Study guide & lexicon
The Epistemic Theater
How science knows what it knows, and what counts as science: a companion to the narrated program.
Each chapter lists the key positions and who holds them, the predation case files the chapter uses (anchored on Wilmers et al. 2025), formulas where they matter, every numbered lexicon term from the video with the time it appears, and practice prompts. For each prompt, separate what was observed, what was inferred, and what can be claimed. Timestamps match the video and the interactive player.
Ivideo 1:37–5:37
Aims of Science
Explain, predict, understand, or intervene?
1:44 Case file: the Northern Range elk · 2:44 What is an explanation? · 3:36 True, or merely useful? · 4:31 Observe, infer, claim
Key positions
| Idea | Who | In one line |
|---|---|---|
| Deductive-nomological model | Hempel & Oppenheim 1948 | Explanation is deduction from laws plus conditions; explanation and prediction share one form. |
| Explanatory asymmetry | Bromberger 1966 | The flagpole’s shadow lets you deduce its height but does not explain it. |
| Causal-mechanical explanation | Salmon 1984 | To explain is to show the causal processes and mechanisms that produced the event. |
| Unification | Kitcher 1989 | To explain is to fit many phenomena under few argument patterns. |
| No-miracles argument | Putnam 1975 | Realism is the only view that does not make the success of science a miracle. |
| Pessimistic meta-induction | Laudan 1981 | Many successful theories were false, so success is weak evidence of truth. |
| Constructive empiricism | van Fraassen 1980 | Science aims at empirical adequacy, truth about observables. |
| Structural realism | Worrall 1989 | What survives theory change is mathematical structure. |
| Toulmin model | Toulmin 1958 | Data support a claim through a warrant with backing, a qualifier, and possible rebuttals. |
Predation case files
| Study | System | What it shows |
|---|---|---|
| White et al. 2024 | Northern Range elk | >19,000 just before 1995; 3,915 in 2013; 7,000–9,000 now. |
| Clark & Hebblewhite 2021 | 52 predator-removal experiments | Removals often raised juvenile recruitment but rarely adult survival or abundance. |
| Vucetich et al. 2005 | Northern Range elk, 1961–2004 | First post-reintroduction decade driven mainly by hunting and drought. |
Lexicon
No. 0012:31
Aims of science
Goals that inquiry pursues: description, explanation, prediction, understanding, and intervention. They can conflict.
No. 0022:45
Deductive-nomological model
Hempel and Oppenheim’s account: an explanation is a valid deduction of the event from laws of nature plus initial conditions.
No. 0033:17
Causal-mechanical explanation
Salmon’s account: to explain is to describe the causal processes and mechanisms that produced the phenomenon.
No. 0043:17
Unification
Kitcher’s account: explanation fits many phenomena under a small number of argument patterns.
No. 0053:37
Scientific realism
The view that successful theories are approximately true, including their claims about unobservable entities.
No. 0063:37
No-miracles argument
Realism best explains the predictive success of science; otherwise that success would be a miracle (Putnam).
No. 0073:48
Pessimistic meta-induction
Many successful past theories were false, so success is weak evidence of truth for current ones (Laudan).
No. 0084:01
Constructive empiricism
Van Fraassen’s view that science aims only at empirical adequacy: truth about observables.
No. 0094:32
Observation
A record produced by senses or instruments; the data a claim rests on. Never wholly free of theory (Chapter VI).
No. 0104:32
Inference
Reasoning from observations to conclusions through models and assumptions.
No. 0114:32
Knowledge claim
A statement offered as knowledge, with explicit scope and confidence, open to scrutiny.
No. 0124:50
Warrant (Toulmin)
The general rule licensing the step from data to claim, supported by backing and limited by qualifiers and rebuttals.
Exam prompts
- Distinguish explanation from prediction. Give an ecological model that predicts without explaining, and one that explains but predicts poorly.
- State the no-miracles argument and the pessimistic meta-induction. Which position do you hold about “predation risk” or “carrying capacity”?
- Lay out the claim “wolves caused the Northern Range elk decline” as a Toulmin argument, including a qualifier and a rebuttal.
IIvideo 5:37–9:55
Knowing Nature
What is knowledge, and how do we seek it?
5:45 Knowledge and justification · 6:50 Four questions · 7:37 Positivism and postpositivism · 8:32 Interpretivism · 9:23 One herd, two lenses
Key positions
| Idea | Who | In one line |
|---|---|---|
| Justified true belief | Classical analysis | Knowledge as belief that is true and held for good reasons. |
| Gettier problem | Gettier 1963 | Justified true beliefs that are true by luck are not knowledge. |
| Foundationalism vs coherentism | — | Justification rests on basic beliefs, or on mutual support among beliefs. |
| Problem of induction | Hume 1748 | Past experience cannot logically guarantee the future. |
| Fallibilism | Peirce 1877 | Any belief may be wrong; inquiry stays open to correction. |
| Naturalized epistemology | Quine 1969 | Epistemology is continuous with science: the study of how knowers actually learn. |
| Research paradigm | Guba & Lincoln 1994, 2005 | Answers to ontology, epistemology, methodology, and axiology. |
| Positivism; verification principle | Comte; Vienna Circle; Ayer 1936 | Knowledge from observation; meaningful only if verifiable or analytic. Self-undermining. |
| Postpositivism; critical realism | Bhaskar 1975 | A real world, imperfectly known through fallible, probabilistic inquiry. |
| Interpretivism; Verstehen | Dilthey 1883; Weber 1922; Taylor 1971 | Human action is understood through meanings, not only explained by causes. |
| Social construction | Berger & Luckmann 1966 | Many social categories and institutions are created and sustained by people. |
Predation case files
| Study | System | What it shows |
|---|---|---|
| Haggerty & Travis 2006 | Elk on private land outside Yellowstone | Tolerance of elk on irrigated fields shaped herd distribution: meanings shape management. |
| Wilmers et al. 2025 | Social carrying capacity | The number of carnivores people tolerate is often below ecological carrying capacity. |
Lexicon
No. 0135:46
Epistemology
The study of knowledge: its nature, sources, justification, and limits.
No. 0145:46
Justified true belief
The classical analysis of knowledge: a belief that is true and held for good reasons.
No. 0155:58
Gettier problem
Cases of justified true belief that are true by luck, showing the classical analysis is incomplete.
No. 0166:15
Foundationalism vs coherentism
Justification rests on basic, self-justifying beliefs, or on the mutual support of a whole system of beliefs.
No. 0176:33
Fallibilism
Any belief, however well supported, might be mistaken and remains open to revision (Peirce).
No. 0186:51
Ontology
Assumptions about what exists and what kind of reality is being studied.
No. 0196:51
Methodology
The logic of how inquiry should proceed, as distinct from particular methods or techniques.
No. 0206:51
Axiology
The values and ethical commitments of a research tradition, including what researchers owe to others.
No. 0217:07
Research paradigm
A tradition’s combined answers to questions of ontology, epistemology, methodology, and axiology (Guba and Lincoln).
No. 0227:38
Positivism
Comte’s doctrine that genuine knowledge comes from observation and the laws summarizing it, not metaphysics.
No. 0237:38
Verification principle
Logical positivism’s criterion: a statement is meaningful only if empirically verifiable or true by definition.
No. 0248:12
Postpositivism
The fallibilist successor to positivism: a real world, known imperfectly, approached through design, statistics, and replication.
No. 0258:12
Critical realism
The view that a mind-independent reality exists and can be known, though only partially and through fallible theory.
No. 0268:33
Interpretivism
The tradition holding that social life must be understood through the meanings people give it, not only explained by causes.
No. 0278:33
Verstehen
Interpretive understanding of action from the actor’s point of view (Dilthey, Weber); contrasted with causal explanation, Erklären.
No. 0288:49
Social construction
The process by which categories and institutions are created and sustained by people (Berger and Luckmann).
No. 0299:06
Social carrying capacity
The population size of a species that people will tolerate, often below ecological carrying capacity.
Exam prompts
- Construct an ecological Gettier case of your own and say what it shows about justification.
- Compare postpositivism and interpretivism on ontology, epistemology, methodology, and axiology.
- Why did the verification principle fail, and what replaced it in practice?
IIIvideo 9:55–12:44
Inference
Which inferences preserve truth, and which expand knowledge?
