Effectiveness, and what this site reports instead

This cascade is adapted from the eligibility flowchart in the IUCN regional guidelines. It is an adaptation, not conformance to a standard: no participant-flow reporting standard exists for primary ecological studies. The nearest analogue covers evidence syntheses only. Each step is one of the gates set out under Eligibility.

How this page was produced derived

Every value below is computed at request time from arctic_api.species, arctic_api.species_listing_event and arctic_api.species_iucn by src/lib/effectiveness.ts. The conventions are stated before the values. No value on this page is entered by hand, and none is produced by a model.

Three terms, kept apart

Category movement is the direction of change between a species' IUCN Red List category at two points in time, measured on the 2001 ordinal scale Least Concern, Near Threatened, Vulnerable, Endangered, Critically Endangered. It is arithmetic on two labels. This is what this site reports.

Status trajectory is category movement restricted to genuine changes, those the IUCN Red List identifies as reflecting a real change in extinction risk rather than a taxonomic revision, a criteria revision, or improved knowledge.

This site does not report it, because the source data does not carry the flag it requires. The assessment table records a category and a publication year. It does not record whether a change was genuine. The distinction is therefore not computable here, in either direction, for any of the 43 species.

Effectiveness would be a status trajectory attributable to the Appendix listing. Attribution requires a counterfactual: what would have happened without the listing. This study has none.

It is a before and after design, uncontrolled, with no comparison group, no matched controls and no synthetic control, and because a CITES listing is not assigned by the researcher, it is nonexperimental in the strict sense. The class of inference it supports is descriptive and associational, not causal.

An indicator of status and trend documents progress toward a goal. Alone it cannot reveal whether an intervention affected that progress (Ferraro 2009). The confounds are of the two kinds that literature names: cotemporaneous factors correlated with both the listing and the outcome, which for these species means habitat loss, climate change and differing national management regimes, and selection bias, since species are not listed at random but because something was already wrong.

The cost is measurable. Before and after designs recover the true direction and magnitude of a biodiversity response 2.9 to 4.2 times less often than controlled before, after, control, impact designs, and a larger sample does not fix it: more data increases the precision of a biased estimate without moving it (Christie et al. 2019). Forty-three species would not have become forty-three matched pairs.

This site reports no effectiveness score, and there is no number on it that should be read as one.

Eligibility: 4 of 43 species have two comparable endpoints

A species can be scored only if both of its endpoints are on the same scale. Five gates, applied in order. A species that fails a gate is listed with the gate it failed, never omitted and never blank.

Three of these five gates follow an adopted standard. Two are this project's own reasoning. The difference is marked on every gate below, because a page that presented them as equally backed would be overstating its case.

1. Gate 1. Two global assessments have been published since the listing, a baseline and a later endpoint. 2 species fail.Adopted standard The Red List Index admits a species with insufficient information only once it has been assessed a second time (SDG Indicator Metadata 15.5.1, section 4.c).
2. Gate 2. The baseline and current categories are risk categories, not Data Deficient. 1 species fails: 0 at the baseline, 1 currently.Adopted standard The Red List Index denominator excludes species assessed as Data Deficient in the current period (SDG Indicator Metadata 15.5.1, section 4.c).
3. Gate 3. The baseline category is on the 2001 ordinal scale. 33 species fail.Own reasoning No adopted standard was located that covers comparison across the pre-2001 Red List systems. The argument is set out in the finding document.
4. Gate 4. Both endpoints were published after the 2001 criteria revision. 3 species fail.Own reasoning No adopted standard was located. The species it removes carry a modern category code at a pre-2001 vintage: the code survived the revision, the criteria behind it did not.
5. Gate 5. The change is flagged as genuine rather than a criteria revision or improved knowledge. 43 species fail. No species passes.Adopted standard The Red List Index counts only genuine category changes (SDG Indicator Metadata 15.5.1, section 4.c). The requirement is correct and this database cannot meet it.

