Research method
Mixed-Effects Model
A mixed-effects (multilevel) model estimates fixed effects of interest while allowing random intercepts or slopes for grouping units — children in schools, plots along transects, trees in a stand — so repeated measures are not treated as independent. In this set the same heading also covers a metapopulation PVA that compares invader-suppression scenarios. The random-effects structure is a statistical assumption about clustering, not proof that the grouping factor caused the outcome.
Environmental analysts use mixed models when the data are nested: yearly BMI on the same child, SOC along a heath–forest transect, tree-ring isotopes in one mountain range. The method answers 'after allowing for clustering, does this exposure or vegetation state still track the outcome?' Its main limitation is that a significant slope is not a randomised trial, space-for-time is not a future shrub-expansion experiment, and a PVA quasi-extinction time is not a field test that managers can hit those removal rates.
Evidence
What the evidence shows
Drawn from 4 studies in this library. Each finding starts with a plain-language takeaway, then the denser detail. Supports means evidence for a finding; Challenges means evidence against a stated position; Qualifies marks scope with a short note on each study’s contribution. Challenged positions are labeled — they are not findings.
Multilevel growth models of measured BMI in a school cohort (mean age 6.6 years; BMI rose ~2.6 units over 5 years) assigned baseline traffic density and CALINE-like freeway versus non-freeway NOₓ, adjusting for SES, smoking, greenness, fast food and street connectivity. Traffic within 150 m was confounded (especially the Spanish questionnaire). Non-freeway NOₓ stayed positively associated with BMI level and slope; freeway NOₓ was not.
Space-for-time mixed structure along 12 Abisko and 7 Vassijaure forest–heath transects treated shrub expansion as an analogue. At Abisko, organic-horizon SOC was 7.03 kg m⁻² on heath versus 2.98 (shrub) and 2.04 (forest). Total SOC fell from 9.01 to 4.51 kg m⁻² heath to forest. Shrub plots had the highest respiration; per kilogram of organic SOC, shrub and forest plots respired about three times heath rates. Hyphal in-growth rose from heath to forest edge.
Tree-ring δ¹³C, δ¹⁸O and δ¹⁵N plus growth for European beech, Norway spruce and silver fir in old-growth Šumava forests showed spruce iWUE saturation and growth declines in acidic, drought-prone periods; beech resilient to acid deposition; fir growth accelerating once air quality improved; beech gaining dominance. Thirty-five trees in one mountain range cannot represent all temperate forests or isolate a single driver experimentally.
A metapopulation PVA for invasive brown trout in Colorado River and Bright Angel Creek compared electrofishing, harvest, redd destruction, mechanical removal and climate/reservoir scenarios. Best combinations reached quasi-extinction in 5.7–6.3 years; skipping age-0 methods dropped QE probability to 29%; warming can shrink the mainstem but tributaries persist. The model does not field-test that managers can actually hit those suppression rates every year.
Open questions
Tensions and limits
Some items are genuine disagreements on the same question. Others mark different assays, populations, or outcomes — limits on how far one study travels — not a forced fight between papers.
Mixed-effects and adjacent hierarchical papers here are not one invasive-species result. Children's BMI growth is a traffic-pollution question (non-freeway NOₓ associated; freeway NOₓ not). Abisko space-for-time finds SOC falling from 9.01 to 4.51 kg m⁻² heath to forest as a shrub-expansion analogue. Šumava tree rings track beech versus spruce under shifting air chemistry in 35 trees. The brown-trout PVA is a suppression forecast (QE in 5.7–6.3 years; 29% if age-0 methods are skipped), not a mixed model of field counts proving those rates. The method label mixes four questions.
- Non-freeway NOx tracks children's BMI growth
- Shrubs and trees empty subarctic soil carbon
- Beech outcompetes spruce as air chemistry shifts
- Suppressing invasive brown trout networks
Study Role Design N Population Outcome Non-freeway NOx tracks children's BMI growth Supports CohortLongitudinal multilevel BMI growth models vs traffic density and CALINE-like NOx N=4550 · 4,550 children in 13 Southern California communities Schoolchildren in Southern California communities BMI level and 4-year growth related to non-freeway vs freeway NOx and near-road traffic Shrubs and trees empty subarctic soil carbon Supports Cross-sectionalSpace-for-time forest–heath transects comparing SOC and respiration under shrub expansion N=19 · 12 Abisko + 7 Vassijaure forest–heath transects Subarctic treeline heath, shrub, and birch forest plots in northern Sweden Soil organic carbon stocks and respiration rates across vegetation types Beech outcompetes spruce as air chemistry shifts Supports OtherMulti-century tree-ring isotopes and growth chronologies in Šumava old-growth forests N=35 · Isotopic dataset: 35 trees spanning ~500 years since 1501 CE European beech, Norway spruce, and silver fir in Šumava mountain forests Growth and water-use efficiency under historic acid deposition, N, drought, and warming Suppressing invasive brown trout networks Supports Computational / modellingMetapopulation PVA of invasive brown trout under removal and climate/reservoir scenarios Simulation scenarios (18 under current climate); not an empirical sample-size study Modeled brown trout in Colorado River and Bright Angel Creek subpopulations Time and probability of quasi-extinction under multi-life-stage suppression Even within vegetation change, Abisko and Šumava do not tell the same carbon or competitive story. Shrub and forest plots at Abisko held less SOC and respired about three times heath rates per kilogram of organic carbon. Šumava emphasises species replacement (beech gaining, spruce declining under acid/drought, fir recovering with cleaner air) rather than a measured SOC collapse along a shrub transect.