10:02 Case file: the Kaibab deer · 10:47 Three kinds of inference · 11:49 Strong inference
Key positions
| Idea | Who | In one line |
|---|---|---|
| Deduction | — | Truth-preserving, not ampliative. Valid if the conclusion must follow; sound if valid with true premises. |
| Induction | Hume 1748 | Ampliative and fallible; cannot be justified without assuming nature is uniform. |
| Abduction / IBE | Peirce 1878; Lipton 2004 | Infer the hypothesis that would best explain a surprising fact; “loveliest” is not always “likeliest”. |
| Hypothetico-deductive method | — | Deduce observable consequences of a hypothesis and test them. |
| Multiple working hypotheses | Chamberlin 1890 | Hold rivals simultaneously to avoid attachment to one. |
| Strong inference | Platt 1964 | Devise rival hypotheses and crucial experiments that exclude some; iterate. |
Predation case files
| Study | System | What it shows |
|---|---|---|
| Leopold 1949 | Kaibab Plateau deer | Deer irruption and range collapse attributed to predator removal. |
| Hairston, Smith & Slobodkin 1960 | Green world hypothesis | The world is green because predators hold herbivores in check. |
| Caughley 1970 | Kaibab records | Data “unreliable and inconsistent”; cattle removal an untested rival. |
| Ford & Goheen 2015 | Trophic-cascade research | Call for strong inference and mechanism. |
Lexicon
No. 03010:18
Green world hypothesis
Hairston, Smith and Slobodkin (1960): the world is green because predators hold herbivores below levels that would strip vegetation.
No. 03110:18
Trophic cascade
Indirect effects of predators that propagate down a food chain, for example to plants by suppressing herbivores.
No. 03210:48
Deduction
Inference in which true premises guarantee a true conclusion; truth-preserving but not ampliative.
No. 03310:48
Validity and soundness
An argument is valid if its conclusion must follow from its premises, and sound if it is valid and its premises are true.
No. 03411:05
Induction
Ampliative inference from observed cases to general conclusions or unobserved cases; fallible.
No. 03511:05
Problem of induction
Hume’s argument that induction cannot be justified without circularly assuming that nature is uniform.
No. 03611:25
Abduction
Inference from a surprising observation to the hypothesis that would best explain it (Peirce); inference to the best explanation (Lipton).
No. 03711:50
Hypothetico-deductive method
Propose a hypothesis, deduce observable consequences, and test them; failure counts against the hypothesis.
No. 03812:21
Multiple working hypotheses
Chamberlin’s method of holding several rival explanations at once to avoid premature attachment to one.
No. 03912:21
Strong inference
Platt’s procedure: devise alternative hypotheses, design crucial experiments that exclude some, and iterate.
Exam prompts
- Classify each step of the Kaibab story as deduction, induction, or abduction, and identify the untested rival.
- Derive opposing predictions for fear-mediated versus density-mediated aspen recovery.
- What does Platt’s strong inference add to the hypothetico-deductive method?
IVvideo 12:44–15:45
Falsification
Can a single failed prediction refute a theory?
12:50 Popper’s asymmetry · 13:31 Case file: does fear lower pregnancy? · 14:23 Duhem, Quine, and the protective belt
Key positions
| Idea | Who | In one line |
|---|---|---|
| Logical asymmetry | Popper 1959 | No number of confirmations proves a universal claim; one counterinstance refutes it. |
| Corroboration | Popper 1963 | Survival of severe tests; not proof or probability. |
| Falsifiability as demarcation | Popper 1963 | Science makes risky predictions; theories that explain anything explain nothing. |
| Duhem–Quine thesis | Duhem 1906; Quine 1951 | Hypotheses are tested with auxiliaries, so a failure does not say what is false. |
| Research programmes | Lakatos 1970 | Hard core, protective belt; progressive if adjustments predict novel facts. |
| NHST is not falsification | Johnson 1999; Anderson et al. 2000 | Rejecting a null of no effect is not a risky test of a research hypothesis. |
Predation case files
| Study | System | What it shows |
|---|---|---|
| Creel et al. 2007 | Elk progesterone vs elk:wolf ratio | Reported lower progesterone where wolves were relatively more numerous. |
| White et al. 2011; Proffitt et al. 2014 | Direct pregnancy measurements | Pregnancy uniformly high regardless of wolf risk. |
| Middleton et al. 2013 | GPS-collared elk and wolves | Encounters too infrequent to disrupt foraging. |
| Kauffman et al. 2010; Beschta & Ripple 2013 | Risk and aspen browsing | No risk effect found; reply: elk sense risk where they meet wolves, not where wolves kill. |
Lexicon
No. 04012:51
Logical asymmetry
No finite set of confirming instances proves a universal claim, but one genuine counterinstance refutes it.
No. 04113:03
Corroboration
Popper’s term for a theory’s record of surviving severe attempts at refutation; not proof or probability.
No. 04213:21
Falsifiability
The property of a theory that some possible observation would contradict it; Popper’s mark of science.
No. 04313:32
Nonconsumptive effects
Effects of predators on prey through fear: changes in behavior or physiology that may reduce fitness without killing.
No. 04414:04
Auxiliary hypothesis
An assumption, such as how an instrument or index works, needed alongside the main hypothesis to derive a prediction.
No. 04514:24
Duhem–Quine thesis
Hypotheses are tested only in conjunction with auxiliary assumptions, so a failed prediction cannot by itself identify what is false.
No. 04614:56
Research program
Lakatos: a hard core of central commitments, shielded by a protective belt of revisable auxiliary hypotheses.
No. 04714:56
Progressive vs degenerating
A program is progressive when its adjustments predict novel facts that are confirmed, degenerating when they only accommodate failures.
No. 04815:29
Null hypothesis significance testing
Testing whether data are improbable under a null model, usually of no effect; rejecting it does not test a substantive hypothesis severely.
Exam prompts
- Is the hypothesis that fear lowers elk pregnancy falsified? Identify the auxiliaries and say what Lakatos would ask.
- Explain why Neptune and Vulcan illustrate the Duhem–Quine problem.
- Why is rejecting a null hypothesis not a Popperian test? What would a severe test of a cascade look like?
Vvideo 15:45–18:27
Paradigms
How does science change its mind?
15:51 Normal science and revolution · 16:39 Case file: from vermin to keystone · 17:44 Choosing between theories
Key positions
| Idea | Who | In one line |
|---|---|---|
| Normal science; puzzle-solving | Kuhn 1962 | Research within a shared paradigm; anomalies set aside until crisis. |
| Disciplinary matrix; exemplar | Kuhn 1970 | The two senses of paradigm: shared commitments, and model solutions students imitate. |
| Incommensurability; gestalt switch | Kuhn 1962 | Rival paradigms lack a common measure; change resembles seeing the duck–rabbit flip. |
| Theory-choice values | Kuhn 1977 | Accuracy, consistency, scope, simplicity, fruitfulness, weighed differently. |
| Epistemological anarchism | Feyerabend 1975 | “Anything goes”: no method survives the history of science intact. |
| Research traditions | Laudan 1977 | Traditions judged by problem-solving success. |
| Balance of nature to patch dynamics | Wu & Loucks 1995 | A candidate paradigm shift in ecology. |
Predation case files
| Study | System | What it shows |
|---|---|---|
| Leopold 1949 | “Thinking Like a Mountain” | A conversion from predator control to predator appreciation. |
| Estes et al. 1998; Terborgh et al. 2001; Crooks & Soulé 1999 | Exemplars of top-down control | Otters–kelp; predator-free islands; mesopredator release and songbirds. |
| Ripple & Beschta 2003, 2006, 2012 | Yellowstone cascade | Normal science extending the pattern to cottonwood, willow, and aspen. |
Lexicon
No. 04916:03
Paradigm (Kuhn)
A community’s shared disciplinary matrix of theory, methods, and values, anchored by exemplary problem solutions.
No. 05016:03
Exemplar
A concrete, model problem solution that students learn to imitate and that defines a paradigm in practice.
No. 05116:03
Normal science
Puzzle-solving research within an accepted paradigm, which sets aside anomalies rather than testing the paradigm itself.
No. 05216:21
Gestalt switch
Kuhn’s image for a paradigm change: the same data seen as a different whole, like the duck–rabbit figure.
No. 05316:21
Incommensurability
Rival paradigms lack a common measure: they differ in problems, standards, and meanings, so no neutral test decides.
No. 05417:09
Mesopredator release
An increase in mid-sized predators after apex predators decline, through reduced predation or competition.
