DoAssignment study guide

A1.8 · Test predictions against data and identify uncertainty or bias

Learn to test predictions against data and identify uncertainty or bias through clear examples and targeted practice.

Ontario Grade 10 Science

Scientific Investigation and Careers

Comparing observations and noticing uncertainty or bias

Suppose you predict that a plant near a window will grow taller than a similar plant farther away. The prediction is not yet evidence. You need observations, such as recorded heights, to check it. In this lesson, you will compare predictions with data and notice limits in the evidence. Examples with invented values are labelled hypothetical. They are not reported experimental results.

What you will learn

Explain the difference between a prediction and data.
Compare observations with a prediction and decide whether the data support it.
Identify uncertainty and possible bias in a simple investigation.
Describe how uncertainty or bias affects the strength of a conclusion.

1. From observations to a testable prediction

In earlier science learning, you used observations to describe what happened and evidence to support an explanation. A prediction is a statement about what you expect to happen. Data are the observations or measurements collected during an investigation. A prediction comes before the test; data come from the test.

A useful prediction is specific enough to compare with observations. “The plant will do better” is unclear because “better” could mean taller, greener, or having more leaves. “After two weeks, the plant receiving more light will be taller” names a result and a time. The prediction might still be wrong. Its purpose is to guide a test, not to guarantee the outcome.

To test a prediction, identify what it says should happen. Collect relevant data, check that the records are clear, and compare the observed pattern with the predicted pattern. Data support a prediction when the observations fit it. This means the observations from this test agree with the prediction; it does not prove that the same result will occur in every situation.

A prediction states what you expect; data record what you observe.
Make the expected result clear enough to compare with the data.
Evidence can support a prediction without proving it in every case.

2. Compare the data with the predicted pattern

Data may be written as words, counts, or measurements. A table can make a comparison easier by placing a prediction beside the observations. Look at all the relevant data. Do not choose only the observation that best agrees with what you expected.

Imagine predicting that objects exposed to more light will show a larger change than objects exposed to less light. If the recorded changes mostly follow that pattern, the data may support the prediction. If the results are mixed, the evidence may be unclear. If the pattern goes the other way, the data do not support the prediction in this test.

When comparing data, describe what the observations show in plain language. You do not need an extra calculation if the pattern is already clear. If different observations point in different directions, explain that rather than hiding the disagreement.

Compare the data with the exact result your prediction described.
Consider all relevant observations, not only a convenient result.
Mixed results can make a conclusion uncertain.

3. Notice uncertainty and possible bias

Uncertainty is doubt about how exact or reliable an observation is. A ruler reading may be uncertain if a mark lies between two scale lines. A person may also record a changing event at slightly different moments each time. Uncertainty does not make every observation useless. It means the result should not be presented as more exact than the method allows.

Bias is an influence that unfairly pushes observations or choices in one direction. For example, someone who expects one group to do better might judge its results more generously or record its measurements differently. Bias can also come from setting up a comparison so that one group gets an advantage unrelated to the question.

When reviewing data, ask how sure you are about the observations and whether the way the test was done or judged could have favoured one result. These questions are not accusations. They help explain how much confidence the evidence deserves. Name a possible source of uncertainty or bias and explain how it could affect the comparison.

Clear records and consistent methods help reduce avoidable problems. Use the same measuring approach for each group, record observations as they happen, and apply the same rule throughout the test. These choices cannot make every observation perfectly certain, but they can make the comparison fairer and easier to interpret.

Uncertainty is doubt about the exactness or reliability of observations.
Bias can unfairly favour a result or affect how evidence is collected or judged.
Name the possible problem and explain how it could affect the conclusion.

4. Make a conclusion that fits the evidence

A conclusion should answer the question and match the strength of the data. If the observations fit the prediction, say that they support it in this test. If they do not fit, say that the prediction was not supported by these data. If uncertainty is large or the pattern is mixed, say that the result is unclear. Do not change the prediction after seeing the data just to make it seem correct.

A conclusion is more convincing when observations are relevant, recorded consistently, and not clearly affected by bias. Even then, keep the claim within the limits of the test. A small set of observations from one situation does not automatically describe every situation. Report what the evidence says without claiming more than it can show.

Science investigations must also be safe. Follow your teacher’s instructions and use only approved materials and equipment. Do not carry out chemical or other potentially hazardous tests at home. Safe planning allows observations to be collected responsibly.

State whether the data support, do not support, or leave the prediction unclear.
Mention important uncertainty or bias when explaining the result.
Keep the conclusion limited to the evidence collected and follow safety directions.

A simple way to compare a prediction with data

Question to ask	What to look for
What did the prediction expect?	A clear result or pattern stated before the test.
What do the data show?	The recorded observations, including the full set rather than selected results.
How well do they match?	A matching, opposite, mixed, or unclear pattern.
Could evidence be limited?	Uncertainty in observations or possible bias in collecting or judging them.
What can be concluded?	A statement that fits the data and notes important limits.

Worked example

A prediction about reading time

Hypothetical classroom data: A student predicts that a quiet room will help classmates finish a short reading task faster. In a quiet room, three students take 8, 9, and 10 minutes. In a room with conversation, three students take 9, 10, and 11 minutes. Do the data support the prediction, and what should the student avoid claiming?

State the predicted pattern
The prediction expects shorter completion times in the quiet room. That gives us a direct way to compare the two sets of observations.
Compare the full sets
The quiet-room times range from 8 to 10 minutes. The conversation-room times range from 9 to 11 minutes. The ranges overlap, so it would be inaccurate to say every quiet-room time is lower than every conversation-room time. When the times are compared in order, each quiet-room time is one minute shorter. This pattern supports the prediction in this small hypothetical test.
Limit the conclusion
There are only three students in each group, and we do not know whether the groups were otherwise similar. The student should say the data support the prediction in this test, not that quiet rooms always help everyone finish faster.