Study Role Design N Population Outcome Shrubs and trees empty subarctic soil carbon Supports Cross-sectionalSpace-for-time forest–heath transects comparing SOC and respiration under shrub expansion N=19 · 12 Abisko + 7 Vassijaure forest–heath transects Subarctic treeline heath, shrub, and birch forest plots in northern Sweden Soil organic carbon stocks and respiration rates across vegetation types Beech outcompetes spruce as air chemistry shifts Supports OtherMulti-century tree-ring isotopes and growth chronologies in Šumava old-growth forests N=35 · Isotopic dataset: 35 trees spanning ~500 years since 1501 CE European beech, Norway spruce, and silver fir in Šumava mountain forests Growth and water-use efficiency under historic acid deposition, N, drought, and warming
Common misconceptions
Random intercepts for school or child remove confounding, so the BMI–NOₓ slope is as good as an RCT.
Clustering is handled; diet was not fully measured, baseline-address exposure misses residential mobility, and freeway versus non-freeway NOₓ already split (only non-freeway stayed associated). Observational growth is not a randomised pollution trial.
Because forest plots have 2.04 kg m⁻² organic SOC versus 7.03 on heath, future shrub advance will lose the same carbon.
Space-for-time cannot prove that. Moth defoliation likely understated forest ectomycorrhizal growth; root C was not fully inventoried; permafrost and waterlogged tundra may respond differently.
Quasi-extinction in 5.7–6.3 years means brown trout will be gone on that timetable if managers try.
Those times assume suppression combinations the model was fed, including age-0 methods (without them QE probability fell to 29%). The analysis does not field-test that crews can hit those rates every year, and tributaries can persist even if warming shrinks the mainstem.
Exam-style questions
Short-answer questions that ask you to explain or compare, not recall.
Children's BMI rose ~2.6 units over 5 years from a mean age of 6.6. Non-freeway NOₓ stayed associated; freeway NOₓ did not; traffic within 150 m was confounded. What does a mixed growth model still require you to say in an exam answer?
The random structure accounts for repeated BMI on nested children/schools. The freeway split and 150 m confounding show that 'traffic' is not one exposure. Diet was incomplete and baseline addresses miss moves, so the slope is not a causal trial of NOₓ.
At Abisko, total SOC fell from 9.01 to 4.51 kg m⁻² heath to forest, with shrub/forest respiration about 3× heath per kg organic SOC. Why is that not a forecast of Arctic carbon loss under shrubification?
It is a space-for-time analogue with uninventoried root C, moth effects on forest ECM, and likely different responses on permafrost or waterlogged tundra. Mixed models of transect plots describe current contrasts, not a future experiment.
Šumava used 35 trees. Fir growth accelerated once air quality improved; spruce showed iWUE saturation and declines in acidic drought; beech was resilient. What causal claim is too strong?
That air-chemistry recovery alone caused assemblage change across temperate forests. One mountain range and 35 trees cannot isolate a single driver experimentally or represent all stands.
Best brown-trout suppression combinations reached QE in 5.7–6.3 years; skipping age-0 methods dropped QE probability to 29%. How should a manager read those two numbers together?
Age-0 methods are load-bearing in the model: without them, quasi-extinction becomes unlikely (29%). The times are scenario outputs, not a guarantee, and they still need field rates managers may not achieve; tributaries can persist under warming even if the mainstem shrinks.
The studies
4 studies in this library bear on Mixed-Effects Model, ordered by citations.
- Non-freeway NOx tracks children's BMI growth
In 4,550 Southern California children, non-freeway NOx associated with BMI at age 10 and 4-year growth; freeway NOx and near-road traffic density were weaker after confounding control.
- Shrubs and trees empty subarctic soil carbon
At a Swedish subarctic treeline, shrub and birch plots stored far less soil organic carbon than nearby heaths and respired that carbon faster, implying woody expansion can accelerate soil C loss.
- Suppressing invasive brown trout networks
Only metapopulation-wide, multi-life-stage suppression quickly drove invasive brown trout toward quasi-extinction.
- Beech outcompetes spruce as air chemistry shifts
Since 1501, spruce hit water-use saturation under acid rain and drought, while beech kept growing as nitrogen and temperatures rose.
Learn alongside