Passing gates 1 to 4: 4 species. The cascade is 43, then 41, then 40, then 7, then 4.

Gate 3 removes 33 of 43 species, and this is the central limitation of the construct.

Most Arctic species were listed before the modern Red List existed. Their first post-listing assessment uses a category from the pre-1994 or 1994 systems, and three of those categories are not weaker labels for modern ones, they are different kinds of statement.

K, Insufficiently Known, is a claim about knowledge rather than risk. Monodon monoceros runs K in 1988 to Least Concern today. Reading that as a narwhal recovering rather than as an assessor finding out would be an error of a whole category.

LR/cd, Lower Risk conservation dependent, was discontinued in 2001 with no successor. Eschrichtius robustus runs LR/cd in 1996 to Least Concern. Whether that is an improvement depends on what LR/cd is read as. IUCN published no successor, so any reading is a convention: this project uses a published crosswalk between the two systems, which places it at Near Threatened (the manuscript uses the same one), and the species timelines mark such a point hollow and name both codes so a reader can see that it was placed rather than published.

NR and NA mean the species was not assessed. Branta hutchinsii carries NR at 1988, 1994, 2000 and 2004, then Least Concern. That is not a trajectory of any shape.

For the pre-1994 codes above, no mapping onto the ordinal scale is published and this site does not invent one: they carry no position, on this page or on any chart. LR/cd is the single exception and it is an adopted convention rather than an invented one, cited where it is used. Placing a category and scoring a movement are not the same act. The timeline places LR/cd so a reader can see where the species sat; this ledger still refuses to score the movement across it, at gate 3, because a category change over a criteria revision is not a change in extinction risk however the old code is read.

The ledger derived

All 43 species. A species that was not scored shows the gate it failed. There are no blanks and no species is omitted. derived

Among the 4 species passing gates 1 to 4: 0 improved, 2 stable, 2 deteriorated. Among the 7 passing gates 1 to 3: 1 improved, 3 stable, 3 deteriorated. Both are reported, with their gates named, because the difference between them is what gate 4 costs. Neither is an effectiveness result and neither has a denominator of 43.

Which way the exclusions bias the result

The 39 excluded species are not a random sample of the 43. Gate 3 removes 33, every one of them listed between 1975 and 1979, because early listing is exactly what puts a baseline in a superseded Red List system. Gate 1 removes two listed between 2012 and 2023, at the other extreme. So the four scorable species are the ones whose first assessment after listing already used the 2001 criteria, and their observation windows are short: 3 to 12 years, against listings that begin in 1975.

If long-listed species have systematically different trajectories from recently-listed ones, and there is no reason to assume they do not, then the direction of change among these four is not a smaller version of the portfolio's. It is a different sample. This is stated rather than corrected, because correcting it would require the mapping that gate 3 exists to refuse.

Which of the five possible meanings each number here speaks to

"Effectiveness" can mean five distinct things in this study. This table states which of them this site carries evidence for.

A note on the form of this page

Reporting a construct per species with explicit gates, rather than as one portfolio score, follows the practice of the major biodiversity indicators. The Red List Index excludes Data Deficient species from its denominator and admits a newly assessed species only once it has a second assessment. The Living Planet Index states its restrictions inline with its sample size and quantifies the bias each one introduces. That precedent is at the indicator level, not the primary-study level. No standard requires a single study to report this way, and this page is an argument by analogy rather than conformance to a rule. The analogy is stated so a reader can judge it.

References

Two works carry the argument above. The site links to nothing outside itself, so they are given in full rather than as links.

• Ferraro, P. J. (2009). Counterfactual thinking and impact evaluation in environmental policy. New Directions for Evaluation, 2009(122), 75-84. doi:10.1002/ev.297
• Christie, A. P., Amano, T., Martin, P. A., et al. (2019). Simple study designs in ecology produce inaccurate estimates of biodiversity responses. Journal of Applied Ecology, 56(12), 2742-2754. doi:10.1111/1365-2664.13499