Science: Evaluating Models and Viewpoints

This is Unit Twenty-One of ACT Preparation: evaluating models and viewpoints. Unit twenty judged how data were produced; this unit judges the explanations built on them, and what happens when two explanations disagree. One rule governs it: evidence decides between explanations only when they predicted it differently. By the end you will be able to tell a hypothesis from a theory. To name inductive and deductive reasoning. To choose the finding that supports or weakens a claim. To decide whether data can show cause. And to compare conflicting viewpoints and judge which a new finding favors.

Seven entries on how science explains: method, theory, models and reasoning.

No single method.

ACT 2001 No Single Method

No single scientific method fits every study; what science requires is skepticism, a tentative explanation, then tests of its predictions. A student who suspects a plant leans toward light can turn the pot and watch. The steps from unit twenty describe a common pattern, not a rule.

Hypothesis and theory.

ACT 2002 Hypothesis and Theory

A hypothesis is a tentative explanation awaiting tests; a theory is a tested and confirmed explanation for observations. A new fertilizer's benefit is a hypothesis until tested; an explanation confirmed across many studies is a theory. Everyday speech uses theory to mean a guess; science does not.

Models.

ACT 2003 Models

A scientific model is a simplified representation built to explain observations and make predictions. No model is perfect, but an imperfect one can be useful: a lake model tracking only temperature and oxygen can still predict when fish are stressed. Judge a model by what it explains and predicts.

From many cases to a rule.

ACT 2004 Inductive Reasoning

Inductive reasoning moves from many related observations to a general conclusion, common in descriptive science. Every one of 200 tagged turtles returning to the same beach supports a conclusion about that population. One new observation that does not fit can overturn it.

From a rule to a case.

ACT 2005 Deductive Reasoning

Deductive reasoning moves from a general principle to a specific prediction, which holds as long as the principle holds; it is the logic of hypothesis-based science. If plants need light to make food, a plant kept in darkness should weaken. Deduction runs general to specific; induction, specific to general.

Two pathways of science.

ACT 2006 Descriptive and Hypothesis-Based Science

Descriptive science observes and discovers, usually by induction; hypothesis-based science tests a proposed answer, usually by deduction. Cataloguing a meadow's insects is descriptive; testing whether mowing reduces them is hypothesis-based. Most research combines both, and the boundary is blurred.

Inductive or deductive, step by step.

ACT 2007 Inductive or Deductive?

Find where the reasoning starts, many observations or a general principle, and where it ends, a general conclusion or a specific prediction. An invented researcher who finds frogs only in ponds with cattails across sixty ponds, and concludes the frogs need cattails, reasons inductively. A second who reasons that a new cattail pond should attract them reasons deductively. One study can use both; name the step the question points to.

Seven entries on evidence: what supports a claim, what weakens it, and when it shows cause.

Evidence that supports.

ACT 2008 Evidence That Supports a Claim

Evidence supports a claim when it matches what the claim predicts, especially when the alternatives would not. If a model predicts fewer fish in warmer water, a survey finding fewer at warm sites supports it. Evidence every explanation predicts supports none of them over the others.

Evidence that weakens.

ACT 2009 Evidence That Weakens a Claim

Evidence weakens a claim when it contradicts what the claim predicts, and a single well-measured contradiction can outweigh several results that merely fit. Healthy fish in the warmest pond weaken the claim that warmth alone harms them. Check that it was measured under the conditions the claim describes.

Claims that can be tested.

ACT 2010 Testable Claims

A scientific claim must be testable: some practical observation could show it false. "Unseen beings move the stones and leave no trace" cannot be tested; "frost cracks the stones" can. An untestable claim is not shown true by the lack of evidence against it.

Which finding supports, step by step.

ACT 2011 Which Finding Supports the Claim?

State the claim and its prediction, check each finding against it, and prefer the one unlikely if the claim were false. For an invented claim that a bird nests on cliffs to avoid ground predators, its cliff nesting is the observation the claim explains, not new support. Ground nesting on an island with no ground predators fits the claim's reason and is hard to explain otherwise, so it supports best. A claim cannot be tested by what it was built to explain.

Which finding weakens, step by step.

ACT 2012 Which Finding Would Weaken It?

State the claim's prediction, imagine a result that would be surprising if it were true, and choose the finding that matches, under the same conditions. Cliff nesting even on predator-free islands weakens the cliff claim, because the reason is gone and the behavior remains. A finding about a different species does not. Weakening is not disproving; one contrary finding raises doubt.

