RM - Year 12

Involves manipulation of IV to measure effect on DV. Types of experiments = lab, field, natural and Quasi

A general statement of what the researcher intends to investigate & purpose of the study

Clear, precise, testable statement that states relationship between variables

States direction of difference of relationship

Doesn't state direction of difference or relationship

Aspect of investigation that is manipulated by researcher (or changes naturally) so effect on DV can be measured

Variable that is measured by researcher. Any effect on DV should be caused by IV

Clearly defining variables in terms of how they can be measured

Any variable other than IV that may affect DV if not controlled. EVs are nuisance variables that don't vary systematically with IV

Any variable other than IV that may affect DV so we don't know what affects the DV. They vary systematically with IV

Any cue researcher or research situation that may be interpreted by pps as revealing purpose of investigation. May lead to pps changing behaviour

Any effect of investigator's behaviour on DV. Includes design, sample, interaction etc.
2 types = direct and indirect effects

Using chance to control bias when deciding/designing order of conditions

Using same formalised procedures and instructions for all participants to avoid investigator effects

Different ways in which pps can be organised in relation to experimental conditions

Consequence of experimenter interacting with pp

Consequence of experimenter's design of study

Researcher manipulates IV and records affect on DV, while maintaining control of EVs in a controlled setting

Researcher manipulates IV and records affect on DV in a natural setting

IV is naturally occurring and isn't brought about by researcher. Affect on DV is measured

IV hasn't been determined by anyone, the variables simply exist

Does the experiment test what it's supposed to test?

Can the results be generalised?

Can we generalise to real life?

Can we generalise to other people?

Can we generalise to other time periods

Group of people who are focus of researcher's interest. Smaller sample drawn from population

Group of people who take part in research investigation. Sample drawn from target population and presumed to be representative

Method used to select people from population

+ High levels of internal validity due to high control. Cause and effect can be established
+ Easy to replicate so validity can be assessed
- Low external validity as it lacks ecological validity
- Demand characteristics so behaviour may be unnatural

+ Higher mundane realism than lab so higher external validity
+ Less demand characteristics due to natural setting
- Low control due to EV's. Cause and effect less easy to establish
- Precise replication not possible
- Ethical issues as pps don't know they're being studied

+ Opportunity for research that otherwise wouldn't be possible
+ High external validity as they are real life issues
- Few opportunities as natural events are rare
- Pps not randomly allocated to each condition so there may be CV's

+ Normally controlled conditions so high levels of internal validity and replication possible so validity can be assessed
- Pps not randomly allocated to each condition so there may be CV's

How much an experiment mirrors the real world (common place and ordinary)

Sample divided in groups and each group does 1 condition

+ Order effect don't occur as pps only do 1 condition
- More pps needed
- Pp characteristics might affect results

Pps do all conditions

+ Fewer pps needed
+ Pp characteristics which could act as EV's are kept constant between conditions
- Order effects such as fatigue, boredom and practice effect results

Sample divided into groups of pps who have been matched for relevant characteristics. Each group does 1 condition

+ Order effects don't occur as pps do only 1 condition
+ Fewer pp characteristics
- Matching pps is very time consuming
- Not possible to match all participant characteristics
- More pps needed

Attempt to control pp variables in independent groups which ensures each pp has same chance of being in each condition

In repeated measures, a CV arises from order of conditions eg. practice or boredom

Controls order effects
Half do A ----> B
Rest do B ----> A
(Remember *ABBA*)

Extent to which findings and conclusions from a study can be applied to the population. Possible if pp sample is representative of population

These arise when a conflict exists between rights of pps and goals of research to produce authentic and valid data

Quasi-legal document produced by BPS that instructs psychologist what behaviour is and isn't acceptable with human pps. The key principles are respect, integrity, competence and responsibility (RICR)

Informed consent
Deception
Confidentiality
Protection from harm
Right to withdraw

Demand characteristics

Presumptive
Prior general
Retrospective

Debrief pps:
Discuss aims
Allow for questions
Ask if they want to withdraw their data

Anonymity:
Number pps
Use pseudonyms
(Unless law states otherwise or an arrangement is agreed in advance)

