PSYCH1000: CHAPTER 2: STUDYING BEHAVIOUR
SCIENTIFICALLY
1. Collecting data
- random vs representative sampling, sample size
2. Research methods
- descriptive (case study, naturalistic, survey)
- Correlational (involves variables)
- Experiments (involves variables)
3. Pitfalls and Problems
- lack of control of variables confounds findings
4. Ethics
- where/how do we draw the line?
Methods of Research
• Descriptive Research: how organisms behave (natural settings)
• Case Studies: studying one exceptional case —> suggest ideas/hypotheses to test, not
conclusive causal evidence
• Naturalistic/Observation: describe behaviour/relations between variables in real-life
settings, observer must not influence
• Surveys: representative samples can estimate results for an entire population,
unrepresentative samples=distorted results, interviewer biases/interviewee bias
• Correlational Research: degree of relationship between naturally occurring variables
• Still no manipulation, just measuring 2 things and comparing
• +/- correlations, +1.0 and -1.0 = strongest correlations
• Correlation =/= causation (3rd variable problem), but X can predict Y
• Experimental Research: best way to determine cause/effect relations
• One / more independent variable(s) manipulated to study impact on dependent variable
(other variables controlled)
• Between groups design (dif people randomly assigned either control or experiment group)
• Repeated measures/within groups design (all ppl assigned to all conditions)
• Randomly order conditions to counterbalance potential order effect
Threats to the Validity of Research
• High internal validity = permits clear causal conclusions
• Confounding = independent variable mixed with uncontrolled variable (can’t link one to
effect) —> ruins internal validity
• Placebo Effect = behaviour/reaction changed by expectations—> weak internal validity
• Experimenter Expectancy = subtle influence by researcher conveys hypothesis (w.i.v)
• Demand characteristics = cues in experiment that convey hypothesis (subjects try to “help”
researcher)
• Double blind procedure: removes biases in both parties (nobody knows expected results)
• External validity = ability to generalize cause/effect findings
• Researchers use replication to determine (try to achieve same results again)
• Meta-analysis (combine many studies testing same variables) to assess generality
Ethical Principles in Human and Animal Research
• Informed consent, right to privacy, benefits must outweigh risk, deception = controversial
• Animal research - humane treatment, risks justified by potential importance of research
• Research must be approved by ethics review board (w/ nonscientists) before conducting
SCIENTIFICALLY
1. Collecting data
- random vs representative sampling, sample size
2. Research methods
- descriptive (case study, naturalistic, survey)
- Correlational (involves variables)
- Experiments (involves variables)
3. Pitfalls and Problems
- lack of control of variables confounds findings
4. Ethics
- where/how do we draw the line?
Methods of Research
• Descriptive Research: how organisms behave (natural settings)
• Case Studies: studying one exceptional case —> suggest ideas/hypotheses to test, not
conclusive causal evidence
• Naturalistic/Observation: describe behaviour/relations between variables in real-life
settings, observer must not influence
• Surveys: representative samples can estimate results for an entire population,
unrepresentative samples=distorted results, interviewer biases/interviewee bias
• Correlational Research: degree of relationship between naturally occurring variables
• Still no manipulation, just measuring 2 things and comparing
• +/- correlations, +1.0 and -1.0 = strongest correlations
• Correlation =/= causation (3rd variable problem), but X can predict Y
• Experimental Research: best way to determine cause/effect relations
• One / more independent variable(s) manipulated to study impact on dependent variable
(other variables controlled)
• Between groups design (dif people randomly assigned either control or experiment group)
• Repeated measures/within groups design (all ppl assigned to all conditions)
• Randomly order conditions to counterbalance potential order effect
Threats to the Validity of Research
• High internal validity = permits clear causal conclusions
• Confounding = independent variable mixed with uncontrolled variable (can’t link one to
effect) —> ruins internal validity
• Placebo Effect = behaviour/reaction changed by expectations—> weak internal validity
• Experimenter Expectancy = subtle influence by researcher conveys hypothesis (w.i.v)
• Demand characteristics = cues in experiment that convey hypothesis (subjects try to “help”
researcher)
• Double blind procedure: removes biases in both parties (nobody knows expected results)
• External validity = ability to generalize cause/effect findings
• Researchers use replication to determine (try to achieve same results again)
• Meta-analysis (combine many studies testing same variables) to assess generality
Ethical Principles in Human and Animal Research
• Informed consent, right to privacy, benefits must outweigh risk, deception = controversial
• Animal research - humane treatment, risks justified by potential importance of research
• Research must be approved by ethics review board (w/ nonscientists) before conducting
