Research Methods

Experimental Method

The manipulation of the IV so that there is an effect on the DV

Aim

A statement outlining what the researcher intends to investigate

It also states the purpose of the study

Hypothesis

A statement that states the relationship between the IV an DV

Alternative - Shows relationship

Null - Shows no relationship

Directional Hypothesis

Hypothesis stating the direction of relationship between variables

E.g. The more sleep a participant has the better their memory performance

ONE TAILED

Non-Directional Hypothesis

A hypothesis that doesn’t state the direction of relationship between variables

E.g. The difference in the amount of sleep of the participant, will mean that they will have varying performance scores

TWO TAILED

IV

Aspect that has been changed to have an effect on the DV

DV

Aspect being measured in the study and caused by changes to the IV

Operationalisation

Defining the variables in terms of how they are measured

E.g. Participants getting at least 4 hours of sleep will get better scores than those who get less than four hours of sleep

(Add a quantifiable value to the hypothesis)

Extraneous Variables

Any variables that is not the IV affecting the DV if it is not controlled

Confounding Variables

Variables that can influence the DV and affect the IV therefore cannot tell if the change is due to the IV or extraneous variables

Demand Characteristics

Cues from the researcher or the research situation that can make the participant feel as though they know the aims of the investigation causing them to change their behaviour rather than acting naturally

* please you effect
* screw you effect

Investigator Bias

Unwanted influence from the researchers behaviour (conscious or unconscious) that can have an effect on the DV

Randomisation

The use of chance to decide the experimental method and conditions to control effects of investigator bias

Standardisation

Using the exact same formalised procedures and instructions for all participants in the study

Including standardised instructions

Lab Experiment

Experiment in a controlled environment where the IV is manipulated and the effects on the DV is recorded whilst maintaining the extraneous variables

Strength

* High degree of control
* Replicability
* High internal validity

Limitations

* Lack generalisability
* Low external validity
* Low mundane realism
* Participants may react with demand characteristics

Field Experiment

Experiment that takes place in a natural setting where the researcher manipulated the IV and measures the effects of the DV

Strength

* Higher mundane realism
* High external validity

Limitations

* Ethical issues

Right to withdraw, Informed consent, Invasion of Privacy
* Loss of control over extraneous and confounding variables

Difficult to establish cause and effect
* Difficult to replicate

Natural Experiment

Experiment where the IV is not manipulated by the researcher (and would of occurred had they not been there) and records the effect on the DV

Strength

* Provide research for unethical or events that wouldn’t have occurred otherwise due to practicality

E.g. Romanian Orphans
* High external validity

Limitations

* Not replicable
* Cannot control extraneous and confounding variables

Quasi Experiment

Experiment where no IV has been established instead it naturally occurs (age, gender)

Strength

* Carried out in controlled conditions
* High degree of control 
* Replicability 


* High internal validity 

Limitations 

* Lack generalisability 
* Low external validity 
* Low mundane realism 
* Participants may react with demand characteristics 
* Hard to establish cause and effect

Opportunity Sampling

Participants are conveniently recruited for being in the right place at the right time

Strength

* Time effective
* Cheap
* Convenient

Limitations

* Not representative of whole population

Cannot be generalised
* Researcher bias

Researcher may approach or avoid particular people

Random Sampling

All members of target population have an equal chance of being selected

Each person is allocated a number and then a random generator or lottery method is used to select the participants

Strength

* Eliminate researcher bias
* Higher internal validity

Limitations

* Difficult to obtain individuals information
* Time consuming
* Costly
* Volunteer Bias

Individual may not want to participate therefore end up with an unrepresentative sample

Systematic Sampling

Predetermined system where every nth numbered participant is selected

Strength

* Avoid researcher bias
* Numerical therefore objective

Limitations

* Difficult to obtain individuals information 
* Time consuming 
* Costly 
* Volunteer Bias 

Individual may not want to participate therefore end up with an unrepresentative sample 

Stratified Sampling

The composition of the sample reflects the target population with the same proportions of people in each subgroup (strata)

First identify strata’s, then calculate proportion of each needed based on the target population, select sample at random from each strata

