research methods

• one tailed directional hypothesis

  participants in [IV condition 1] will ... significantly more/less [DV] than participants [IV condition 2]
• non directional two-tailed hypotheses

  there will be a significant difference in the [DV] in the participants [IV condition 1] compared to participants [IV condition 2]
• null hypotheses

  there will be no significant difference in the [DV] of the participants [IV condition 1] compared to participants [IV condition 2]. Any difference will be due to chance
• lab experiments have:

  high level of control over variables

  standardised procedures

  independent variable manipulated by researcher
• field experiment is conducted in natural setting

  independent variable manipulated by researcher
• quasi experiment

  independent variable already exists in participants

  cannot randomise between conditions of experiment
• Natural experiment

  researcher records change in dependent variable between two levels of independent variable

  happens when an event causes people to form into levels of IV such as natural disasters or political events
• lab experiment evaluation

  highly controlled so high internal validity, can establish cause and effect

  standardized so can replicate

  lacks ecological validity and mundane realism

  demand characteristics may be present
• field experiment evaluation

  high ecological validity and mundane realism

  less demand characteristic

  lack of control over extraneous variables so low internal validity

  difficult to randomly assign participants to different conditions
• natural experiment evaluation

  allows research unable in controlled experiments due to ethics/cost

  high ecological validity

  cannot control extraneous variable or establish cause and effect

  cannot be replicated
• quasi experiment evaluation

  only way to study factors that are pre existing characteristics in participants

  some factors related to level of IV cannot be controlled

  confounding variables present
• independent measures design

  different participants in two or more conditions

  randomly allocated to avoid researcher bias

  produces unrelated data
• repeated measures design

  same participants in each condition

  produces related data
• order effects

  when participants perform better with practice or worse with fatigue

  reduce by counterbalancing

  ABBA format in repeated measures design
• matched pairs design

  different participants in each condition

  participants assessed and ranked on characteristic

  top 2, then following two, randomly assigned to each condition

  produces related data
• independent measures design evaluation

  participant less likely to work out aim

  reduces demand characteristic

  no order effects

  participant variables
• repeated measures evaluation

  does not need as many participants

  less participant variables if they participate in both conditions

  participants likely to work out aim

  demand characteristics

  order effects
• matched pairs evaluation

  reduce participant variables as characteristics are matched

  no order effects

  time consuming

  participants are similar NOT identical
• single blind control

  participants do not which condition of the IV they are in

  cannot alter behaviour

  reduces demand characteristuics
• double blind control

  participant and researcher do not know who is in what condition
• validity

  degree to which a measure accurately assesses a specific concept, trait or construct as it claims
• internal validity

  whether the study design, conduct and analysis answers the researchers questions without bias
• face validity

  whether a test appears to measure what it is supposed to measure
• temporal validity

  validity of findings in relation to the progression of time
• concurrent validity

  extent to which results of a particular test/measurement correspond to those of previously established measurements for the same construct
• external validity

  whether the findings can be generalised to other settings/contexts
• population validity

  whether you can generalise findings from your sample to a larger group of people
• ecological validity

  extent to which your findings can be generalised to real world situations
• reliability

  consistency of measurement, instrument or procedure of yielding the same results on repeated trials
• internal reliability

  internal consistency of a measure
• external reliability

  consistency of a measure from one use to another
• inter-rater reliability

  extent to which different raters agree in assessing a specific phenomenon, behaviour or characteristic
• test/re-test reliability

  administer same test twice over a period of time with the same participants

  correlate scores
• split half method

  test split into two halves

  score each half seperately

  correlate scores using statistical method (pearson correlation)
• informed consent

  participants told purpose of study and risks

  researcher may not fully disclose due to potential demand characteristics
• types of consent
• types of consent

  presumptive- gather if similar people would consent and generalise

  prior general- participant consents for many studies

  retrospective- consent after study during debrief

  parental consent- for those under 16
• right to withdraw at any point in study

  data can be removed
• confidentiality

  protection of personal data

  replace anything that can be linked to identities when publishing

  replace names with numbers
• privacy

  right to control how participants information is released/used

  difficult to not invade privacy in a field study

  maintain anonymity

  replace names with numbers