Research methods- book 1

• Independent variable


  Variable which you change
• Dependent variable


  variable that you measure
• Confounding variable


  Are not controlled because they either can't be controlled or haven't been identified. Can't be certain IV effects the DV.
• Extraneous variables


  variables which might impact on the DV which are then identified and controlled.
• Standardisation


  - Minimises the possibility of a variable other than the IV having and influence on the DV.

  - Each participant must follow the exact same instructions, any deviation will impact results.
• Aim


  a statement about what the experiment is trying to find out
• Hypothesis


  a prediction of the outcome of the experiment. Can be research, experimental and correlational.
• Research hypothesis


  all types of research including correlational analysis and experiments.
• Alternative hypothesis


  they predict the outcome, can be directional or non-directional and experimental or correlational.
• Null hypothesis


  Hypothesis that states there will be 'No significant difference' or there will be 'no correlation.' Opposite of alternative hypothesis.
• Directional hypothesis (one-tailed hypothesis)


  pedicts the direction of the outcome, used when there is previous research and so there is some idea what the outcome is likely to be.
• Non-directional hypothesis (two-tailed hypothesis)


  predicts that the IV will have an impact on the DV but does not predict the direction of the outcome. Used when there no previous reseach or previosu conflicting research.
• Correlational hypothesis


  Only write if the text includes: correlation, relationship, link or association.
• Non-directional correlational hypothesis


  Relationship between variable A and variable B.
• Directional correlation hypothesis


  There is a negative/positive relationship between variable A and variable B.
• Null correlational hypothesis


  there is no relationship between variable A and variable B.
• Experimental hypothesis


  When researched is looking for difference between two conditions. Two independent variables (2 conditions) and dependent is being measured.
• Non-directional experimental hypothesis


  There is a significant difference in (dependent variable) between condition A and condition B.
• Directional experimental hypothesis


  (The dependent variable) will be higher/lower in condition A than condition B.
• Null experimental hypothesis


  There is no significant difference in (the dependent variable) between condition A and condition B.
• Operationalised variables


  make it clear what the independent and dependent variables are and how they are going to be measured.
• Experimental method
• Laboratory experiments


  In a lab, under controlled conditions, extraneous variables are identified and controlled, cause and effect can be established.
• Control groups


  Used when they want to compare two groups of participants to see whether the manipulation of the IV has an effect on the DV. Control group will be used.
• Randomisation


  Minimise investigator effect and helps to reduce bias within the results. Investigator might have an idea of the outcome of the study causing bias towards the results.
• Replicability


  highly controlled nature of a lab experiment, can be repeated over and over again.
• External validity


  refers to the extent at which results can be generalised beyond the research setting. Lab experiments often lack external validity because they are so controlled.
• Field experiments


  carried out in natural settings and so have higher mundane realism because behaviour is more authentic. Good external validity.
• Natural experiment


  type of field experiment, cause and effect can't be established as IV can't be manipulated, high external validity
• Quasi-experiments


  There is an IV but it can't be manipulated directly. Still possible to see impact of IV on DV. No random allocation to each condition. There is no control group. Cause and effect can't be established. E.g. gender.
• Demand characteristic and investigator effects
• Demand characteristics


  when participants work out what a study is trying to achieve and they change their behaviour to either under or over perform.

  - They are less likely to occur when participants don't know they are taking part in an experiment in a natural setting, field experiment.

  - Please-U effect

  - Screw-U effect

  - easiest way to prevent is through deception, participants are unaware of the full extent of the study.
• Investigator effects


  - Investigator has an impact on the study

  - May not be consciously aware of it

  - Experimenter might lead a participant to perform a task in a certain way because the results are anticipated
• Blind study


  - helps avoid investigator effects

  - person observing behaviour or administering the experiment has no knowledge of the hypothesis and cannot read data in a biased way.
• Double-blind study


  - helps avoid investigator effect

  - neither the experiment nor the participants are aware which condition they are in.
• Experimental designs
• Independent groups


  - Random allocation

  - Different participants are used in each condition

  - Done by putting participants names into a hat

  - Means theres no research bias in the allocation of participants to each condition.

  - S- same material can be used in each condition, no order effects, participants more likely to remain naïve.

  - L- Requires more participants, might be participant variables that change the results.
• Repeated measures


  - The same participants are used in both conditions

  - S- eliminated participant variables, fewer participants needed.

  - L- Order effects such as practice and fatigue, performance could either be better or worse in second condition.
• Counterbalancing


  - used to over come order effects

  - half the participants complete condition A and condition B while the other half condition B then condition A.

  - L- participants work out the aim of the experiment by completing both conditions, same sets of material cannot be used in both conditions. So need to find two similar sets of material.
• Matched pairs


  - participants are matched on variables which are important in terms of the piece of research.

  - Age, IQ, gender etc.

  - S- reduces participant variables, no order effects

  - L- very time consuming, impossible to eliminate all participant variables which may change results.