Last year the question was whether the evidence distinguishes between the explanations on the table. That question has not gone away. But chemistry hands you numbers, and a number brings a second question with it that a Grade 10 argument could usually dodge: is the difference you are arguing about bigger than the uncertainty in the measurement?
Until you can answer that, you do not yet know whether you disagree.
The question
Two groups ran the same test in Building the Activity Series and came away with different accounts of what happened. Neither group made an obvious mistake. What would settle it, and what would only feel like settling it?
Then the sharper version. Two groups measured the mass change in the same reaction and got 0.42 g and 0.38 g, on balances that read to 0.01 g. Is that a disagreement, or is it two measurements of the same thing? Now change the numbers to 0.42 g and 0.41 g and answer again.
Only one of those pairs needs explaining. Knowing which is the whole skill.
Five claims to rank
- One carefully run trial, by a group whose technique you respect, never repeated.
- A result you reproduced yourself at the bench in twenty minutes.
- A pattern that held for every group in the room, with nobody able to say why.
- A mechanism that fits everything you know about bonding, with no measurement behind it yet.
- A value printed in a data table, compiled from measurements nobody in this room made or watched.
Rank them from most to least convincing. Then find someone who ranked them differently and work out why.
Claim 5 is the new one, and it is the one this course will make you argue about most. The table is a summary of measurements far better than any you can make on this bench — and it describes conditions your bench does not meet. When is the printed number the strongest evidence in the room, and when is your own reading the stronger of the two? See Reading a Data Table before you decide you know.
The questions we will hold every claim to
| Ask this | Weak answer | Strong answer |
|---|---|---|
| What was measured? | ”The results” | The quantity, the instrument, the resolution |
| How well? | ”Pretty accurately” | A stated uncertainty, in the same units |
| Compared with what? | Nothing | A control differing in one thing only |
| How many trials? | One | Enough to see the spread, and the spread quoted |
| Is the difference real? | ”They looked different” | The gap exceeds the combined uncertainty |
| What else could explain it? | Not considered | Alternatives named, and ruled out by a test |
| What would count against it? | Nothing could | A stated result that would sink it |
Rows two and five are the Grade 11 rows. A claim with no uncertainty attached is not a careful claim that happens to be missing a detail — it is a claim that has not yet said how much it is worth.
Two errors that behave completely differently
This is worth arguing about on its own, because the two look identical in a data table and want opposite responses.
Random error scatters your readings around the true value: where you judge the meniscus, when exactly you call the endpoint, small draughts on the balance. Repeat the trial and the scatter starts to average out. More trials genuinely help.
Systematic error shifts every reading the same way: a balance that was never zeroed, a solution that is more dilute than its label says, consistently reading the top of the meniscus instead of the bottom. Repeating does nothing at all — it makes you more confident in a number that is wrong, which is worse than being uncertain.
Precision is not accuracy, and this is where it bites
Five trials agreeing to three decimal places tells you your technique is repeatable. It tells you nothing about whether the value is right. Tight agreement is exactly what a systematic error produces, and it is the most persuasive-looking data you will ever generate. The only way to catch it is a comparison against something outside your own procedure — another group’s method, a standard, a check calculation.
Two traps worth naming out loud
“Consistent with” is not “supports”. Almost anything is consistent with a vague enough prediction. Data that is hard to explain any other way is the thing that carries weight.
Landing on the accepted value is not proof your method was sound. Two errors of opposite sign cancel, and the result looks excellent. If you cannot account for how your procedure produced that number, agreeing with the table is luck you have not yet distinguished from skill.
There is no answer key here
There is a real case for ranking claim 5 first and a real case for ranking it last, and the case depends entirely on what you are trying to decide. If the question is “what is the molar mass of magnesium”, the table wins and it is not close. If the question is “what happened in this flask, on this bench, this afternoon”, the table cannot answer it and your reading is the only evidence in existence.
Bring your ranking to the discussion in writing, and bring one number from your own work with an uncertainty attached to it. Afterwards, add an entry to your Chemistry Journal naming one place your ranking moved and what moved it. Related: Mistakes Are Data, Significant Figures in Practice, and Writing a Lab Report.
Curriculum connection
A1.8
synthesize, analyse, interpret, and evaluate qualitative and quantitative data; solve problems involving quantitative data; determine whether the evidence supports or refutes the initial prediction or hypothesis and whether it is consistent with scientific theory; identify sources of bias and error; and suggest improvements to the inquiry to reduce the likelihood of error
Link to original