Association and cause.

ACT 2013 Association and Cause in Science

Two measurements that rise and fall together are associated, but that alone does not show cause; testing cause takes a controlled experiment, ideally randomized. In invented data, sites with more moss also had more snails, and both may follow the shade. Ask what third factor could drive both.

Can the data show cause, step by step.

ACT 2014 Can the Data Show Cause?

Decide whether the data were observed or produced by assigning conditions, name a third factor for observational data, and choose the conclusion the design allows. An invented survey finds gardens with more bees grow more squash, but sunnier gardens might attract bees and grow squash; the survey supports association. Netting some plants against bees at random could test the cause. "Causes" must be earned by the design.

Eight entries on conflicting viewpoints, ending with a full passage.

Two explanations that disagree.

ACT 2015 Conflicting Viewpoints

A conflicting viewpoints passage presents two or more explanations of one phenomenon that disagree, resting on different premises or incomplete data: two invented scientists blame a lake's fish deaths on low oxygen and on a heat spike. Neither view is labeled correct; the questions test understanding, not picking a winner.

Shared ground.

ACT 2016 Shared Ground

Competing viewpoints usually share some facts or assumptions, and finding them narrows the dispute: both scientists agree the fish died within one week in August, in unusually warm water. A question asking what both accept wants the shared fact.

What each view explains.

ACT 2017 What Each Viewpoint Explains

Each viewpoint explains some observations and may struggle with others. Low oxygen explains why bottom-feeders died first; a heat spike explains why shallow-water fish died too. A view is weakened only by data it predicts wrongly, not by data it never claimed to explain.

Summarizing each view, step by step.

ACT 2018 Summarizing Each Viewpoint

State each view's claim in a sentence, list its evidence, and note what it predicts that the other does not; taking these notes first, as the ACT's own guide suggests, speeds the comparison questions. Scientist one blames an algae bloom's decay for low oxygen, citing a thick bloom, and predicts very low bottom oxygen. Scientist two blames heat directly, citing a record-warm week, and predicts that fish in deep, cool water survived. Keep each summary in its own terms.

Agreement and difference, step by step.

ACT 2019 Where the Viewpoints Agree and Differ

Place the summaries side by side, mark what both make or assume, and mark what each would deny in the other. Both scientists accept the August die-off and the warmth; they differ on the mechanism, whether oxygen loss was the step between warmth and death. Viewpoints can agree on a cause and differ on how it acts.

When new evidence decides.

ACT 2020 New Evidence Decides

New evidence favors one viewpoint when one predicted it and the other did not, or when it contradicts one but not the other. Healthy fish in deep, cool water fit the heat view; low oxygen near the bottom fits the oxygen view. Evidence both predict decides nothing.

Judging a new finding, step by step.

ACT 2021 Which Viewpoint a New Finding Supports

State what each view predicts about the finding's situation, compare, and choose the view whose prediction it matches and the other's it contradicts. An invented finding of normal bottom oxygen during the die-off contradicts scientist one's prediction and leaves scientist two's standing, so it favors scientist two; a very warm week, which both accept, favors neither. Ask what each view predicted before the finding, not what it could be stretched to explain afterward.

The whole unit in one passage.

ACT 2022 A Conflicting Viewpoints Passage

Evidence decides between explanations only when they predicted it differently. In an invented passage, a ridge's north side keeps its snow three weeks longer. Student one says it receives less direct sunlight; student two says thicker trees shade it. Both accept that less melting energy reaches the north side; they differ on whether the cause is the slope's angle or the trees. A bare, treeless clearing on the north side that also keeps its snow favors student one, whose view predicts it, while student two's does not. Colder night air on the north side decides nothing, because both views can accept it. Summarize each view in its own terms, find the shared ground first, and let a finding decide only when the views predicted it differently.

Look at what you can now do. You can tell a tentative hypothesis from a confirmed theory, and judge a model by what it predicts. You can name reasoning that runs from cases to a rule, inductive, or from a rule to a case, deductive. You can choose the finding that supports or weakens a claim, under the claim's own conditions. You can say whether a design earns the word cause, or only association. And you can compare conflicting viewpoints and judge what a new finding favors. Unit twenty-two fills in the background science that passages assume.