No. 05517:45
Theory-choice values
Kuhn’s shared criteria of accuracy, consistency, scope, simplicity, and fruitfulness, weighed differently by different scientists.
No. 05618:02
Epistemological anarchism
Feyerabend’s thesis that no fixed methodological rule holds across the history of science: “anything goes.”
Exam prompts
- Is the move from predator control to trophic cascades a Kuhnian revolution? Use exemplars and anomalies in your answer.
- Distinguish Kuhn’s paradigm from a research paradigm in Guba and Lincoln’s sense.
- How can theory choice be rational if paradigms are incommensurable?
VIvideo 18:27–21:48
Observation
Is there such a thing as a pure observation?
18:33 Seeing as · 19:11 Case file: the five tallest stems · 20:16 Instruments and definitions · 21:22 Laden, not determined
Key positions
| Idea | Who | In one line |
|---|---|---|
| Theory-ladenness; seeing-as | Hanson 1958 | What is observed depends on concepts brought to it (Tycho and Kepler at dawn). |
| Data vs phenomena | Bogen & Woodward 1988 | Theories explain stable phenomena inferred from data, not the data themselves. |
| Triangulation | Hacking 1983 | Agreement among independent methods built on different theories. |
| Imperfect detection | MacKenzie et al. 2002 | Counts confound abundance with detection; non-detection ≠ absence. |
| Reflexivity | Interpretive research | Declaring the researcher’s position because it shapes what is noticed. |
Formulas
Expected count
E[C] = N · p (abundance × detection probability)
Occupancy, one site
P(detected at least once in K surveys) = ψ[1 − (1 − p)K]
Predation case files
| Study | System | What it shows |
|---|---|---|
| Ripple & Beschta 2007 | Five tallest aspen per stand | Designed to detect early recovery among the leading stems. |
| Painter et al. 2015 | Random plots and the five tallest stems, 2012 | Up to 65% of stands had a sapling above 200 cm; the range-wide inference rested largely on the random samples. |
| Kauffman et al. 2010 | All young aspen in random plots | No difference with predation risk. |
| Brice et al. 2022, 2024 | Both designs on the same stands | Both detect a cascade; tallest stems show whether, random samples show how strong. |
| Hobbs et al. 2024 | Willow sampling | Studies at well-watered sites overstated recovery. |
| Kohl et al. 2018, 2019 | GPS on elk, wolves, pumas | Diel landscape of fear invisible to sparse VHF data. |
Lexicon
No. 05718:51
Theory-ladenness
The dependence of what is observed, described, and recorded on the observer’s concepts, expectations, and instruments (Hanson).
No. 05818:51
Data vs phenomena
Data are particular records; phenomena are stable features inferred from them. Theories explain phenomena (Bogen and Woodward).
No. 05919:49
Sampling design
The rule for choosing which units to measure. It determines which population, and which question, the data can answer.
No. 06020:03
Sampling bias
Systematic difference between the sampled units and the population of interest, which distorts estimates.
No. 06120:34
Landscape of fear
The spatial and temporal pattern of predation risk perceived by prey, which can shape where and when they forage.
No. 06220:51
Operational definition
A definition of a concept in terms of the procedure used to measure it, such as “risk” as distance to a kill site.
No. 06321:06
Imperfect detection
Animals present may go unrecorded, so counts confound abundance with detection probability (MacKenzie et al.).
No. 06421:23
Triangulation
Agreement among independent methods with different assumptions, which reduces the chance a result is an artifact (Hacking).
No. 06521:39
Reflexivity
Critical attention to how the researcher’s position, values, and presence shape the research.
Exam prompts
- Explain how the tallest-stem and random-plot designs each embody a question, and which question each can answer.
- Give three operational definitions of “predation risk” and the theory each carries.
- Does theory-ladenness mean anything goes? Answer with Hacking’s triangulation.
VIIvideo 21:48–24:25
Models
How can a false model teach us something true?
21:54 The map and the territory · 22:49 Case file: the three-box cascade · 23:40 Models that act
Key positions
| Idea | Who | In one line |
|---|---|---|
| All models are wrong | Box 1976 | Models are useful because they omit; the question is whether they are wrong in ways that matter. |
| Three kinds of idealization | Weisberg 2007 | Galilean (tractability), minimalist (difference-makers), multiple-models. |
| Models as mediators | Morgan & Morrison 1999 | Partly independent of theory and world, so they can teach about both. |
| Generality, realism, precision | Levins 1966 | No model maximizes all three; robustness analysis across models. |
| Estimand vs estimate | — | The quantity of interest versus the model output; the link is an assumption. |
Predation case files
| Study | System | What it shows |
|---|---|---|
| Metz et al. 2020 | Pumas vs wolves, 1998–2004 | Pumas killed a larger share of Northern Range elk than wolves in most years. |
| Barber-Meyer et al. 2008 | Elk calves | Bears were the main predators of calves. |
| Ripple et al. 2025 vs a critique cited in Wilmers et al. 2025 | Willow crown volume | +1,500% estimate vs concern that the volume model was calibrated on lightly browsed plants. |
| Messier 1994; National Research Council 1997 | Predator pit | Little support for alternative stable states in wolf–moose systems. |
Lexicon
No. 06622:08
Idealization
Deliberate distortion or omission in a model to make it tractable or to isolate the factors that matter (Weisberg).
No. 06722:08
Models as mediators
Models are partly independent of both theory and data, so they can be used to learn about each (Morgan and Morrison).
No. 06822:27
Generality–realism–precision
Levins’s tradeoff: a model can maximize at most two of generality, realism, and precision.
No. 06922:27
Robustness analysis
Checking whether a result holds across models with different simplifying assumptions.
No. 07023:19
Extrapolation
Applying a model outside the range of conditions used to build or calibrate it.
No. 07123:19
Estimand vs estimate
The estimand is the quantity a study aims to know; the estimate is what a model and data produce. The link between them is an assumption.
No. 07223:41
Predator pit
A hypothesized low-density prey equilibrium maintained by predation, separated from a high-density equilibrium.
No. 07323:58
Alternative stable states
Different self-reinforcing states a system can occupy under the same conditions; shifting between them requires a large perturbation.
Exam prompts
- What is wrong, and what is right, about the three-box wolf–elk–aspen cascade?
- Explain extrapolation using the willow crown-volume dispute.
- How did the predator-pit model act in the world, and what did tests find?
VIIIvideo 24:25–28:55
Causality
What would have happened otherwise?
24:32 Association is not explanation · 25:31 Would that elk have died anyway? · 26:38 Drawing the causes · 27:37 Designs that break the symmetry
Key positions
| Idea | Who | In one line |
|---|---|---|
| Confounding; reverse causation; collider bias | — | Three non-causal sources of association. |
| Simpson’s paradox | — | Subgroup trends reverse when pooled; resolving it needs a causal model. |
| Regularity, counterfactual, interventionist | Hume 1748; Lewis 1973; Woodward 2003 | Three accounts of what a cause is. |
| Potential outcomes | Rubin 1974; Holland 1986 | Each unit has outcomes with and without treatment; only one is observed. |
| DAGs; back-door criterion; ladder | Pearl 2009; Pearl & Mackenzie 2018 | Block non-causal paths; association, intervention, counterfactuals. |
| Randomization; factorial designs | Fisher 1935 | Balance confounders in expectation; reveal interactions. |
| BACI; Hill’s viewpoints | Stewart-Oaten et al. 1986; Hill 1965 | Weighing observational evidence when experiments are impossible. |
Formulas
Individual causal effect
τi = Yi(1) − Yi(0) (never observed directly)
Average causal effect
ATE = E[Y(1)] − E[Y(0)], identified by randomization
Back-door adjustment
P(y | do(x)) = Σz P(y | x, z) P(z)
Predation case files
| Study | System | What it shows |
|---|---|---|
| Brodie et al. 2013 | 45 elk populations | Higher survival with both pumas and wolves, most likely because hunting quotas were cut; harvest dominated. |
| MacNulty et al. 2020 | Ages of killed cow elk | Wolf kills modally ~16 years old; hunter kills ~4: compensatory vs additive. |
| Ford & Goheen 2015; Wilmers et al. 2025 | Three causal links of a cascade | Herbivory–plants established by exclosures; predators–herbivores least certain; net effect mixed. |
| Bilyeu et al. 2008; Hobbs et al. 2024 | Dams × exclosures | Together, willows passed browse height in ~5 years; controls never did. |
| Hurley et al. 2011; Bishop et al. 2009 | Mule deer | Predator removal barely changed growth; winter feeding raised it 13% per year. |
| Wilmers et al. 2006 | Isle Royale parvovirus | Wolf crash as a natural experiment; moose grew faster and higher. |
Lexicon
No. 07424:33
Confounding
A common cause of exposure and outcome that creates or distorts an association between them.
