Research Methods - Validity, reliability, extraneous variables, aims and hypothesis, sampling and ethics

Research methods are systematic procedures used to collect, analyse and observe data to investigate human behaviour and mental processes.

Reliability

The extent to which a test produces consistent results

Validity

The extent to which a test produces legitimate results

Internal Validity

The extent to which data measures what it set out to measure

External validity

The extent to which data can be generalised beyond the research setting

Ecological validity

The extent to which data can be generalised to real life

Temporal Validity

The extent to which data can be generalised to other times

Variables

Things that are open to change in an experiment/research

Independent variable

This variable is manipulated by the researcher. This is operationalised by clearly expressing it so it can be manipulated to create a minimum of two conditions (in a piece of reseach)

Dependent Variable

This variable is measured by the researcher to observe if manipulating the IV causes change. This is operationalised by clearly expressing what is being measured (in a piece of research)

Extraneous Variables

Also known as control variables. These variables can affect the dependant variable if not controlled. They are operationalised by being identified and controlled so the experiment has a high internal validity.

Variables in experiment - Music + concentration test - operationalise!

Independent Variable = One group is under the condition of working in silence whilst completing a word search for 5 minutes. The other group listens to music at a fixed volume whilst doing the same

Dependent Variable = The number of words found in the word search

Extraneous variables = The level of distraction in the room, the time allowed for the word search, time allowed to prepare for word search, level of food + sleep

Operationalisation

Clearly defining a variable so they can be manipulated (IV) or measured (DV). Must be specific to the given experiment.

Participant Variables

Characteristics of participants that can act as extraneous variables if an independent groups design is used and the participant variables are relevant to the dependant variables

Control of Participant variables

Randomisation can mean that the likelihood of one group being made up of a large amount of people with one characteristic is smaller, but this can still have an effect. The best method is the matched pairs design - this pairs people with the same characteristics together and splits the pair between the conditions so PV has no effect.

Examples of Participant variables

Gender, IQ, age, eyesight

Situational Variables

Features of the environment that are only act as extraneous variables if they are relevant to the DV

Control of Situational Variables

Standardisation (making the same) of situational variables means all participants are subject to the same environmental conditions

Examples of situational variables

Light level, noise level, temperature, time of day

Order Effects

Only extraneous variables when the same groups of participants is tested in all variables (repeated measures design) as they can overperform (practise) or underperform (tired/bored)

Control of Order Effects - counterbalancing

Counterbalancing is used so half of the participants experience condition A then condition B and the other half experience it in reverse. This means any order effects cancel each other out as for both groups in their condition 2 they are both over/under performing and when comparing the conditions this can be ignored

Demand Characteristics

An extraneous variable when participants behave in unnatural ways in order to either please the researcher (please-U effect) or sabotage the experiment (screw-U effect) when they are able to guess the aim of the experiment due to cues. This is most common in Repeated measures designs.

Control of Demand Characteristics

Very difficult to control. The best way is to use the independent group design or single blind designs so that it is harder for participants to guess the aim of the experiment.

Investigator Effects

An extraneous variable when the researcher’s characteristics or behaviour has an influence over the research outcome. (e.g. participant shares less with a researcher of the opposite gender OR researcher design experiment to be easier

Control of Investigator Effects

Very difficult to control! Double blind design where the researcher is unaware of the condition the participant is in is best OR randomisation of testing so it is not designed to be easier or harder

Independent Groups Design

When all the Participants are split into separate groups. Each group experiences one condition. Can cause issues with participant variables and situational variables

Matched Pairs design

Method of controlling participant variables - the participants are put into pairs with the same characteristics and then these pairs are split between two groups and subject to different conditions. These groups have the same split of participant characteristics so can be compared.

Single blind testing

Used to control demand effects. The participants (NOT THE RESEARCHER) in the experiment are kept unaware of the conditions they are in so they cannot assume what is being tested and change their behaviour.

Double Blind testing

This controls investigator effects. The investigator AND the participants are unaware of the conditions they are in. This means the investigator’s behaviour towards the participants does not change between the groups.

Control Groups

This group is not exposed to the variable under investigation and it is used as a baseline that the experimental group can be compared to (like the group given a placebo in a drug trial).

Experimental groups

This group is exposed to the variable under experimentation (independent variable). If this causes significant changes in the data compared to the control group (dependent variable) we can determine this change was as a result of the independent variable (if extraneous variables were controlled).

