[RsR] use of lmrob() on ecological time series
Dear Emily Let me first ask the technical question: Do you still get convergence problems and the like when you apply lmrob(..., init=lmrob.lar, psi="lqq") ? It is of course helpful if papers dealing with the same problem use the same analysis procedures to make them comparable. On the other hand, alternative methods may give more precise and more interpretable answers. Thus, it would be interesting to compare the results that you get following the other papers' R script with those you get by an alternative procedure. If I understand correctly, you use the estimated slope your variable of interest, but then you classify these estimated slopes according to their statistical significance. With the three classes, significant positive, non-significant, significant negative you then do further analyses. This is at least my understanding of your analysis (1). Have you removed the non-significant series from the analysis? In (2), have you used the estimated slopes directly? Are these analyses also taken from the other papers, or does that only refer to the method of estimating the slopes? In any case, I have proposed that you use the estimated slope directly as the y variable (as you probably did in (2)) for two or more groups comparisons or in regression models. It will be more easy to explain what I mean if you send some of your time series as examples. Best regards. Werner Werner Stahel M +41 79 784 9330 | P +41 44 364 6424
Von: Emily Klein <emily.klein04 at gmail.com>
Gesendet: Dienstag, 27. August 2019 18:01:24
An: M?chler Martin
Cc: Stahel Werner A.; r-sig-robust at r-project.org
Betreff: Re: [RsR] use of lmrob() on ecological time series
Dear Martin and Werner,
Thank you very much for your responses here; I appreciate your time!
I know our initial use of the robust linear models is because of their use in previous research. I apologize for my naivete, but my understanding is that lmrob() fits a robust linear model to the data, and it then provides the model slope, which in my reading of previous work has been interpreted by others as the trend for that data and to indicate whether a population is increasing (positive slope) or decreasing (negative slope). That is, my reading suggests these authors used the earlier function lmRob() on logged abundance, and directly classified trends from that assessment. There was no further analyses beyond that (according to their R scripts).
We have followed that process, and then have assessed (1) the proportion of species within terrestrial, marine fish, or marine mammal groups under different management regimes (recreational hunt, commercial harvest, no management) that have positive or negative slopes, and (2) if the average slope for each group is significantly different from other groups (e.g. is the average slope for terrestrial species under X management significantly different from the average slope for terrestrial species under Y management, but also marine species under X or Y management). Does this seem to make sense, or is there a bit we're missing?
Unfortunately, we have very few species that have both managed and non-managed regions, as management doesn't really follow best experimental practices (which would be both really interesting but also potentially a problem for species being harvested or the people who would be banned from hunting them), although we have several that have had different management regions over time (and have looked at these more specifically - but only to compare slopes). In addition, we recognize the variability among species themselves (e.g. in response to environmental impacts, etc), but as we can't get around that, have instead aimed at using averages across many time series in each grouping.
In terms of the "non-significant change" I have also followed previous researches and pulled the confidence intervals, using these to remove time series with 0 inside that interval.
I may be missing more from your responses - apologies! Sometimes it takes me a bit longer to think through things. I very much appreciate your time and thoughts.
Emily
On Mon, Aug 26, 2019 at 4:19 AM Martin Maechler <maechler at stat.math.ethz.ch<mailto:maechler at stat.math.ethz.ch>> wrote:
>>>>> Stahel Werner A
>>>>> on Sat, 24 Aug 2019 20:53:22 +0000 writes:
> Dear Emily
> This is a late answer to your message from July 15.
> The first issue is the use of robust linear regression of log(abundance) on the year.
> I think that this is a very reasonable way to summarize the time series -- as long as
> a log-linear trend appears appropriate.
> The trend is then measured by the slope. A great advantage of using the log is that
> slopes are then on a common scale for rare and abundant species, as a certain slope
> corresponds to a certain percent increase or decrease per year.
> You then go on to question the classification into significantly increasing, and so on.
> This classification is common but unreasonable.
> A 5% increase per year can be significant for one species and insignificant for another,
> just because the former shows less random fluctuations than the latter.
> We should focus on estimation and supply confidence intervals for characterizing the
> (im-)precision.
> (You may have read about the controversy about "null hypothesis significance testing"
> and p values.)
> Since the slopes are used for further analysis, the classification is not needed nor helpful
> at all.