No. 07524:33
Collider bias
A spurious association created by conditioning or selecting on a variable caused by both exposure and outcome.
No. 07625:18
Simpson’s paradox
An association within every subgroup that reverses when the subgroups are combined.
No. 07725:32
Counterfactual
A claim about what would have happened had some condition been different; the basis of Lewis’s analysis of causation.
No. 07825:32
Interventionist causation
C causes E if an intervention changing C, holding other things fixed, would change E (Woodward).
No. 07925:50
Fundamental problem of causal inference
Each unit reveals only one of its potential outcomes, so individual causal effects are never directly observed (Holland).
No. 08026:05
Additive vs compensatory mortality
Additive deaths raise total mortality; compensatory deaths replace deaths that would have occurred from other causes.
No. 08126:39
Directed acyclic graph
A diagram of causal assumptions in which arrows point from causes to effects and no path loops back on itself.
No. 08226:57
Back-door criterion
Pearl’s rule for which variables to adjust for: block every non-causal path that enters the exposure through an arrow pointing into it.
No. 08326:57
Ladder of causation
Pearl’s three levels of causal questions: association (seeing), intervention (doing), and counterfactuals (imagining).
No. 08427:38
Randomization
Assigning treatments by chance so that confounders are balanced across groups in expectation (Fisher).
No. 08527:38
Interaction
When the effect of one factor depends on the level of another, revealed by factorial designs.
No. 08628:21
Natural experiment
A comparison created by events outside the researcher’s control that approximates random assignment.
No. 08728:38
BACI design
Before-after-control-impact: compares change at impact sites with change at control sites over the same period.
Exam prompts
- Draw a DAG for the Northern Range and identify a back-door path between wolves and elk survival.
- Why is compensatory mortality a counterfactual claim? What evidence bears on it?
- Design a study that separates top-down from bottom-up limitation of a deer population.
IXvideo 28:55–31:49
Underdetermination
What if the evidence fits more than one theory?
29:02 One body of evidence, many theories · 29:50 Case file: one decline, many causes · 31:01 Breaking the tie
Key positions
| Idea | Who | In one line |
|---|---|---|
| Underdetermination | Duhem 1906; Quine 1951 | Evidence may not fix which theory is true. |
| Transient vs strong | — | Resolvable by further evidence, or empirically equivalent forever. |
| Equifinality | Beven & Freer 2001 | Many parameter sets fit the same record. |
| Pattern–process problem | Levin 1992 | Different processes generate the same pattern. |
| Multimodel inference | Burnham & Anderson 2002 | Rank and average candidate models by information criteria. |
| Scale dependence | Prugh & Sivy 2020; Sivy et al. 2017 | Positive associations locally, negative across landscapes. |
Formulas
AIC
AIC = −2 ln L + 2k; Δi = AICi − AICmin
Akaike weight
wi = exp(−Δi/2) / Σj exp(−Δj/2)
Predation case files
| Study | System | What it shows |
|---|---|---|
| Vucetich et al. 2005 | Elk decline, first decade | Hunting and drought; no comparable apportionment since. |
| MacNulty et al. 2020 | Winter range outside the park (20% → 80%) | Demographic, not behavioural: survival was higher outside. |
| Brice et al. 2024 | Aspen recruitment | Risk effects 8–28 times weaker than density effects. |
| Hobbs et al. 2024 vs Ripple et al. 2025 | Willow | Stalled system (possible alternative stable state) vs slow recovery, from the same data. |
Lexicon
No. 08829:03
Underdetermination
The evidence available is insufficient to determine which of several rival theories is true.
No. 08929:19
Transient vs strong underdetermination
Transient: resolvable by further evidence. Strong: rivals are empirically equivalent, so no possible evidence decides.
No. 09029:35
Equifinality
Many different model structures or parameter sets reproduce the same observations equally well (Beven).
No. 09129:35
Pattern–process problem
Different processes can generate the same pattern, so patterns alone rarely identify their cause (Levin).
No. 09231:02
Theoretical virtues
Non-empirical criteria such as simplicity, scope, fruitfulness, and coherence used to choose among theories that fit the data.
No. 09331:13
Multimodel inference
Ranking and averaging a set of candidate models by information criteria such as AIC, rather than selecting one (Burnham and Anderson).
No. 09431:30
Scale dependence
The direction or strength of an ecological relationship changes with the spatial or temporal scale of study (Wiens; Levin).
Exam prompts
- Which parts of the Northern Range story show transient underdetermination that has been resolved, and which remain open?
- Explain how scale can make two contradictory carnivore studies both correct.
- What does multimodel inference do that a single best model does not?
Xvideo 31:49–35:15
Bayesian Evidence
How much should evidence change belief?
31:56 Degrees of belief · 32:43 The strength of evidence · 33:34 Case file: updating on the cascade · 34:45 Priors and their critics
Key positions
| Idea | Who | In one line |
|---|---|---|
| Credence; Dutch book | Ramsey 1926 | Degrees of belief should obey probability, or a set of bets loses for sure. |
| Conditionalization; confirmation | — | Update to P(H | E); evidence confirms by raising probability. |
| Likelihood ratio | Royall 1997 | The strength of evidence; Royall’s three questions: evidence, belief, action. |
| Base-rate fallacy | — | Ignoring how rare the hypothesis is. |
| Washing out of priors | — | Different reasonable priors converge with enough shared data. |
| Severe testing | Mayo 2018 | Support only when a claim passed a test it would likely have failed if false. |
Formulas
Bayes’ theorem
P(H | E) = P(E | H) P(H) / P(E)
Odds form
posterior odds = likelihood ratio × prior odds; LR = P(E | H1) / P(E | H2)
Camera-trap example (illustrative)
P(wolf | flag) = (0.95 × 0.001) / (0.95 × 0.001 + 0.01 × 0.999) ≈ 0.087
Flat-prior reading of a CI (illustrative)
SE ≈ (0.05 + 0.41)/3.92 ≈ 0.117; P(effect < 0) ≈ Φ(0.18/0.117) ≈ 0.94
Predation case files
| Study | System | What it shows |
|---|---|---|
| McLaren & Peterson 1994 | Isle Royale tree rings | Fir growth tracked wolves and moose: a prior for cascades. |
| Kauffman 2010 → Hobbs 2024 | Yellowstone studies in sequence | Belief in some aspen cascade survived; belief that fear drives it fell. |
| Wilmers et al. 2025 | 16 large carnivore–mesocarnivore pairs | Mean correlation −0.18 (95% CI −0.41 to 0.05; p = 0.11). |
Lexicon
No. 09531:57
Credence
A degree of belief, represented as a probability between zero and one.
No. 09631:57
Dutch book argument
Credences that violate the probability axioms expose you to a set of bets that loses in every outcome (Ramsey).
No. 09732:16
Bayes’ theorem
P(H | E) = P(E | H) P(H) / P(E): the posterior is proportional to the prior times the likelihood.
No. 09832:30
Conditionalization
Updating by adopting, as the new credence in H, the old conditional probability of H given the evidence.
No. 09932:44
Likelihood ratio
P(E | H1) / P(E | H2): how much more probable the evidence is under one hypothesis than another; the measure of evidential strength.
No. 10032:58
Base-rate fallacy
Judging a hypothesis by the accuracy of a test while ignoring how rare the condition is to begin with.
No. 10134:46
Washing out of priors
With enough shared evidence, posteriors from different reasonable priors converge.
No. 10235:00
Severe testing
Mayo’s error-statistical criterion: data support a claim only if the claim passed a test it would probably have failed were it false.
Exam prompts
- Work the camera-trap base-rate example and explain why most flagged “wolves” are coyotes.
- What does a non-significant result with interval −0.41 to 0.05 not show? Answer in Bayesian and in error-statistical terms.
- How should a prior for the strength of the Yellowstone cascade change across the studies in this chapter?
XIvideo 35:15–38:47
Replication
Would it happen again, and elsewhere?
