I have what I hope is a simple question - is it possible to do time series analysis on a small data set specifically only four data points? I have collected human threat data (mean number of threats per kilometre walked/ survey) every 3 months in eight different sites (four with an experimental element and four control sites). I am trying to determine the best way to determine if there is a trend in the data though from what I''ve seen of TSA, our data set is too small. I did a simple paired t-test on the first and average of following three surveys that shows no significant difference but it of course ignores the detail of the latter three surveys. I am not sure how to organize the data for importation into R. I''ve looked at Crawley and one of Zuur publications but the TS data they work with is much longer term and not over several (independent) sites. Any advice, help, or direction is much appreciated! Best, Nicole ________________________ Nicole Gross-Camp, PhD Senior Research Associate University of East Anglia Norwich Rwanda: +250 (0)783647398 Skype: nicolegross-camp Based in Rwanda from September 2010 through July 2011: BP 512 Butare Rwanda personal website: http://www.uea.ac.uk/dev/People/Research+Associate/Nicole+Gross-Camp#info project website: www.ReDirectRwanda.com [[alternative HTML version deleted]]
Although this is not an R question.... actually, if I get such a question in our consulting centre, my advice is: Present a table of your 4 observations and that's it from a quantitative point of view. There is no sense in modelling or testing with 4 observations. Best, Uwe Ligges On 09.05.2011 16:20, Nicole Gross-Camp wrote:> I have what I hope is a simple question - is it possible to do time series > analysis on a small data set specifically only four data points? > > > > I have collected human threat data (mean number of threats per kilometre > walked/ survey) every 3 months in eight different sites (four with an > experimental element and four control sites). I am trying to determine the > best way to determine if there is a trend in the data though from what I've > seen of TSA, our data set is too small. I did a simple paired t-test on the > first and average of following three surveys that shows no significant > difference but it of course ignores the detail of the latter three surveys. > > > > I am not sure how to organize the data for importation into R. I've looked > at Crawley and one of Zuur publications but the TS data they work with is > much longer term and not over several (independent) sites. > > > > Any advice, help, or direction is much appreciated! > > > > Best, > > Nicole > > > > ________________________ > > Nicole Gross-Camp, PhD > > Senior Research Associate > > University of East Anglia > > Norwich > > Rwanda: +250 (0)783647398 > > Skype: nicolegross-camp > > > > Based in Rwanda from September 2010 through July 2011: > > BP 512 > > Butare > > Rwanda > > > > personal website: > http://www.uea.ac.uk/dev/People/Research+Associate/Nicole+Gross-Camp#info > > > > project website: > > www.ReDirectRwanda.com > > > > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.
On Mon, May 9, 2011 at 10:20 AM, Nicole Gross-Camp <n.gross-camp at uea.ac.uk> wrote:> I have what I hope is a simple question - is it possible to do time series > analysis on a small data set specifically only four data points? > > > > I have collected human threat data (mean number of threats per kilometre > walked/ survey) every 3 months in eight different sites (four with an > experimental element and four control sites). I am trying to determine the > best way to determine if there is a trend in the data though from what I've > seen of TSA, our data set is too small. I did a simple paired t-test on the > first and average of following three surveys that shows no significant > difference but it of course ignores the detail of the latter three surveys. > > > > I am not sure how to organize the data for importation into R. I've looked > at Crawley and one of Zuur publications but the TS data they work with is > much longer term and not over several (independent) sites. > > > > Any advice, help, or direction is much appreciated! >This i panel data where you have 8 time series, not one time series, right? You could try a gls model with an AR1 correlation error structure for each of the 8 series. See the analysis of the Owl data set in Zuur. -- Statistics & Software Consulting GKX Group, GKX Associates Inc. tel: 1-877-GKX-GROUP email: ggrothendieck at gmail.com