PSYC201 Lectures 3, 4, 5, 6

• Operational definitions: a useful definition of a concept that specifies the operation used to measure or manipulate the concepts
• Manipulated variables:

  • Confounds - manipulating a factor other than the I.V. unintentionally
  • number of levels
  • strength of manipulation
• Strong manipulation:

  Adv:

  • more likely to reveal the effect- gives I.V. a good chance to work

  Dis:

  • ethical issues, generalisability
• Measured variables:

  • Self-report
  • Behavioural
  • Psychological
• 3 Approaches to research:

  • Naturalistic - field studies, cannot make predictive nor causal observations
  • Experimental - comparisons with control group, random assignment
  • Non-Experimental/correlational - casual observation
• Experimental: when differences on variable A are produced, then differences on variable B are measured
• Correlational: When differences on variable A are observed, then differences on variables B are measured
• Naturalistic observations:

  • discovery oriented, not proving through experimentation
  • invisible observers
  • observer bias
  • hard to distinguish cause from coincidence
  • low internal validity
  • high external validity
• Invisible observers: researchers must become participants or concealed - some behaviours occur in specific places, way to gain confidence around people to get more honest answers
• Observer bias: subjective rating of behaviour if they more acquainted with participants, their perspectives ultimately change
• Internal validity: behaviours being the product of the manipulation of the I.V and nothing else - in a lab for an experiment
• External validity: behaviours being observed will likely be those observed in a natural setting
• Single-blind study: participants do not fully know the conditions they are going to be subjected to
• Double-blind study: neither the researchers nor the subjects know what conditions the subjects are in - eliminates experimenter bias
• Non-equivalent groups: when the groups themselves provide the differences in results, instead of the manipulation of the I.V.
• Random Assignment: every participant has an equal chance of being put in groups
• Manipulation of the I.V:

  • Establish at least two conditions that differ systematically on the I.V.
  • Control all other variables so that none of them systematically differs across conditions (so that there is no confounding variable)
• Controlling confounding variables:

  1. Experimental control 1.a. Elimination
  2. Randomisation
  3. Counterbalancing
  4. Matching participants then random assignment
  5. Statistical control
• Experimental method: determining whether variables are related, where IV is manipulated and all other variables are controlled, either by randomisation or by direct experimental control
• Non-experimental/Correlational method: measurement of variables to determine whether variables are related to one another
• Factorial design- relationship between independent variables- to see how many conditions the study could lead to and which outcome is most/least likely to occur, when considering our hypothesis
• Problems with causal conclusions in Non-experimental research:

  • direction of cause and effect - unknown usually, whether A causes B or vice versa
  • The ‘third factor’ problem - an extra variable that could explain the supposed relationship between two variables
• Techniques for controlling third factors in non experimental research:

  • Holding constant- keeping the potential third factor constant - hard to control
  • Statistical control - statistically remove it from the equation
  • Matching - when assigning participants to the IV in an experiment. In a non-experimental study, it occurs at the selection into the study
• Applications of Non-Experimental Research:

  • Early stages
  • Ethical and practical limitations
  • The goal is prediction not cause - Prediction v/s control
• Classifying research:

  • causal statements require experimental demonstration
  • Non experimental data cannot establish causation
  • Naturalistic observations is inadequate to establish causation
• The ‘third factor’ problem: an extra variable that could explain the supposed relationship between two variables

  • C may cause either A or B, or A may have an effect on C which affects B or, B may have an effect on C which could affect A
• Are variables manipulated or measured:

  • decide based on hypothesis
  • if manipulated, create levels and how to expose them to participants
  • if measured, decide how to do so precisely
• Responsibility to colleagues and public:

  • data fabrication
  • misleading research reports
  • duty to correct errors
  • plagiarism
  • offensive behaviours failure to keep promises