35:23 Four ways to hold up · 35:56 The replication crisis · 36:56 Case file: thirty kilometers apart · 38:13 Remedies and other standards
Key positions
| Idea | Who | In one line |
|---|---|---|
| Reproducibility; replicability | National Academies 2019 | Same data and code vs new data; plus generalizability and robustness. |
| Many analysts | Silberzahn et al. 2018 | 29 teams, one data set, estimates from slightly negative to strongly positive. |
| Replication crisis | Ioannidis 2005; Open Science Collaboration 2015 | 97% of originals significant; 36% of replications. |
| P-hacking; HARKing; forking paths | Kerr 1998; Gelman & Loken 2014 | Flexible analysis inflates false positives. |
| Pseudoreplication | Hurlbert 1984 | Non-independent measurements treated as replicates. |
| Preregistration; transferability | Nosek et al. 2018; Yates et al. 2018 | Fix plans before data; test models in new places. |
| Trustworthiness criteria | Lincoln & Guba 1985 | Credibility, transferability, dependability, confirmability. |
Predation case files
| Study | System | What it shows |
|---|---|---|
| Garrott et al. 2008 | Madison Headwaters elk | Cow survival down 5–15%; numbers down 5- to 10-fold. |
| Klein 1995; Roffler et al. 2023 | Coronation and Pleasant islands | Wolves nearly eliminated deer. |
| Griffin et al. 2011 | 11 elk populations, neonates | Only grizzly predation additive. |
| Brodie et al. 2013 | 45 populations, adult females | Human harvest dominated. |
| Prugh & Sivy 2020; Wilmers et al. 2025 | Where studies are vs where carnivores live | ~70% of interaction studies in protected areas; 15% of range protected, 4% strictly. |
Lexicon
No. 10335:24
Reproducibility
Obtaining consistent results from the same data using the same methods and code.
No. 10435:24
Replicability
Obtaining consistent results across studies that address the same question with new data.
No. 10535:38
Generalizability
The extent to which results apply in contexts, places, or populations beyond those studied.
No. 10635:38
Robustness
Whether a result survives reasonable alternative analyses of the same data.
No. 10736:18
Publication bias
Selective publication of positive or significant results, which inflates effects in the literature.
No. 10836:18
P-hacking
Trying analyses, subsets, or stopping rules until a result crosses a significance threshold.
No. 10936:18
HARKing
Hypothesizing after the results are known and presenting the hypothesis as if it had been made in advance (Kerr).
No. 11036:18
Garden of forking paths
Data-dependent analytic choices that inflate false positives even without deliberate fishing (Gelman and Loken).
No. 11136:40
Pseudoreplication
Treating non-independent measurements, such as repeated samples of one experimental unit, as independent replicates (Hurlbert).
No. 11237:44
External validity
Whether a result holds beyond the conditions, sites, or populations in which it was obtained.
No. 11337:59
Scope conditions
The circumstances under which a generalization is expected to hold.
No. 11437:59
Apparent competition
Two prey share a predator; one prey sustains more predators, which then depress the other (Holt).
No. 11538:14
Preregistration
Recording hypotheses, design, and analysis plan before data are collected or examined.
No. 11638:33
Trustworthiness criteria
Lincoln and Guba’s standards for interpretive research: credibility, transferability, dependability, and confirmability.
Exam prompts
- Distinguish reproducibility, replicability, generalizability, and robustness with one ecological example each.
- Why did wolves limit elk in the Madison Headwaters but not clearly on the Northern Range? State the scope conditions.
- What does the mismatch between where carnivore studies occur and where carnivores live imply for external validity?
XIIvideo 38:47–42:18
Values
Can science be value-free, and should it be?
38:53 The argument from inductive risk · 39:41 Case file: whose costs? · 40:38 Whose baseline? · 41:16 Objectivity as a social achievement
Key positions
| Idea | Who | In one line |
|---|---|---|
| Value-free ideal; discovery vs justification | Reichenbach 1938 | Values may guide questions and uses, not the weighing of evidence. |
| Argument from inductive risk | Rudner 1953; Douglas 2000, 2009 | How much evidence is enough depends on the cost of error. |
| Epistemic vs non-epistemic values | Douglas 2009 | Non-epistemic values may set thresholds but never substitute for evidence. |
| CUDOS norms | Merton 1942 | Communalism, universalism, disinterestedness, organized skepticism. |
| Transformative criticism | Longino 1990 | Objectivity as a property of critical communities. |
| Strong objectivity | Harding 1991 | Starting from marginalized standpoints exposes hidden assumptions. |
| Crisis discipline | Soulé 1985 | Conservation must act before evidence is complete; its goals are values. |
Predation case files
| Study | System | What it shows |
|---|---|---|
| Lamb et al. 2024; Serrouya et al. 2021 | Mountain caribou | −51% (1991–2023); 15 of 41 subpopulations functionally extirpated; apparent competition. |
| Raynor et al. 2021 | Wolves in Wisconsin | Deer–vehicle collisions −24%; benefits ≈ 63× livestock losses. |
| Gilbert et al. 2017 | Puma recovery, eastern US | Projected 155 human deaths averted over 30 years. |
| Wilmers et al. 2025 | Baselines | Disagreement over recovery partly reflects pre-European vs pre-Indigenous baselines. |
Lexicon
No. 11738:54
Value-free ideal
The view that non-epistemic values should play no role in the internal reasoning of science, only in choosing problems and applying results.
No. 11838:54
Discovery vs justification
Reichenbach’s distinction between how hypotheses arise and how they are evaluated against evidence.
No. 11939:08
Inductive risk
The chance of error in accepting or rejecting a hypothesis; weighing its consequences requires values (Rudner, Douglas).
No. 12039:23
Epistemic vs non-epistemic values
Values bearing on truth, such as accuracy and simplicity, versus values bearing on consequences, such as safety or livelihoods.
No. 12140:02
Type I and Type II errors
Rejecting a true null (a false alarm) versus failing to reject a false one (a miss); in management, each carries different costs.
No. 12240:39
Ecological baseline
The reference condition against which change or recovery is measured; choosing one is partly a value judgment.
No. 12340:58
Crisis discipline
Soulé’s description of conservation biology: a value-driven science that must act before evidence is complete.
No. 12441:17
CUDOS norms
Merton’s norms of science: communalism, universalism, disinterestedness, and organized skepticism.
No. 12541:17
Transformative criticism
Longino’s account of objectivity as a property of communities with venues for criticism, uptake, shared standards, and equality of authority.
No. 12641:36
Strong objectivity
Harding’s view that starting inquiry from marginalized standpoints exposes assumptions and so increases objectivity.
Exam prompts
- Apply the argument from inductive risk to a decision to cull wolves for caribou.
- Which values enter when choosing an ecological baseline for “recovery”?
- Explain how the Yellowstone debate illustrates objectivity as a social achievement.
XIIIvideo 42:18–46:05
Demarcation
What counts as science, and who decides?
42:26 Why the line matters · 43:05 A century of criteria · 44:21 The answer depends on the criterion · 45:15 Case file: the Kaibab story, revisited
Key positions
| Idea | Who | In one line |
|---|---|---|
| Verifiability; falsifiability | Ayer 1936; Popper 1959 | Early criteria; each fails as necessary and sufficient. |
| Puzzle-solving; progress | Kuhn 1962; Lakatos 1970 | Criteria based on traditions and programmes. |
| Stagnation test | Thagard 1978 | Pseudoscience stagnates while rivals progress, and its practitioners ignore that. |
| Methodological naturalism | McLean 1982; Kitzmiller 2005 | The working rule courts applied to creationism and intelligent design. |
| Daubert standard | Daubert 1993 | Testability, peer review, error rates, standards, general acceptance. |
| Demise of demarcation | Laudan 1983 | No necessary and sufficient conditions survive. |
| Cluster concept; systematicity | Pigliucci & Boudry 2013; Hoyningen-Huene 2013 | Looser revivals; science as more systematic knowledge. |
| Wissenschaft | — | German for all systematic knowledge, humanities included. |
Predation case files
| Study | System | What it shows |
|---|---|---|
| Leopold 1949 → Caughley 1970 | The Kaibab story, revisited | A weakly tested account lived in textbooks until criticism corrected it: demarcation concerns practices. |
Lexicon
No. 12742:27
Demarcation problem
The problem of distinguishing science from non-science and pseudoscience, and of saying on what grounds.
