>
> I would consider comparing only the observations for "like" months in each
> year so that you are limiting the evaluation to data you actually have.
> Similarly, you'd have a good rationale for limiting the analysis to only
> those months of observation if indeed no other observations occur outside
> the timeframe.
>
> As to which statistical approaches to take, I think that perhaps a
> difference-in-differences approach could be considered, or perhaps some type
> of correlation approach (i.e., intra-class, controlling for within year
> serial dependency)
>
> Just some food for thought...
>
>  Ariel
>
> Date: Tue, 22 Mar 2011 21:27:45 +0000
> From: rachel grant <rachelannegrant@gmail.com>
> Subject: st: Question on a quasi-time series
>
> Hi
> I have daily count data over a number of years totalling 317 cases.
> However it's not a true time series because I have not got a full year for
> each year, just a month or two.
> The data are likely to be partially serially dependent within years but not
> between years. So I am not sure how to correct for the possible serial
> dependency.
> What I have tried is ranking the data 1-317 and then using this ranking to
> use the command "tsset". Will this work? Alternatively should I simply
> increase my confidence level to p= 0.01 and not bother trying to correct for
> the autocorrelation? Thanks!
>
> - --
> regards, Rachel Grant
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>
> ------------------------------
>
> Date: Tue, 22 Mar 2011 21:39:56 +0000
> From: Nick Cox <n.j.cox@durham.ac.uk>
> Subject: st: RE: Question on a quasi-time series
>
> I wouldn't tackle it that way at all. Ranking values has nothing to do with
> time or dependence structure! Time series problems are rarely such that you
> can just tweak a critical level!
>
> You have a time series, just one with lots of missing data. You can still
> -tsset- on daily date and look at e.g. the autocorrelation function. You can
> still look for trends.
>
> The bigger question is why you have gaps. If the gappiness is capricious as
> far as the phenomena are concerned, that is the best news. On the other
> hand, it seems far more likely that either the organisms or the observers
> were visible or active in the field at certain times, e.g. seasonality for
> the organisms or it was research time with good weather and light and
> absence of teaching or committee work for the observers.
>
> (I am interpolating here a memory that your data are essentially ecological;
> I may be misremembering, or this problem may be different. Either way, just
> abstracting a problem from its context often obscures it and makes good
> advice more difficult.)
>
> Nick
> n.j.cox@durham.ac.uk
>
> rachel grant
>
> I have daily count data over a number of years totalling 317 cases.
> However it's not a true time series because I have not got a full year for
> each year, just a month or two.
> The data are likely to be partially serially dependent within years but not
> between years. So I am not sure how to correct for the possible serial
> dependency.
> What I have tried is ranking the data 1-317 and then using this ranking to
> use the command "tsset". Will this work? Alternatively should I simply
> increase my confidence level to p= 0.01 and not bother trying to correct for
> the autocorrelation? Thanks!
>
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/



--
regards, Rachel




--
regards, Rachel

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/