Evidence boundary
Use supplied property datasets, graphs, and approved digital tools only.
Explicit science instruction
A periodic trend is an evidence pattern, not a command that every adjacent value must move smoothly. Graph the actual property values before applying a memorized trend.
A graph can reveal both the broad direction and local deviations. The electron-structure model is then used to explain the pattern while exceptions remain visible.
Predictions should include uncertainty: a directional prediction may be well supported even when an exact numerical value is not.
Vocabulary: trendline; deviation; residual; prediction; model limit.
Worked example with reasoning
Synthetic Period 3 radius values decrease overall from 186 pm toward 99 pm, but not every adjacent change is identical. The broad decrease supports a trend while the uneven steps show that the trend is not a perfectly linear law.
Guided question
Given five property values, sketch a graph and state whether the dominant pattern is increasing, decreasing, or non-monotonic.
Independent work
- Graph one supplied periodic-property dataset with axes and units.
- Annotate the broad trend and at least one deviation.
- Use the pattern to make one bounded prediction.
- State what extra data would increase confidence.
Investigation / Data Lab
Preserve the raw table, graph, trend statement, exception annotation, prediction, and one model revision prompted by the evidence.
Misconception check
A trend is not invalid because one point deviates; exceptions may reveal model limits or additional structure.
Formative understanding check
Why should a deviating point remain visible instead of being deleted?
Reveal response
It is part of the evidence and may reveal measurement issues, model limits, or additional chemical structure.
Mentor evidence conference
Ask Archaemenes to challenge whether your visual trend is stronger than the numerical data justify.
Teacher guidance
Foundation: Plot supplied points and describe direction.
Core: Analyze trend plus deviation and make a bounded prediction.
Extended: Compare two simple models for the same dataset.
Return to Thursday Evidence