No. 12842:46
Daubert standard
The US test for admitting expert scientific testimony: testability, peer review, error rate, standards, and general acceptance.
No. 12943:06
Pseudoscience
A practice presented as science that fails its standards, for example by stagnating while ignoring progressive rivals (Thagard).
No. 13043:29
Methodological naturalism
The working rule that scientific explanations appeal only to natural causes, whatever one believes beyond them.
No. 13144:00
Cluster concept
A concept defined by a family of overlapping features, none necessary or sufficient on its own (Pigliucci).
No. 13244:00
Systematicity
Hoyningen-Huene’s thesis that scientific knowledge differs from everyday knowledge chiefly by being more systematic.
No. 13344:00
Wissenschaft
German for organized, systematic knowledge of any kind, including the humanities; broader than English “science”.
Exam prompts
- Apply three demarcation criteria to postpositivist ecology, interpretive human-dimensions research, and astrology. Where do the answers differ?
- Is the label “science” a definitional choice? What follows from choosing a narrow or a broad definition?
- Why does the Kaibab story matter for demarcation?
XIVvideo 46:05–48:59
Synthesis
How do you audit a knowledge claim?
46:13 Auditing a famous claim · 47:38 Seven questions for any claim · 48:19 Exam rehearsal
Seven-question claim audit
- What was observed, and how?
- What was inferred, under which assumptions?
- Which alternatives remain?
- How strong is the evidence, and what would change your mind?
- Has it been replicated, and where does it generalize?
- Which values and stakes shape the claim?
- What would another tradition of inquiry ask that this one did not?
Lexicon
No. 13447:39
Claim audit
Seven questions: what was observed and how; what was inferred and how; alternatives; strength of evidence; replication and scope; values; other traditions.
Rehearsal answers
What separates induction from abduction, and where does strong inference fit?
Induction generalizes from observed cases to unobserved ones; abduction infers the hypothesis that would best explain a surprising fact. Strong inference disciplines abduction by generating rival explanations and designing tests that exclude some, so the loveliest story must also survive elimination.
Why does the Duhem–Quine thesis complicate falsification?
A prediction follows from a hypothesis plus auxiliary assumptions, so a failed prediction refutes only the conjunction. Logic cannot say whether the hypothesis or an auxiliary (an index, an instrument, a measure of risk) is at fault; Lakatos asks whether rescuing adjustments predict novel facts.
What did Kuhn mean by incommensurability?
Rival paradigms differ in problems, standards, and meanings of terms, so there is no neutral measure by which a single test decides between them; choice relies on shared values weighed by judgment.
Why is compensatory mortality a counterfactual claim?
It asserts that animals killed by a predator would have died anyway from other causes: a claim about an unobserved outcome. Evidence such as the age and condition of killed animals (wolf-killed cow elk ~16 years, hunter-killed ~4) bears on it indirectly.
What is the argument from inductive risk?
Because evidence never makes a hypothesis certain, accepting it requires judging the evidence sufficient, and sufficiency depends on the costs of false acceptance and false rejection. Non-epistemic values therefore legitimately set evidential thresholds (Rudner; Douglas), though they must not replace evidence.
How does the answer to “what counts as science?” change with the criterion?
Under falsifiability even astrology qualifies; under institutional norms and progress interpretive research qualifies and astrology does not; under cluster, systematicity, or Wissenschaft accounts both research traditions qualify. The practices do not change; the label does.
References
Predation casebook
- Barber-Meyer, S.M., Mech, L.D. & White, P.J. (2008) Elk calf survival and mortality following wolf restoration to Yellowstone National Park. Wildlife Monographs 169: 1–30.
- Beschta, R.L. & Ripple, W.J. (2013) Are wolves saving Yellowstone’s aspen? A landscape-level test of a behaviorally mediated trophic cascade: comment. Ecology 94: 1420–1425.
- Bilyeu, D.M., Cooper, D.J. & Hobbs, N.T. (2008) Water tables constrain height recovery of willow on Yellowstone’s northern range. Ecological Applications 18: 80–92.
- Bishop, C.J., White, G.C., Freddy, D.J., Watkins, B.E. & Stephenson, T.R. (2009) Effect of enhanced nutrition on mule deer population rate of change. Wildlife Monographs 172: 1–28.
- Boyd, D.K. et al. (2023) North American wolves. In T.L. Hiller et al. (eds) Wild Furbearer Management and Conservation in North America. Wildlife Ecology Institute.
- Brice, E.M., Larsen, E.J. & MacNulty, D.R. (2022) Sampling bias exaggerates a textbook example of a trophic cascade. Ecology Letters 25: 177–188.
- Brice, E.M., Larsen, E.J., Stahler, D.R. & MacNulty, D.R. (2024) The primacy of density-mediated indirect effects in a community of wolves, elk, and aspen. Ecological Monographs 95: e1627.
- Brodie, J. et al. (2013) Relative influence of human harvest, carnivores, and weather on adult female elk survival across western North America. Journal of Applied Ecology 50: 295–305.
- Carson, R. (1962) Silent Spring. Houghton Mifflin.
- Caughley, G. (1970) Eruption of ungulate populations, with emphasis on Himalayan thar in New Zealand. Ecology 51: 53–72.
- Clark, T.J. & Hebblewhite, M. (2021) Predator control may not increase ungulate populations in the future: a formal meta-analysis. Journal of Applied Ecology 58: 812–824.
- Creel, S., Christianson, D., Liley, S. & Winnie, J.A. Jr. (2007) Predation risk affects reproductive physiology and demography of elk. Science 315: 960.
- Crooks, K.R. & Soulé, M.E. (1999) Mesopredator release and avifaunal extinctions in a fragmented system. Nature 400: 563–566.
- Estes, J.A., Tinker, M.T., Williams, T.M. & Doak, D.F. (1998) Killer whale predation on sea otters linking oceanic and nearshore ecosystems. Science 282: 473–476.
- Ford, A.T. & Goheen, J.R. (2015) Trophic cascades by large carnivores: a case for strong inference and mechanism. Trends in Ecology & Evolution 30: 725–735.
- Gaillard, J.-M., Festa-Bianchet, M., Yoccoz, N., Loison, A. & Toïgo, C. (2000) Temporal variation in fitness components and population dynamics of large herbivores. Annual Review of Ecology and Systematics 31: 367–393.
- Garrott, R.A., White, P.J. & Rotella, J.J. (2008) The Madison Headwaters elk herd: transitioning from bottom–up regulation to top–down limitation. In R.A. Garrott, P.J. White & F.G.R. Watson (eds) The Ecology of Large Mammals in Central Yellowstone. Elsevier.
- Gilbert, S.L. et al. (2017) Socioeconomic benefits of large carnivore recolonization through reduced wildlife–vehicle collisions. Conservation Letters 10: 431–439.
- Griffin, K.A. et al. (2011) Neonatal mortality of elk driven by climate, predator phenology and predator community composition. Journal of Animal Ecology 80: 1246–1257.
- Haggerty, J.H. & Travis, W.R. (2006) Out of administrative control: absentee owners, resident elk and the shifting nature of wildlife management in southwestern Montana. Geoforum 37: 816–830.
- Hairston, N.G., Smith, F.E. & Slobodkin, L.B. (1960) Community structure, population control, and competition. American Naturalist 94: 421–425.
- Hobbs, N.T., Johnston, D.B., Marshall, K.N., Wolf, E.C. & Cooper, D.J. (2024) Does restoring apex predators to food webs restore ecosystems? Large carnivores in Yellowstone as a model system. Ecological Monographs 94: e1598.
- Hurley, M.A. et al. (2011) Demographic response of mule deer to experimental reduction of coyotes and mountain lions in southeastern Idaho. Wildlife Monographs 178: 1–33.
- Kauffman, M.J., Brodie, J.F. & Jules, E.S. (2010) Are wolves saving Yellowstone’s aspen? A landscape-level test of a behaviorally mediated trophic cascade. Ecology 91: 2742–2755.
- Kauffman, M.J., Varley, N., Smith, D.W., Stahler, D.R., MacNulty, D.R. & Boyce, M.S. (2007) Landscape heterogeneity shapes predation in a newly restored predator–prey system. Ecology Letters 10: 690–700.
