Colored Light and Bean Seedlings: A Confounded Investigation

Confound 1: unequal natural light (and likely temperature) at the windowsill. The red tray receives sunlight the other trays do not, plus whatever warmth the window supplies. Sunlight is full-spectrum and far more intense than a classroom lamp, so "red" is really "red lamp plus direct sun." Either factor alone could account for the 12.4 cm mean, which means the red result cannot be attributed to light color.

Confound 2: inconsistent watering. The blue tray got an extra watering on day 9. Overwatering can slow growth by limiting root oxygen, so the blue mean may be depressed by the watering schedule rather than by blue light. The effect goes in the opposite direction from confound 1 - it makes blue look worse than blue light actually is.

A third, quieter problem: no replication. One tray per condition means a single tray of poor seeds or a single shallow spot on the shelf could produce the whole difference. With n = 1 per condition, the tray and the treatment are the same variable.

Why this matters for the conclusion: the observed ordering is consistent with the confounds alone, so the data do not discriminate between "red light is best" and "sunlight is best" or "blue light is fine but that tray got drowned."

Redesign essentials:

1. One location for all trays - same shelf, same distance from each lamp, no windowsill. Block the window or use a windowless corner if possible.
2. Hold everything constant except light color: identical lamp wattage and distance, identical pots, soil, seed lot, and planting depth; the same measured volume of water on the same schedule (use a graduated cup, not "every other day" by feel).
3. Replicate: at least three trays per condition, with trays assigned to positions at random and rotated periodically, so a position effect is spread across conditions rather than stacked on one.
4. Measure the same way every time: same ruler, same time of day, same definition of height, and record temperature in each position.
5. Include the no-lamp control in the same location, which tells you how much growth is attributable to the lamps at all.

The comparison this makes valid: with location, water, and replication controlled, any consistent difference among red, blue, and green can be attributed to light color, and a difference between the lamped trays and the no-lamp tray can be attributed to the presence of lamp light. Right now neither comparison is interpretable.

Response to the parent: the red tray was on the windowsill, so it got direct sunlight the other trays never received. What the class actually observed is "the tray in the sun grew tallest" - which is a real finding, but not a finding about red light. The blue tray also got overwatered once, so it may have been set back for a reason that has nothing to do with color. I would not spend money on red bulbs based on this run.

What the claim would require:

• Replicated trays per color (three or more), so a single bad tray cannot drive the result.
• Every tray in the same spot, with the same lamp distance and wattage, the same soil and seed lot, and the same measured watering.
• Temperature logged, since lamp color often comes bundled with heat differences.
• The test repeated, ideally with a different plant species and a second run, because one three-week trial in one classroom is a single data point no matter how many seeds it contains.

Modeling the reasoning for students: the honest classroom conclusion is narrower than the students' first claim - "the trays under lamps grew more than the tray with no lamp, but our color comparison is confounded." That narrower claim is still a real result and is exactly the kind of statement a scientist would defend. A good follow-up question for the class: which single change to our setup would have made the biggest difference to what we could conclude?