Control of Variables (AQA A-Level Psychology): Revision Notes
Control of Variables
In psychological research, unwanted factors can interfere with the relationship between independent and dependent variables, potentially distorting results. Researchers have developed several methods to tackle this issue and maintain experimental control.
Types of variables affecting research
Extraneous variables
Extraneous variables are any factors other than the independent variable that might influence the dependent variable if left uncontrolled. These are often called "nuisance variables" because they don't change systematically alongside the independent variable.
Examples include participant age, laboratory lighting, or room temperature. While these variables may add "noise" to the data and make it harder to detect genuine effects, they don't completely invalidate the findings since they affect all conditions equally.
The key characteristic of extraneous variables is that they create random variation rather than systematic bias. This means that while they might make it harder to detect true effects, they don't fundamentally compromise the validity of your conclusions.
Confounding variables
Confounding variables present a more serious problem because they change systematically with the independent variable. This makes it impossible to determine whether changes in the dependent variable are due to the independent variable or the confounding factor.
Research Example: Energy Drink Study
Consider an energy drink study where the first ten participants (assigned to the water condition) happen to be introverted individuals, while the next ten participants (assigned to the energy drink condition) are extroverted.
The Problem: Personality becomes a confounding variable because extraversion varies systematically with the drink condition, potentially explaining any differences in chattiness between groups.
Why it's confounding: You can't tell if increased chattiness is due to the energy drink or because extroverted people were assigned to that condition.
Critical Distinction: Unlike extraneous variables that add random noise, confounding variables create systematic bias that can completely invalidate your results. They make it impossible to draw valid conclusions about cause and effect.
Demand characteristics
Demand characteristics occur when participants pick up cues from the researcher or experimental situation that reveal the study's purpose. Participants then modify their behaviour accordingly, making their responses less natural.
Participants actively try to make sense of unfamiliar experimental situations. They may look for clues about what behaviour is expected and either:
- Over-perform to please the experimenter (the "please-U effect")
- Deliberately under-perform to sabotage results (the "screw-U effect")
Either response means participant behaviour is no longer natural, creating an unwanted influence on the dependent variable.
Demand characteristics highlight the importance of keeping participants "blind" to the study's true purpose whenever ethically possible. This preserves the natural quality of their responses.
Investigator effects
Investigator effects refer to any influence the researcher's behaviour (conscious or unconscious) has on research outcomes. This encompasses everything from study design choices to participant selection and interaction during data collection.
For example, if researchers expect the energy drink group to be more talkative, they might unconsciously smile more or encourage greater chattiness from those participants. Hugh Coolican (2006) notes this can include expectancy effects and unconscious cues, as well as deliberate design decisions like leading questions.
Methods of control
Randomisation
Randomisation involves using chance wherever possible to reduce researcher influence on experimental design. This helps control investigator effects by removing subjective decision-making.
Examples include:
- Randomly generating word order in memory experiments rather than letting the experimenter choose positions
- Randomly determining the order of experimental conditions for each participant
- Using random assignment to allocate participants to different groups
In multi-condition studies, randomisation ensures that order effects don't systematically bias results. This provides an alternative to counterbalancing techniques.
Standardisation
Standardisation ensures all participants experience identical environmental conditions, information, and procedures. This involves creating detailed protocols specifying exactly what will happen during the study.
Key elements include:
- Standardised instructions read to every participant
- Identical physical environments for all sessions
- Consistent researcher behaviour across participants
- Fixed procedures that don't vary between conditions
Proper standardisation prevents procedural variations from acting as extraneous variables.
Standardisation and randomisation work together as complementary control methods. Standardisation eliminates unwanted variation, while randomisation distributes any remaining variation equally across conditions.
The Hawthorne effect
The Hawthorne effect describes how some people work harder and perform better when participating in research, simply because they're receiving attention from researchers rather than due to any experimental manipulation.
This phenomenon was first identified by Henry A. Landsberger in the 1950s during his analysis of 1920s-1930s experiments at the Hawthorne electric company. The original studies investigated whether changing workplace lighting affected worker productivity.
Researchers found that productivity increased during the experimental period but then decreased once the study ended. This suggested that worker attention from the research team, rather than lighting changes, caused the productivity improvements. Landsberger defined the Hawthorne effect as a short-term performance improvement caused by observing workers.
The Hawthorne effect represents a specific form of demand characteristics where participant behaviour changes due to awareness of being studied.
Key principles summary
| Variable Type | Description |
|---|---|
| Extraneous variables | "Nuisance" variables that don't vary systematically with the IV and can often be controlled before the experiment begins |
| Confounding variables | Variables that vary systematically with the IV, making it impossible to determine what caused changes in the DV |
| Demand characteristics | Participants interpret cues from the experimenter and research situation, potentially changing their behaviour |
| Randomisation | Using chance to reduce researcher influence on investigation design |
| Standardisation | Ensuring all participants receive identical instructions and experiences |
Key Points to Remember:
- Extraneous variables are controllable "nuisance" factors that don't systematically vary with your independent variable
- Confounding variables are the real problem - they change systematically with your independent variable, making results uninterpretable
- Demand characteristics occur when participants guess the study's purpose and alter their natural behaviour
- Randomisation uses chance to eliminate researcher bias in design decisions
- Standardisation ensures every participant experiences identical conditions, preventing procedural variations from affecting results