Research methods

Experimental method

Scientific approach = allows researchers to establish cause-and-effect relationships

Systematically manipulating variables whilst maintaining strict control over other factors

Operationalisation

Converting abstract concepts into measurable variables

Memory → word list

Extraneous variables

Any factors that might affect DV (apart from IV)

Confounding variables

Uncontrolled extraneous variables can skew results

Types of extraneous variables

Participant → individual differences between participants = age, personality traits, prior experience

Situation → environmental factors in experimental setting = time of day, noise

Experimenter → characteristics of researcher that may influence results = age, appearance, personality

Demand characteristics

Participants deduce aim of study or expected outcomes and change behaviour due to wanting to do what is right or purposely skew results

Unnatural behaviour due to nerves, evaluation anxiety, or displaying socially desirable responses

Investigator effects

Changes in participant behaviour due to the researcher's actions, biases, or characteristics → comfort, ethnicity, tone

Unconscious bias in data interpretation → find evidence that supports their expectations

Single blind procedure

Participants do not know which experimental condition assigned to → reduce demand characteristics

Double blind procedure

Participants + researcher do not know which experimental condition assigned to → reduce demand characteristics + investigator bias

Types of experiment: laboratory

Controlled environment = allow manipulation of IV + controlling potential confounding variables

+ Facilitate replication

+ Establish cause and effect

+ High control = high internal validity

- Low ecological validity

- May have demand characteristics

Types of experiment: field

Real world settings

IV remains manipulated by experimenter

Other variables controlled as much as possible

+ High ecological validity

+ No demand characteristics

- Less control over extraneous variables = difficult to establish cause and effect

- Difficult to replicate

- Ethics arise → participants are unaware + IV is manipulated

- Sample bias → participants aren’t randomly allocated to groups = may not be comparable to others

Types of experiments: natural

IV varies naturally

Researcher records effect on DV without direct manipulation of variables

+ Good when manipulating IV would be unethical → poverty or trauma affects on development

+ High ecological validity

+ No demand characteristics

- Less control over extraneous variables = difficult to establish cause and effect

- Difficult to replicate

- Sample bias → participants aren’t randomly allocated to groups = may not be comparable to others

Types of experiments: quasi

Researchers cannot freely manipulate IV

Cannot randomly allocate participants to different conditions

Often examine naturally occurring variables → gender differences, males + females are compared on various measures

+ Good when manipulating IV would be unethical → poverty or trauma affects on development

+ Useful for real-world applications

+ - Moderate ecological validity

- Limited ability to establish cause and effect

Observational techniques

Methods used to observe and record behaviours without direct manipulation of variables

May include naturalistic or controlled observations and offers insights into real-world behaviours

Types of observations: naturalistic

Observing + recording behaviour in environment where it would typically take place, without any interference from the researcher

+ High external validity because findings can often be applied to everyday life, as behaviour is studied in its natural context

+ Participants behave authentically since the environment is familiar

- Difficult to replicate due to lack of control over the research situation

- Many uncontrolled extraneous variables can make it challenging to identify clear patterns of behaviour

Types of observations: controlled

Observing and recording behaviour in a structured setting where the researcher manages certain aspects of the environment or variables

+ Allows for replication due to controlled conditions = less extraneous variables

- Lower external validity as the environment may not reflect real-life situations

- Risk of demand characteristics where participants alter behaviour because they know they are being observed = Hawthorne effect

Types of observations: covert

Participants' behaviour is observed and recorded without their awareness or permission

+ Eliminate participant reactivity (where knowledge of being observed changes behaviour)

+ Ensure natural behaviour is captured

+ Increase the validity of collected data = no Hawthorne effect

- Ethical concerns

Types of observations: overt

Observing and recording participants' behaviour after obtaining their knowledge and informed consent

+ No deceptive ethical issues

- May lead to ppt reactivity = Hawthorne effect

Types of observations: participant

Researcher joins and becomes part of the group they are studying

+ Provide researchers with direct experience of the situation

+ Offer increased insight into participants' lives and behaviours

+ May enhance the validity of findings through deeper understanding

- Risk of the researcher losing objectivity by identifying too strongly with the group