Aims

The intended purpose of a psychological investigation, outlines what the researcher is trying to discover

How to write an aim

Always starts with the phrase “to discover”

Experiment = To discover the effect of (independent variable) on (dependent variable)

Correlation = To discover the relationship between (co-variable 1) and (co-variable 2)

Hypothesis

A statement that predicts the outcome of a psychological investigation - always start a piece of research with two (null and alternate) and one is chosen after collecting and analysing data - must operationalise the IV and DV

Alternate Hypothesis

A statement which predicts a difference or correlation in results

Directional alternative hypothesis

It states the direction of the predicted difference or correlation in results - used if there is pre-existing research

eg) Participants who are in silence will find more words that participants who listen to music / there will be a positive/negative correlation between

Non-directional alternate hypothesis

It does NOT state the direction of the predicted difference or correlation - used if there is NO pre-existing research

eg) There will be a difference in the number of words found between the participants who are in silence and the participants who listen to music / there will be a correlation between

Null Hypothesis

A statement that predicts no difference or correlation in results

eg) There will be no difference in the number of words found between the participants who are in silence and the participants who listen to music / there will be no correlation between 

Population

The entire group that a researcher is interested in studying

Sample

A smaller, manageable subset of the population that data is collected from

Unbias sample

A sample that is representative of the population the researcher is interested in studying.

Bias sample

A sample that is not representative of the population the researcher is interested in studying.

Opportunity Sampling

Selects participants based on them being easily available at the time of study - it is quick rapid + easy but usually creates a  biased sample because it is drawn from a very small part of the population

Volunteer Sampling

Uses an advertisement places in a location members of the chosen population will see it to get participants to volunteer to be in the study by responding to the advert - It can give access to a wide variety of participants but it is likely to create a bias sample due to volunteer bias

Volunteer Bias

The fact that most volunteer participants are highly motivated and have extra time, which is not likely to generalise with the whole population

Random Sampling

All members of a population are assigned a number, participants are then selected by random selection (eg. Random number generator) from the population - likely to create an unbias sample as all have an equal chance of selection, but it requires a list of all members of a population which may be difficult to generate for large populations

Systematic Sampling

Participants are selected using a predetermined system (eg. every nth population member) - using an objective system is likely to create an unbias sample, but it is not truly unbias unless the numerical interval is selected by random

Stratified Sampling

Subgroups (strata) within a population are identified and a pre-determined number of pts. are taken from each of these subgroups in proportion to their representation in population. This number of participants are then selected from each subgroup using random sampling. Unbias sample, it is proportional and randomly selected representation of subgroups within the sample, BUT it is time consuming

Ethical issues

Conflicts between the rights of participants and what researchers must do in order to conduct valid, useful research. These include: Informed consent, deception, protection from harm and privacy + confidentiality

British Psychological Society’s code of ethics

Implemented by ethics committees in research institutions in order to deal with ethical issues. If psychologists fail to observe and follow these guidelines they violate their professional duty and risk disbarment from the society

Informed Consent

Participants, in order to make an informed decision on participating, should be aware of: research aims and procedures, their rights and the right to withdraw and what their data will be used for

HOWEVER - this can lead to demand characteristics that cause abnormal behaviours and invalidate the study

Guidelines on Informed Consent

Participants are given a consent for detailing relevant information that may affect their decision to participate, if they agree this is signed. Under 16s need parental consent.

Consent can also be obtained through other means with an ethics committee’s approval: presumptively, prior agreement to be deceived, retrospectively

Deception

The deliberate misleading or withholding of information from a participant, meaning they cannot give informed consent. This is because informed consent can cause demand characteristics and invalidate data

Guidelines on Deception

The need for deception must be approved by an ethics committee that make a cost-benefit judgement. After the study pts should be given a full debrief of the research and be informed of: research aims, other unsupplied details and what data will be used for. The option to withdraw their data must then be given.

Protection from Harm

Participants should not be exposed to more risk of physical (injury) or psychological (embarrassment, stress) harm than in their daily lives.

HOWEVER - studying more important psychological questions may involve a degree of distress to pts. OR it can be difficult to predict the outcomes of certain procedures so it is difficult to guarantee protection from harm

Guidelines on Protection from Harm

The need for harm must be approved by an ethics committee that make a cost-benefit judgement. Participants should be reminded of their right to withdraw, be debriefed after the study if harmed and the researcher should provided counselling for any psychological harm experienced

Privacy and Confidentiality

Privacy is the right of pts to control the flow of information about themselves. Confidentiality is the right to have any personal information protected

HOWEVER - it may be difficult to avoid invasion of privacy when studying participants without their awareness (to avoid demand characteristics) + when publishing results confidentiality could be breached.

Guidelines on Privacy and Confidentiality

Pts. should not be studied without their awareness unless they are in a public space. Confidentiality is a legal right under the Data Protection Act, so if personal details are held (which is uncommon as most researchers maintain anonyminity by referring to pts by numbers or initials), they must be protected.

Cost-benefit Judgement

Made by an ethics committee that weigh up the cost of the research to the participants against the benefits of the study. This is done in order to decide if the study can be justified