- Klein, D.R. (1995) The introduction, increase, and demise of wolves on Coronation Island, Alaska. In L.N. Carbyn, S.H. Fritts & D.R. Seip (eds) Ecology and Conservation of Wolves in a Changing World. University of Alberta Press.
- Kohl, M.T. et al. (2018) Diel predator activity drives a dynamic landscape of fear. Ecological Monographs 88: 638–652.
- Kohl, M.T. et al. (2019) Do prey select for vacant hunting domains to minimize a multi-predator threat? Ecology Letters 22: 1724–1733.
- Lamb, C.T. et al. (2024) Effectiveness of population-based recovery actions for threatened southern mountain caribou. Ecological Applications 34: e2965.
- Leopold, A. (1949) A Sand County Almanac. Oxford University Press.
- MacNulty, D. et al. (2020) Population dynamics of northern Yellowstone elk after wolf reintroduction. In D.W. Smith, D.R. Stahler & D.R. MacNulty (eds) Yellowstone Wolves: Science and Discovery in the World’s First National Park. University of Chicago Press.
- McLaren, B.E. & Peterson, R.O. (1994) Wolves, moose, and tree rings on Isle Royale. Science 266: 1555–1558.
- Messier, F. (1994) Ungulate population models with predation: a case study with the North American moose. Ecology 75: 478–488.
- Metz, M. et al. (2020) Wolf predation on elk in a multi-prey environment. In D.W. Smith, D.R. Stahler & D.R. MacNulty (eds) Yellowstone Wolves. University of Chicago Press.
- Middleton, A.D. et al. (2013) Linking anti-predator behaviour to prey demography reveals limited risk effects of an actively hunting large carnivore. Ecology Letters 16: 1023–1030.
- National Research Council (1997) Wolves, Bears, and Their Prey in Alaska. National Academies Press.
- Painter, L.E., Beschta, R.L., Larsen, E.J. & Ripple, W.J. (2015) Recovering aspen follow changing elk dynamics in Yellowstone: evidence of a trophic cascade? Ecology 96: 252–263.
- Peterson, R.O., Vucetich, J.A., Bump, J.M. & Smith, D.W. (2014) Trophic cascades in a multicausal world: Isle Royale and Yellowstone. Annual Review of Ecology, Evolution, and Systematics 45: 325–345.
- Proffitt, K.M., Cunningham, J.A., Hamlin, K.L. & Garrott, R.A. (2014) Bottom-up and top-down influences on pregnancy rates and recruitment of northern Yellowstone elk. Journal of Wildlife Management 78: 1383–1393.
- Prugh, L.R. & Sivy, K.J. (2020) Enemies with benefits: integrating positive and negative interactions among terrestrial carnivores. Ecology Letters 23: 902–918.
- Raynor, J.L., Grainger, C.A. & Parker, D.P. (2021) Wolves make roadways safer, generating large economic returns to predator conservation. PNAS 118: e2023251118.
- Ripple, W.J. & Beschta, R.L. (2003) Wolf reintroduction, predation risk, and cottonwood recovery in Yellowstone National Park. Forest Ecology and Management 184: 299–313.
- Ripple, W.J. & Beschta, R.L. (2006) Linking wolves to willows via risk-sensitive foraging by ungulates in the northern Yellowstone ecosystem. Forest Ecology and Management 230: 96–106.
- Ripple, W.J. & Beschta, R.L. (2007) Restoring Yellowstone’s aspen with wolves. Biological Conservation 138: 514–519.
- Ripple, W.J. & Beschta, R.L. (2012) Trophic cascades in Yellowstone: the first 15 years after wolf reintroduction. Biological Conservation 145: 205–213.
- Ripple, W.J., Beschta, R.L., Wolf, C., Painter, L.E. & Wirsing, A.J. (2025) The strength of the Yellowstone trophic cascade after wolf reintroduction. Global Ecology and Conservation 58: e03428.
- Ripple, W.J., Larsen, E.J., Renkin, R.A. & Smith, D.W. (2001) Trophic cascades among wolves, elk and aspen on Yellowstone National Park’s northern range. Biological Conservation 102: 227–234.
- Roffler, G.H., Eriksson, C.E., Allen, J.M. & Levi, T. (2023) Recovery of a marine keystone predator transforms terrestrial predator–prey dynamics. PNAS 120: e2209037120.
- Schmitz, O.J., Beckerman, A.P. & O’Brien, K.M. (1997) Behaviorally mediated trophic cascades: effects of predation risk on food web interactions. Ecology 78: 1388–1399.
- Serrouya, R. et al. (2021) Trophic consequences of terrestrial eutrophication for a threatened ungulate. Proceedings of the Royal Society B 288: 20202811.
- Sivy, K.J., Pozzanghera, C.B., Grace, J.B. & Prugh, L.R. (2017) Fatal attraction? Intraguild facilitation and suppression among predators. American Naturalist 190: 663–679.
- Terborgh, J. et al. (2001) Ecological meltdown in predator-free forest fragments. Science 294: 1923–1926.
- Vucetich, J.A., Smith, D.W. & Stahler, D.R. (2005) Influence of harvest, climate and wolf predation on Yellowstone elk, 1961–2004. Oikos 111: 259–270.
- White, P.J. & Garrott, R.A. (2013) Predation: wolf restoration and the transition of Yellowstone elk. In P.J. White, R.A. Garrott & G.E. Plumb (eds) Yellowstone’s Wildlife in Transition. Harvard University Press.
- White, P.J., Garrott, R.A., Hamlin, K.L., Cook, R.C., Cook, J.G. & Cunningham, J.A. (2011) Body condition and pregnancy in northern Yellowstone elk: evidence for predation risk effects? Ecological Applications 21: 3–8.
- White, P.J., Stahler, D.R., Smith, D.W., MacNulty, D.R. & Garrott, R.A. (2024) Northern Yellowstone Elk Resilience and Adaptation to Changes in Management Policies and the Ecosystem. National Park Service.
- White, P.J., Wallen, R.L. & Hallac, D. (2015) Yellowstone Bison: Conserving an American Icon in Modern Society. Yellowstone Association.
- Wilmers, C.C., Levi, T., Prugh, L.R., Ruprecht, J. & Stahler, D.R. (2025) The ecological impacts of large-carnivore recovery in North America. Annual Review of Ecology, Evolution, and Systematics 56: 337–363.
- Wilmers, C.C., Post, E., Peterson, R.O. & Vucetich, J.A. (2006) Predator disease out-break modulates top-down, bottom-up, and climatic effects on herbivore population dynamics. Ecology Letters 9: 383–389.
Philosophy, methods, and statistics
- Anderson, D.R., Burnham, K.P. & Thompson, W.L. (2000) Null hypothesis testing: problems, prevalence, and an alternative. Journal of Wildlife Management 64: 912–923.
- Ayer, A.J. (1936) Language, Truth and Logic. Gollancz.
- Berger, P.L. & Luckmann, T. (1966) The Social Construction of Reality. Doubleday.
- Beven, K. & Freer, J. (2001) Equifinality, data assimilation, and uncertainty estimation in mechanistic modelling of complex environmental systems using the GLUE methodology. Journal of Hydrology 249: 11–29.
- Bhaskar, R. (1975) A Realist Theory of Science. Leeds Books.
- Bogen, J. & Woodward, J. (1988) Saving the phenomena. Philosophical Review 97: 303–352.
- Box, G.E.P. (1976) Science and statistics. Journal of the American Statistical Association 71: 791–799.
- Bromberger, S. (1966) Why-questions. In R.G. Colodny (ed.) Mind and Cosmos, 86–111. University of Pittsburgh Press.
- Burnham, K.P. & Anderson, D.R. (2002) Model Selection and Multimodel Inference, 2nd edn. Springer.
- Chamberlin, T.C. (1890) The method of multiple working hypotheses. Science 15: 92–96.
- Comte, A. (1830–1842) Cours de philosophie positive, 6 vols. Bachelier.
- Dilthey, W. (1883) Einleitung in die Geisteswissenschaften (English 1989, Introduction to the Human Sciences, Princeton University Press).
- Douglas, H. (2000) Inductive risk and values in science. Philosophy of Science 67: 559–579.
- Douglas, H. (2004) The irreducible complexity of objectivity. Synthese 138: 453–473.
- Douglas, H. (2009) Science, Policy, and the Value-Free Ideal. University of Pittsburgh Press.
- Duhem, P. (1906) La théorie physique: son objet et sa structure (English 1954, The Aim and Structure of Physical Theory, Princeton University Press).