Protection from harm

Debrief and give right to withdraw

Before, during and after experiment

Could lead to a biased sample

Small scale version of investigation before real one to check procedures, materials, measuring scales etc. and for researcher to make any necessary changes

Seeing/listening to what people do without having to ask them. Used in an experiment to assess DV

+ Captures spontaneous behaviour
- Researcher bias as behaviour can be interpreted in different ways (bias reduced if more than 1 observer)

Naturalistic & controlled
Covert & overt
Participant & non-participant

Observation of real behaviour in a natural setting where target behaviour normally occurs. No researcher interference

+ High external validity as behaviour is spontaneous and can be generalised to everyday life
+ No demand characteristics
- Low control of EV's

Observation that has some control of variables, including EV's

+ No demand characteristics as behaviour is natural
- Ethical issues (deception and privacy)

Observer participates

+ Greater insight. Increased validity as researcher experiences situation
- Decreased objectivity and validity as researcher may make subjective interpretations
- Risk of direct investigator effects

Structured and unstructured

Records *everything*

Planned system to record specific behaviours

Target behaviour broken into specific observable categories (operationalises behaviour)

+ Easy and systematic as data is quantitative and easy to compare
- Hard to clarify behaviours as some categories may overlap eg. grinning and smiling

*Event sampling* - target behaviour recorded EVERY time it occurs
*Time sampling* - observations made at set INTERVALS

+ Can record infrequent behaviour
- Complex behaviours over simplified which decreases validity

Any method where a person is asked to state/explain their own feelings, opinions, behaviours and experiences related to a topic

Pps respond to a list of pre-set Q.s. Used to assess pp's thoughts and experiences
Can be used to assess DV or find correlation

Questions where there's no fixed answer and respondents can answer however they want

Questions where there's a fixed choice of answers determined by question setter

Likert scale
Rating scales
Fixed choice option

+ Can reach a lot of pps so lots of data gathered quickly. Cost effective and low effort
+ Pps more willing to open up than in interviews
- Pps may lie to present themselves in a positive way which is a form of demand characteristic called social desirability bias
- Response bias means some pps favour a side (eg. always agree)

Live encounter where 1 person asks questions to assess an interviewee's thoughts and experiences. Questions may be pre-set (structured) or develop along the way (unstructured)

Structured
Unstructured
Semi-structured

Interview schedule reduces interviewer bias with pre-set questions
Quiet room
Rapport
Ethics
NV communication

+ Easy to replicate as it uses a standardised format which decreases differences between interviewers
- Interviewees can't elaborate which could cause frustration

Every member of population has an equal chance of being selected
Uses a random no. generator

Sample is a proportional representation of population
Population broken into smaller groups

Pps who are available at the time

Pps put themselves forward

nth member of population selected

In context of sampling, certain groups may be overnder represented which limits the extent to of generalisation to the target population

+ Unlikely sample will be biased
+ Everyone has equal chance
- Time consuming

+ Representative as whole population is equally represented
- Time consuming

+ Convenient, easy and quick
+ Likely to be ethical
- Less representative

+ Pps motivated
+ Easy
- Pps most likely to all be similar

+ No researcher bias
+ Representative
- Not truly unbiasedepresentative

Strength of relationship between variables. No direct manipulation of IV by experimenter

+ Used when manipulating a variable would be unethical
+ Further investigation if correlation is significant
+ Repeatable so validity can be assessed
- Correlation doesn't always mean causation
- Intervening/3rd variables may affect results
- May lack internal/external validity

Variables investigated within a correlation e.g. height a weight.

As one co-variable increases, so does the other

As one co-variable increases, the other decreases

No relationship between co-variables

Data that's expressed numerically

Data expressed in words

Info observed or collected directly from study

Data originally collected for a different purpose. Often used in correlation study or conducting a *meta analysis*

Making conclusions from previous research

+ Simple to analyse
+ Objective
- Too simplistic and oversimplifies reality

+ Rich detail
+ High external validity
- Difficult to analyse
- Conclusions based on subjective interpretations so there's researcher bias

+ High control over data
+ Data collection is designed to fit aims
- Lengthy and expensive

+ Simple and easy to access
+ Can be analysed using statistical test
- Data doesn't always fit study
- When analysing meta analysis, there may be methodological issues with research studies

Sample drawn should be representative of target population so that findings can be generalised