control in experiments

why is control important

• Control refers to the methods researchers use to minimise the influence of extraneous and confounding variables so that the effect of the independent variable (IV) on the dependent variable (DV) can be clearly observed.

Without control:

• Results lack internal validity (we can’t be sure the IV caused the DV change).

• Findings can’t be replicated (poor reliability).

COUNTERBALANCING

🔹 Definition:

A technique used in repeated measures designs to deal with order effects.

Order effects occur when participants’ performance in one condition affects performance in the next (due to practice, boredom, fatigue, or demand characteristics).

🔹 Purpose:

To balance out order effects across conditions so they don’t systematically bias results.

🔹 Example:

If participants do two tasks — A and B — counterbalancing means:

• Half of participants do A → B

• Half do B → A

This ensures any improvement or fatigue is spread evenly across conditions.

🔹 Types:

• AB/BA design: simplest form — half get each order.

• Latin square: more complex, used if more than two conditions.

🔹 Strengths:

✅ Controls order effects → improves internal validity
✅ Ensures both conditions are equally affected by practice/fatigue

🔹 Weaknesses:

❌ Doesn’t eliminate order effects completely — just balances them
❌ Participants still complete both conditions → risk of demand characteristics

RANDOMISATION

• 🔹 Definition:

  Using chance to determine the order or presentation of conditions, tasks, or stimuli.

  This removes researcher bias in how procedures are carried out.

  🔹 Purpose:

  To prevent the researcher from unintentionally influencing the study (e.g., by always presenting easy questions first).

  🔹 Example:

  • Randomising the order of word lists in a memory test.

  • Randomising the sequence of visual stimuli.

  • Randomising which participant experiences which task first.

  🔹 How to Do It:

  Use random number generators, shuffled decks, or computer-based randomisation.

  🔹 Strengths:

  ✅ Controls order and selection bias
  ✅ Reduces investigator effects
  ✅ Helps maintain objectivity and internal validity

  🔹 Weaknesses:

  ❌ Randomisation doesn’t control all variables — only reduces bias from order/presentation
  ❌ May create uneven or unbalanced sequences by chance

STANDARDISATION

🔹 Definition:

Keeping all procedures, instructions, and environments identical for all participants, except for changes in the IV.

This ensures every participant has the same experience apart from the variable being studied.

🔹 Purpose:

To control situational variables (environmental differences) and investigator effects.

🔹 Example:

• Using a script for instructions

• Testing all participants in the same room, time of day, and duration

• Giving everyone the same materials

🔹 Strengths:

✅ Improves reliability — study can be replicated
✅ Controls extraneous variables → increases internal validity
✅ Reduces investigator effects

🔹 Weaknesses:

❌ Can make experiments feel artificial → reduces ecological validity
❌ May not account for individual needs (e.g., participants might need clarifications)