- Feyerabend, P. (1975) Against Method. New Left Books.
- Fisher, R.A. (1935) The Design of Experiments. Oliver & Boyd.
- Fraser, H., Parker, T., Nakagawa, S., Barnett, A. & Fidler, F. (2018) Questionable research practices in ecology and evolution. PLoS ONE 13: e0200303.
- Gelman, A. & Loken, E. (2014) The statistical crisis in science. American Scientist 102: 460–465.
- Gettier, E.L. (1963) Is justified true belief knowledge? Analysis 23: 121–123.
- Guba, E.G. & Lincoln, Y.S. (1994) Competing paradigms in qualitative research. In N.K. Denzin & Y.S. Lincoln (eds) Handbook of Qualitative Research, 105–117. Sage.
- Guba, E.G. & Lincoln, Y.S. (2005) Paradigmatic controversies, contradictions, and emerging confluences. In N.K. Denzin & Y.S. Lincoln (eds) The SAGE Handbook of Qualitative Research, 3rd edn, 191–215. Sage.
- Hacking, I. (1983) Representing and Intervening. Cambridge University Press.
- Hanson, N.R. (1958) Patterns of Discovery. Cambridge University Press.
- Harding, S. (1991) Whose Science? Whose Knowledge? Cornell University Press.
- Hempel, C.G. & Oppenheim, P. (1948) Studies in the logic of explanation. Philosophy of Science 15: 135–175.
- Hill, A.B. (1965) The environment and disease: association or causation? Proceedings of the Royal Society of Medicine 58: 295–300.
- Holland, P.W. (1986) Statistics and causal inference. Journal of the American Statistical Association 81: 945–960.
- Hoyningen-Huene, P. (2013) Systematicity: The Nature of Science. Oxford University Press.
- Hume, D. (1748) An Enquiry Concerning Human Understanding.
- Hurlbert, S.H. (1984) Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54: 187–211.
- Ioannidis, J.P.A. (2005) Why most published research findings are false. PLoS Medicine 2: e124.
- Johnson, D.H. (1999) The insignificance of statistical significance testing. Journal of Wildlife Management 63: 763–772.
- Kerr, N.L. (1998) HARKing: hypothesizing after the results are known. Personality and Social Psychology Review 2: 196–217.
- Kitcher, P. (1989) Explanatory unification and the causal structure of the world. In P. Kitcher & W. Salmon (eds) Scientific Explanation, 410–505. University of Minnesota Press.
- Kuhn, T.S. (1962) The Structure of Scientific Revolutions. University of Chicago Press.
- Kuhn, T.S. (1970) The Structure of Scientific Revolutions, 2nd edn, with Postscript–1969. University of Chicago Press.
- Kuhn, T.S. (1977) Objectivity, value judgment, and theory choice. In The Essential Tension, 320–339. University of Chicago Press.
- Lakatos, I. (1970) Falsification and the methodology of scientific research programmes. In I. Lakatos & A. Musgrave (eds) Criticism and the Growth of Knowledge, 91–196. Cambridge University Press.
- Laudan, L. (1977) Progress and Its Problems. University of California Press.
- Laudan, L. (1981) A confutation of convergent realism. Philosophy of Science 48: 19–49.
- Laudan, L. (1983) The demise of the demarcation problem. In R.S. Cohen & L. Laudan (eds) Physics, Philosophy and Psychoanalysis, 111–127. Reidel.
- Levin, S.A. (1992) The problem of pattern and scale in ecology. Ecology 73: 1943–1967.
- Levins, R. (1966) The strategy of model building in population biology. American Scientist 54: 421–431.
- Lewis, D. (1973) Causation. Journal of Philosophy 70: 556–567.
- Lincoln, Y.S. & Guba, E.G. (1985) Naturalistic Inquiry. Sage.
- Lipton, P. (2004) Inference to the Best Explanation, 2nd edn. Routledge.
- Longino, H.E. (1990) Science as Social Knowledge. Princeton University Press.
- MacKenzie, D.I. et al. (2002) Estimating site occupancy rates when detection probabilities are less than one. Ecology 83: 2248–2255.
- Mayo, D.G. (2018) Statistical Inference as Severe Testing. Cambridge University Press.
- Merton, R.K. (1942) The normative structure of science. Reprinted in The Sociology of Science (1973), 267–278. University of Chicago Press.
- Morgan, M.S. & Morrison, M. (eds) (1999) Models as Mediators. Cambridge University Press.
- National Academies of Sciences, Engineering, and Medicine (2019) Reproducibility and Replicability in Science. National Academies Press.
- Nosek, B.A. et al. (2018) The preregistration revolution. PNAS 115: 2600–2606.
- Open Science Collaboration (2015) Estimating the reproducibility of psychological science. Science 349: aac4716.
- Pearl, J. (2009) Causality, 2nd edn. Cambridge University Press.
- Pearl, J. & Mackenzie, D. (2018) The Book of Why. Basic Books.
- Peirce, C.S. (1877) The fixation of belief. Popular Science Monthly 12: 1–15.
- Peirce, C.S. (1878) Deduction, induction, and hypothesis. Popular Science Monthly 13: 470–482.
- Pigliucci, M. & Boudry, M. (eds) (2013) Philosophy of Pseudoscience: Reconsidering the Demarcation Problem. University of Chicago Press.
- Platt, J.R. (1964) Strong inference. Science 146: 347–353.
- Popper, K.R. (1959) The Logic of Scientific Discovery (German original 1934). Hutchinson.
- Popper, K.R. (1963) Conjectures and Refutations. Routledge.
- Putnam, H. (1975) Mathematics, Matter and Method. Cambridge University Press.
- Quine, W.V.O. (1951) Two dogmas of empiricism. Philosophical Review 60: 20–43.
- Quine, W.V.O. (1969) Epistemology naturalized. In Ontological Relativity and Other Essays. Columbia University Press.
- Ramsey, F.P. (1926) Truth and probability. In The Foundations of Mathematics and Other Logical Essays (1931). Routledge.
- Reichenbach, H. (1938) Experience and Prediction. University of Chicago Press.
- Royall, R. (1997) Statistical Evidence: A Likelihood Paradigm. Chapman & Hall.
- Rubin, D.B. (1974) Estimating causal effects of treatments in randomized and nonrandomized studies. Journal of Educational Psychology 66: 688–701.
- Rudner, R. (1953) The scientist qua scientist makes value judgments. Philosophy of Science 20: 1–6.
- Salmon, W.C. (1984) Scientific Explanation and the Causal Structure of the World. Princeton University Press.
- Silberzahn, R. et al. (2018) Many analysts, one data set: making transparent how variations in analytic choices affect results. Advances in Methods and Practices in Psychological Science 1: 337–356.
- Soulé, M.E. (1985) What is conservation biology? BioScience 35: 727–734.
- Stewart-Oaten, A., Murdoch, W.W. & Parker, K.R. (1986) Environmental impact assessment: “pseudoreplication” in time? Ecology 67: 929–940.
- Taylor, C. (1971) Interpretation and the sciences of man. Review of Metaphysics 25: 3–51.
- Thagard, P.R. (1978) Why astrology is a pseudoscience. PSA 1978, vol. 1: 223–234.
- Toulmin, S. (1958) The Uses of Argument. Cambridge University Press.
- van Fraassen, B.C. (1980) The Scientific Image. Oxford University Press.
- Weber, M. (1922) Wirtschaft und Gesellschaft (English 1978, Economy and Society, University of California Press).
- Weisberg, M. (2007) Three kinds of idealization. Journal of Philosophy 104: 639–659.
- Woodward, J. (2003) Making Things Happen: A Theory of Causal Explanation. Oxford University Press.
- Worrall, J. (1989) Structural realism: the best of both worlds? Dialectica 43: 99–124.
- Wu, J. & Loucks, O.L. (1995) From balance of nature to hierarchical patch dynamics: a paradigm shift in ecology. Quarterly Review of Biology 70: 439–466.
- Yates, K.L. et al. (2018) Outstanding challenges in the transferability of ecological models. Trends in Ecology & Evolution 33: 790–802.
Legal cases
- Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993).
- Kitzmiller v. Dover Area School District, 400 F. Supp. 2d 707 (M.D. Pa. 2005).
- McLean v. Arkansas Board of Education, 529 F. Supp. 1255 (E.D. Ark. 1982).