Control in Experimental Methods

What is random allocation?

This involves allocating participants to experimental groups in an unbiased way, so they have an equal chance of being selected to take part in each condition. This is to evenly distribute participant chars across the condition to avoid extraneous variables and improves internal validity. It is used to address the problem of participant variables(individual differences) in an independent groups design. This can be done by assigning each participant a number or identifying them by name; the names or numbers are placed in a random number generator/hat ; the first participant drawn is assigned to condition 1 and the next 15 are assigned to condition 2.

What is counterbalancing?

Counterbalancing is used to control for the impact of order effects in repeated measures design and improve internal validity. It involves making sure each condition comes first or second in equal amounts. E.g. Half the participants take part in condition A then B, and the other half take part in conditions B then A. Counterbalancing ensures order effects are distributed evenly across all conditions. Counterbalancing does not reduce order effects because all participants still take part in all conditions therefore rider effects are will present, but their effect is balanced across the conditions.

What is randomisation?

Randomisation refers to the use of chance wherever possible when designing materials and deciding the order of conditions. This is to avoid researcher bias influencing the design of the study; it is an attempt to control investigator effects and improve internal validity. E.g. if in a memory experiment the participants have to learn a list of words; the order of the words should be randomly generated. Or if the experiment involves a number of conditions, the order of the words should be randomly generated. Or if the experiment involves a number of conditions, the order of these conditions should be randomly determined.

What is standardisation?

Standardisation is a way of ensuring all participants have exactly the same environment, information and experience, so individuals experience does not cofound the results, thus improving internal validity. Standardised procedures ensure all participants are tested under the same conditions to enable to deal with experimenter bias and facilitate replication to ensure findings can be checked for reliability. Standardised procedures should include a set of standardised instructions that are given in the same way to all participants to deal with investigator effects.

What are control groups?

A control group consists of participants who experience the control condition of an experiment in which the IV is absent or held constant. The results from the control group are used as a baseline which is compared to the results from the experimental group(s) whose participants experience at least one experimental condition in which the independent variable is varied. E.g. in Bickman’s study, the control group included participants who were approached by a confederate wearing plain clothes. The level of obedience from this control group was compared to participants in the experimental groups who were approached by a confederate wearing either a milkman uniform or security guard uniform.

Improving Reliability of Experiments

When possible, it is better to take more than one measurement from each participant and take a mean score. This reduces the impact of any anomalous score. Pilot studies can be used to check the proposed method measurements works properly and that participants can use any apparatus successfully.

Improving External Validity of Experiments

Ecological validity can be improved by conducting the experiment in a more naturalistic, real life setting e.g. a field setting. Population validity can be improved by taking larger, more varied samples, for example including a wider age group or geographical area.

What are demand characteristics?

Demand characteristics: Are cues in the experimental situation that can make participants aware of what the researcher wants to find, or how they are expected to behave. Demand characteristics may cause participant reactivity, whereby participants change their behaviour by:

• Trying to guess the purpose of the research and acting in a way that they feel is helpful or unhelpful to the researcher

• Acting nervously and out of character because of being evaluated

• Displaying social desirability bias, wishing themselves to be seen in the most favourable light possible

What are ways to reduce demand characteristics? (2)

• Use a single-blind technique: This is where the participant has no idea of the research aims and/or which experimental group they are in

• Use in deception and distractor questions

What are investigator effects?

Investigator effects: Are any unintentional influences of a researcher’s behaviour or characteristics on the performance of the participants. For example:

• Physical characteristics: (e.g. age, gender and ethnic origin)

• Behaviour such as verbal communication( e.g. tone of voice, accent), or non-verbal communication(e.g. smiling, eye-contact)

• Expectations about the outcome of a study

What are ways to reduce investigator effects? (2)

• Use a double-blind technique: Neither the participant nor the investigator is aware of the true aims or details of the investigation, so have no expectations

• Use standardised instructions: This decreases the possibility that the investigator can communicate expectations. Researchers can be trained to use a neutral tone for reading instructions. For interviews a standardised script can be used.

What is a pilot study?

A pilot study is a small-scale trial run of a study which takes place before the real investigation is carried out. Pilot studies allow practical details to be checked, like ensuring instructions can be understood and deciding on the number of participants to use.

What is the aim of pilot studies? (3)

• To check that the procedures, materials and measuring scales work

• To allow the researcher to make changes or modification to the design if necessary

• To improve the quality of research, help avoid unnecessary work, save time and money

What techniques does the pilot study use?

• Questionnaires and interviews: a pilot study can involve trying out questions in advance to be able to re-word or remove any that are confusing

• Observational Studies: a pilot study can check the behavioural categories and coding systems are suitably operationalised and used reliably across observational team.