> In any case, I have not read in detail what is done with the slope. In one paper, it is used
> as the target variable in further regression models.
> I wonder if such regressions make sense when different species are used in the same
> regression. I thought it was a basic paradigm of biology that species have different
> ways to react to environments.
> If one simply want to show that management is helpful, one might compare managed and
> non-managed regions in terms of the number of species (within taxonomic groups?) that
> have recovered -- or directly in terms of average slopes for individual species or taxonomic
> groups.
> Nevertheless, let me add a thought about (2).
> I think the expression "non-significant change" is quite appropriate since a change of 0
> does not exist in real life. It is likely small (unless fluctuations are big and/or the time series
> short, which causes in-significance), but never precisely 0.
> Again, a confidence interval says it all: It contains all plausible values of the true slope.
Indeed! ... and the generic R function for confidence intervals,
confint() also has a good methods for "lmrob" objects, e.g.,
> (RlmST <- lmrob(log.light ~ log.Te, data = starsCYG))
Call:
lmrob(formula = log.light ~ log.Te, data = starsCYG)
\--> method = "MM"
Coefficients:
(Intercept) log.Te
-4.969 2.253
> confint(RlmST)
2.5 % 97.5 %
(Intercept) -11.8375506 1.898775
log.Te 0.7041937 3.802129
>
> Are these thoughts helpful?
> Werner Stahel
> M +41 79 784 9330 | P +41 44 364 6424
> ________________________________
> Von: R-SIG-Robust <r-sig-robust-bounces at r-project.org<mailto:r-sig-robust-bounces at r-project.org>> im Auftrag von Emily Klein <emily.klein04 at gmail.com<mailto:emily.klein04 at gmail.com>>
> Gesendet: Montag, 15. Juli 2019 21:34:45
> An: r-sig-robust at r-project.org<mailto:r-sig-robust at r-project.org>
> Betreff: [RsR] use of lmrob() on ecological time series
> Dear all,
> I am using the lmrob() function from the robustbase package, and I have a
> few questions. To keep the threads clear, I have a general inquiry here,
> and will ask more specific Qs in a second thread. NB: I don't typically
> update in the middle of a project, so am running on R version 3.4.1.
> (1) I am curious the community's thoughts on our approach: We have several
> hundred ecological time series and we're using robust linear models to
> determine if the time series are increasing, decreasing, or not changing,
> by looking at the modeled slope. This approach follows several others,
> including Lotze et al. 2017 (doi: 10.1111/cobi.12957) and Magare et al.
> 2013 (doi:10.1371/journal.pone.0077908). I don't have much experience with
> RLMs, so any thoughts on this approach would be very welcome.
> More specifically, following the work noted above, we are running (with the
> time series indexed with "DBx"):
> lm_test<-lmrob(log(pop_status+1)~observation_year,DBx)
> (2) Previous use of RLMs to identify the direction of ecological time
> series was asked in peer review to use "non-significant change" to
> reference time series with a slope of zero within the 95% confidence
> intervals. I can see excluding time series where there is no agreement on
> the direction of slope, but I think that slope=0 is more "stable" or "no
> change" and is not necessarily "non-significant". Any thoughts?
> Thank you all very much for any feedback you may have. I will start a
> second thread on a few warnings I am getting.
> Emily
> --
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Emily S. Klein, Senior Postdoctoral Associate (she / her / hers)
> The Frederick S. Pardee Center for the Study of the Longer-Range Future |
> Boston University
> *Co-Chair*, ICES Working Group on the History of Fish & Fisheries (WGHIST)
> esklein04 at gmail.com<mailto:esklein04 at gmail.com>
> http://www.bu.edu/pardee/
> http://www.ices.dk/community/groups/Pages/WGHIST.aspx
> _______________________________________________
> R-SIG-Robust at r-project.org<mailto:R-SIG-Robust at r-project.org> mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-robust
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Emily S. Klein, Senior Postdoctoral Associate (she / her / hers)
The Frederick S. Pardee Center for the Study of the Longer-Range Future | Boston University
Co-Chair, ICES Working Group on the History of Fish & Fisheries (WGHIST)
esklein04 at gmail.com<mailto:esklein04 at gmail.com>
http://www.bu.edu/pardee/
http://www.ices.dk/community/groups/Pages/WGHIST.aspx