- The phenomenon of "going native" can occur when the boundary between researcher and participant becomes unclear

‘Going native’

Observer identifies with ppts too much = may cause them to lose objectivity → skewing results ~ may defend or justify group instead of analysing

Types of observations: non-participant

Researcher maintaining distance from the group being studied, observing from the outside

+ Reduces risk of going native

- May limit the understanding of group

Fundamental limitation of observational methods

Inability to establish cause and effect

Only correlations between variables

Observational design: behavioural categories

Dividing complex target behaviours into smaller, observable components that can be accurately measured and recorded

Observational design: event sampling

Researchers identify a specific target behaviour and then systematically record every single occurrence of that behaviour throughout the observation period

Observational design: time sampling

Establishing fixed time intervals (~ every 30 seconds) and recording what behaviour is occurring at those predetermined moments

Observational design: why use behavioural categories, event sampling or time sampling?

Form the foundation of systematic observational research, allowing researchers to transform complex behaviours into measurable, quantifiable data that can be analysed statistically

Observational design: unstructured design

Record everything observed

+ Rich data

- Hard to record everything comprehensively

- Only works for small scale observations

Observational design: structured design

Predetermined behavioural categories and systematic sampling methods

+ Easier than unstructured

+ Allows for larger scale studies

- Less rich data

Observational design: criteria for behavioural categories

• Observable - directly visible without interpretation

• Measurable - capable of being quantified

• Self-evident - clear without additional explanation

• Comprehensive - covering all possible forms of the target behaviour

• Exclusive - non-overlapping to avoid confusion between similar behaviours

Experimental design

• Independent

• Matched

• Repeated → use counterbalancing to reduce order effects

Self report techniques

• Questionnaires

• Interviews

Interview methods

• Structured interviews = standardised as far as possible

  • Fixed + predetermined set of questions is put to every ppt in same order + same way. 

  • Responses = recorded on questionnaire

• Unstructured = informal → attempt to break down barriers

  • No set questions + ppt can steer topic in direction of choice

  • Qualitative

Types of questions

• Open questions → encourage a full + meaningful answer using subject’s own knowledge + feelings = useful for qualitative data

• Closed questions → “yes” or “no” or specific information = useful for gathering specific details

• Postal questionnaires → reduce problem of interviewer bias

Pilot study

• Small scale preliminary study → conducted in order to evaluate the feasibility of the key steps in a future, full-scale project

Meta-analysis

• Statistical procedure used to combine + synthesise findings from multiple independent studies = to estimate average effect size for a particular research question

• Decisions are made about what studies are to be included/excluded

• Strengths: Increases conclusions’ validity → based on a wider range

• Weaknesses: Research designs in studies can vary = not truly comparable

Peer review

• Assess methods + designs used, originality of the findings, the validity of the original research findings and its content, structure and language

• Feedback from the reviewer determines whether the article is accepted

Reliability

• Measure of consistency

• Test-retest reliability: assessing the same person on two different occasions = shows extent to which the test produces the same answers

• Inter-observer reliability: extent to which there is an agreement between two or more observers

Measures of dispersion

• Range

  • + Easy to calculate

  • - Affected by outliers

• Standard deviation = calculates how a set of scores deviates from the mean

  • Low standard deviation = close to mean = reliable

  • + More sensitive → uses all the scores in the data set

  • - Time consuming

Measures of central tendency

• Mean

• Medium

• Mode

Nominal + ordinal + interval data

• Nominal = data put into categories ~ tally chart (basic)

• Ordinal = data ranged in order of ranking

• Interval= equal interval between each unit of measurement (sensitive + sophisticated)

Statistical distributions

Normal distribution → symmetrical = most data clustered around mean + medium + mode are equal

Left skew → mean > medium > mode (hump on right side) = tail on left side

• Negatively skewed distribution = indicating concentration of higher scores + a few low outliers

Right skew → mode > medium > mean (hump on left side) = tail on right side

• Positively skewed distribution = indicating concentration of lower scores + a few high outliers