1.6 Hypotheses, Assumptions, Data & Conclusions: Relevance and Source Reliability
Key Takeaways
- Data are recorded measurements, a hypothesis is a testable prediction stated before testing, an assumption is taken as true without being tested in the study, and a conclusion is the answer the investigator claims the data support.
- Assumptions are usually unstated; surface them by asking what must be true beyond the actual measurements for the conclusion to follow.
- Information is relevant only if it could change the answer, which is why a true-but-irrelevant statement is the most heavily used distractor type on the subtest.
- Source reliability depends on who benefits from the finding, whether independent reviewers checked it, whether the methods can be repeated, and whether the sample is representative.
- Absolute language such as 'proves' or 'guarantees' signals lower reliability, because careful scientific reporting states ranges, uncertainty, and limitations.
Sorting the Sentences in a Science Passage
The HiSET Science framework asks you to distinguish among hypotheses, assumptions, data, and conclusions, to judge whether a piece of information is even relevant to the question asked, and to judge the reliability of a source. These three skills share one habit: before you decide whether a statement is true, decide what kind of statement it is.
| Type of statement | What it is | How to recognize it | Directly measurable? |
|---|---|---|---|
| Observation / data | A recorded measurement or direct sensory record | Contains a number, a unit, or a described observation | Yes |
| Hypothesis | A proposed, testable explanation stated before testing | "We predicted...", "If... then..." | No, but it is testable |
| Assumption | Something taken as true without being tested in this study, and required for the conclusion to hold | "Assuming...", or left unstated entirely | Not in this study |
| Inference | An interpretation that goes beyond what was directly observed | "This suggests...", "The pattern indicates..." | No |
| Conclusion | The answer the investigator says the data support | "Therefore...", "These results show..." | No |
Worked passage. Classify each numbered sentence.
- "Coastal marsh grass has declined across the bay over the past decade." — Observation.
- "We proposed that rising water salinity is responsible for the decline." — Hypothesis.
- "Salinity at the five sampling stations averaged 31 parts per thousand in 2016 and 38 parts per thousand in 2026." — Data.
- "The five stations represent conditions across the whole bay." — Assumption. Nothing in the study tested whether five stations are representative, yet the conclusion depends on it.
- "Rising salinity is driving the loss of marsh grass in this bay." — Conclusion.
Assumptions are the hardest of the five because they are frequently unstated. The reliable way to surface one is to ask: what would have to be true, beyond what was actually measured, for this conclusion to follow? In the passage above, the conclusion also quietly assumes that no other factor changed over the same decade — a second untested assumption.
Determining Relevance
A piece of information is relevant to a question only if it could change the answer. This sounds obvious, but it is the basis of the most heavily used distractor on the entire subtest: the true-but-irrelevant option. Such an option states a perfectly correct piece of science that has nothing to do with the question being asked, and it is attractive precisely because you cannot find anything wrong with it.
Worked example. Question: Which additional measurement would best strengthen the conclusion that rising salinity caused the marsh grass decline?
| Candidate information | Relevant? |
|---|---|
| Salinity records from a nearby bay where marsh grass did not decline | Yes. A comparison site can rule out a bay-wide coincidence |
| Greenhouse data on marsh grass survival across a controlled salinity range | Yes. Directly tests the proposed causal mechanism |
| The chemical formula of sodium chloride | No. True, but it cannot change the answer |
| The average tidal range of the Atlantic coastline | No. True, but not tied to this bay's decline |
Apply a single filter to every answer choice: "If this were false instead of true, would my answer change?" If the answer is no, the option is irrelevant regardless of its accuracy.
Judging the Reliability of a Source
Scientific information reaches the public through very different channels, and the HiSET framework expects you to rank them. Reliability is not about whether a source sounds authoritative; it is about whether its claims can be checked and whether anyone benefits from the result.
- Who produced it, and do they gain from the finding? A supplement manufacturer reporting that its own supplement works has a direct financial interest in the outcome. Disclosed funding does not automatically invalidate a study, but undisclosed interest is a serious warning sign.
- Was it peer reviewed? Publication in a scientific journal means independent specialists scrutinized the methods before release. A press release, an advertisement, and a social media post have passed no such review.
- Are the methods described in enough detail to be repeated? A source that will not say how many subjects were tested, or how the measurement was made, cannot be checked by anyone.
- Is the sample large and representative? A conclusion about a national population drawn from eight volunteers recruited at one clinic cannot support that scope.
- Is it primary or secondary? A primary source reports original data; a secondary source summarizes someone else's work and can introduce error. Prefer the primary source when the exact numbers matter.
- Is it current? In fast-moving areas, a twenty-year-old figure may have been superseded.
- Does it report uncertainty? Careful science states ranges, error, and limitations. Absolute language such as "proves" and "guarantees" is a reliability warning, not a strength.
A Working Ranking
From most to least reliable for a scientific claim:
- Peer-reviewed journal article reporting original data, with methods and limitations stated
- Government or scientific agency report drawing on published research
- Current science textbook or reference work
- Reputable news coverage that names and links its underlying study
- Material published by a company about its own product
- Anonymous blog posts, testimonials, and individual anecdotes
Note the two independent axes at work: a news article accurately summarizing a strong study may be more useful than a poorly designed journal article, so weigh the quality of the underlying evidence, not the format alone.
Exam Strategy
- Label statements before evaluating them. Many items are answered purely by recognizing that a sentence is an assumption rather than data.
- When a question asks what the study assumed, look for what the conclusion needs that the measurements never supplied.
- Test every answer choice for relevance with the "would my answer change?" filter; discard true statements that fail it.
- When ranking sources, ask first who benefits from the claim and second whether anyone independent checked it.
A report on a fishery reads: 'Cod landings in the gulf fell from 14,200 tonnes in 2016 to 3,900 tonnes in 2026. Water temperature at the three monitoring buoys rose by 1.4 degrees Celsius over the same period. The three buoys reflect conditions throughout the gulf. Warming water is therefore responsible for the collapse of the gulf cod fishery.' Which sentence is an assumption rather than data or a conclusion?
A student concludes that a new bicycle chain lubricant reduces rolling resistance after timing one bicycle coasting down a hill twice, once with each lubricant. Which additional piece of information would be most relevant to evaluating this conclusion?
Four sources make claims about whether a dietary supplement improves memory. Which source provides the most reliable evidence, and what feature makes it so?