Answer: The hypothetical data support the prediction in this test, although the ranges overlap. The small groups and any differences between them make a broader claim uncertain.

Check: The conclusion describes the overall pattern without claiming that quiet-room times are lower for everyone in all situations.

Worked example

A result close to the measurement limit

Hypothetical data: A learner predicts that Brand B of paper towel absorbs more water than Brand A. The recorded amounts are 20 mL for Brand A and 21 mL for Brand B. The measuring cup has marks every 10 mL, and the learner estimated amounts between marks. How should the learner report the result?

Compare the recorded values
Brand B has a recorded amount 1 mL higher than Brand A. Taken at face value, this small difference points in the predicted direction.
$21\,\mathrm{mL}-20\,\mathrm{mL}=1\,\mathrm{mL}$
Consider uncertainty
The cup marks are 10 mL apart. Estimates of 20 mL and 21 mL may not be precise enough to distinguish such a small difference. The apparent difference could be smaller than the uncertainty in the readings.
Choose a cautious conclusion
The learner should not claim that Brand B clearly absorbs more based on these readings alone. The result is unclear because the measuring method may not distinguish the small difference. A more suitable measuring method, used safely and consistently, could help test the prediction more clearly.

Answer: The recorded values differ by only 1 mL, but the measuring marks are 10 mL apart. The data do not clearly establish that one brand absorbs more.

Check: The main issue is uncertainty in the measurements, not proof that the brands are identical.

Worked example

Spotting possible bias in observations

Hypothetical investigation: Two groups compare how many seeds sprout under two conditions. The person counting knows which group was expected to do better and decides whether a partly opened seed counts as sprouted. What concern should be identified, and how could it affect the conclusion?

Identify the concern
The person counting knows which group is expected to do better, and the rule for a partly opened seed is unclear. This creates a risk of bias because expectations could affect which seeds are counted.
Explain the possible effect
If borderline seeds in the expected-to-do-better group are more often counted as sprouted, its total may appear higher even if the groups are similar. The observations may then favour the prediction unfairly.
Improve the comparison
Set a clear counting rule before observing the groups and apply the same rule to both. If practical, ask someone who does not know the prediction to count. These steps reduce the chance that expectations shape the recorded data.

Answer: The counting method may be biased because the observer knows the expected result and has no clear rule for borderline cases. Use the same stated rule for both groups and avoid letting the prediction guide the count.

Check: Bias is a possible influence on the observations; it does not by itself prove that the final counts are wrong.

Common mistakes and how to avoid them

Saying a prediction is proven because one observation agrees with it.

Correction: Compare all relevant data. Say that the evidence supports the prediction when the pattern agrees, and keep the conclusion within the test’s limits.

Treating an uncertain measurement as exact.

Correction: Describe how the measuring method limits confidence. Do not claim a clear difference when the observations cannot reliably distinguish it.

Calling every unexpected result biased.

Correction: Bias is a possible unfair influence on how data are collected or judged. An unexpected result may simply mean the prediction was not supported.

Ignoring observations that do not match the prediction.

Correction: Consider all recorded observations. Selecting only results that agree can create a misleading comparison.

Lesson summary

A prediction states an expected result; data are observations collected during a test.
Compare all relevant data with the predicted pattern.
Uncertainty concerns the exactness or reliability of observations. Bias can unfairly influence data collection or judgment.
Conclusions should state whether the data support the prediction and acknowledge important limits.

Check your understanding

Question 1

A prediction says that a shaded area will be cooler than a sunny area. A hypothetical set of temperature readings shows the shaded area cooler each time. Which statement is best?

The data support the prediction in this test.
The prediction is proven for every place and every day.
The readings must be biased because they match the prediction.
The data cannot be used because they are measurements.

Show answer and explanation

The data support the prediction in this test.

The recorded pattern agrees with the prediction, so it supports the prediction in this test. It does not prove the result applies everywhere.

Question 2

A measurement tool has marks 5 units apart. Two results differ by 1 unit. What is the main concern?

The prediction has become data.
The difference may be too small to distinguish reliably with that tool.
The results prove that the two objects are exactly the same.
The observer has definitely changed the results on purpose.

Show answer and explanation

The difference may be too small to distinguish reliably with that tool.

The measurement marks are farther apart than the reported difference. This creates uncertainty about whether the small difference is reliable.

Question 3

A person who expects one group to do better uses a vague rule when deciding which observations to count. What should be considered?

Possible bias in how the observations were judged.
Proof that the prediction is correct.
A reason to ignore all data from both groups.
A guarantee that the groups had equal results.

Show answer and explanation

Possible bias in how the observations were judged.

Knowing the expected result and using an unclear rule could affect which observations are counted. A clear rule applied consistently can reduce this concern.

Key terms

Prediction: A statement about what you expect to happen in a test.
Data: Observations or measurements recorded during an investigation.
Evidence: Information from observations or data used to support or challenge an idea.
Uncertainty: Doubt about how exact or reliable an observation or measurement is.
Bias: An influence that unfairly favours a result or affects how evidence is collected or judged.

Continue through SNC2D

View the complete SNC2D Ontario Grade 10 Science curriculum and lessons

About this lesson and its review

Published by DoAssignment. This reviewed lesson follows Ontario Grade 10 Science (SNC2D), expectation A1.8. It is a study resource, not an official curriculum publication.

Before publication, content is checked for structure, mathematical or chemical notation, calculations, course boundaries, and readability. Errors can still occur, so corrections are welcomed.

Official curriculum reference

Report a correction or ask a question