Control Over Variables

Understanding variables is absolutely crucial for any psychology experiment you'll design or evaluate. Think of it like cooking - if you change multiple ingredients at once, you'll never know which one made your cake taste amazing (or terrible!).

Extraneous variables are sneaky little factors that could mess up your results without you realising. They're not the main thing you're testing (the independent variable), but they might still affect your outcome (dependent variable). These come in two main flavours: participant variables (like how much sleep someone had) and situational variables (like noise in the testing room).

Confounding variables are even worse - they're extraneous variables that have actually gone and affected one condition but not the other. Imagine testing memory techniques where all the participants in Group A happen to be naturally good at memorising, whilst Group B struggles. Your results would be meaningless because you can't tell if your technique works or if you just got lucky with naturally gifted participants.

Key Insight: The golden rule is simple - only your independent variable should influence your dependent variable. Everything else needs controlling!

Demand characteristics occur when participants start playing detective, trying to figure out what your experiment is really about. Once they think they know, they might try to be helpful $the "please-U effect"$ or deliberately sabotage your results $the "screw-U effect"$. Either way, you're not getting natural behaviour anymore.

Investigator effects happen when the researcher accidentally influences the results through their own behaviour. This could be unconscious body language, the way they phrase instructions, or even how they select participants. It's surprisingly easy to bias your own research without meaning to.

Psychology experiments also deal with order effects - when the sequence of conditions affects performance. Participants might get better with practice or worse due to fatigue, which could skew your results completely.

Random allocation helps sort out individual differences between participants by using chance to assign people to different groups. Randomisation uses chance more broadly - like randomly ordering your test materials to avoid bias.

Counterbalancing (the ABBA technique) tackles order effects by having half your participants do Condition A first, then B, whilst the other half does B first, then A. Standardisation ensures everyone gets exactly the same experience through identical procedures and word-for-word instructions.