Strength

* Avoids researcher bias
* Findings can be generalised

Because sample is representative to target population

Limitations

* Difficult to obtain individuals information 
* Time consuming 


* Costly 
* Volunteer Bias 

Individual may not want to participate therefore end up with an unrepresentative sample 
* The identified strata may not reflect all of the differences within the target population

Volunteer Sampling

Self selection where participants join in response to an advert or when they have been asked too

Strength

* Not time consuming
* Willing participants therefore more likely to cooperate in study

Limitations

* Volunteer Bias 

Individual may not want to participate therefore end up with an unrepresentative sample 

Sample may attract a particular person which means their more likely to adhere to demand characteristics

Cannot be generalised

Independent groups design

Participants only perform in on condition of the IV

Two groups experience one of two different conditions (the experimental condition or the control conditions)

Strength

* No order effects
* Demand characteristics eliminated

Less likely to guess the aim of the study

Limitations

* More participants than other designs to generate the same amount of data
* Different abilities of participants in various conditions may lead to changes in the DV

Solutions

* Random allocation ensures each participant has the same chance of being in the experimental condition or control condition

UNRELATED

Repeated Measures

All participants take place in all conditions of the experiment

Strengths

* Higher validity

As participant variables are controlled
* Fewer needed participants

Less time consuming

More cost effective

Limitations

* Order effects

May be bored/Fatigue therefore lead to performance deterioration

Confounding variable
* Demand characteristics

Ppt may figure out or attempt to guess the aim of the study therefore act in accordance to the please you or screw you effect

Solutions

* Counterbalancing

Half ppt do conditions in one order half do conditions in the opposite

Done using random allocation

RELATED

Matched Pairs Design

Ppts are matched together based on variable/s that are found to affect the DV and one will do one condition of the experiment whilst the other does another

Attempts to control extraneous and confounding variables

Strength

* No order effects
* Demand characteristics are reduced

Limitations

* Time consuming
* Expensive
* Large pool of ppts needed
* Difficult to know if ppts are matched on appropriate variables
* Only some attempt to reduce ppts variables as can never fully be matched

RELATED

Pilot Study

Small scale investigation done before the real investigation is done to identify potential problems and procedure modification

Single Blind Procedure

Research method where researchers don’t tell ppts if they are being tested in the IV condition or the control condition

Avoids demand characteristics and participation bias

Double Blind Procedure

Neither researcher nor experimenter knows who is in the IV condition and who is doing the control condition

* Prevent bias and demand characteristics

Placebo

‘Fake’ drugs ext

Control group/conditions

Sets the baseline of which results from the experimental condition will be compared to

* If greater difference researcher can conclude that the IV had an effect on the DV

Naturalistic Observation

Watching and recording behaviour in a setting where the behaviour would typically take place

Strength

* High ecological validity
* High external validity

Therefore can be generalised

Limitations

* Low ecological validity if ppt figure out their being watched
* Unable to replicate
* Uncontrolled confounding and extraneous variables
* Unable to establish cause and effect
* Observational bias

Will need to use inter-observer reliability

Controlled Observation

Watching and recording behaviour in a structured environment (e.g. labs)

Strengths

* Control of variables
* Can be replicated
* Able to focus on a particular behaviour

Limitations

* Unable to establish cause and effect 


* Observational bias 

Will need to use inter-observer reliability 
* Low mundane realism
* Low ecological validity
* Likely to be observing unnatural behaviour
* Likely to present demand characteristics

Overt Observation

Ppts know that they are being watched and that their behaviour is being recorded

Strength

* Ethical

Informed consent given

Limitations

* Unable to establish cause and effect 
* Observational bias 

Will need to use inter-observer reliability 
* Unnatural behaviour as know being observed
* Likely to present demand characteristics

Reducing validity

Covert Observation

Ppts are unaware they are being watched and recorded

Strength

* Natural behaviour
* High internal validity
* Reduced demand characteristics

Limitations

* Unable to establish cause and effect 
* Observational bias 

Will need to use inter-observer reliability 
* Unethical

Informed consent has not been given
* Invasion of privacy

Participant Observation

The observing researcher becomes part of the group in the study (e.g. senior management acting as a worker to understand working conditions better)

Strength

* Higher external validity

Study becomes more insightful

Limitations

* Unable to establish cause and effect 
* Observational bias 

Will need to use inter-observer reliability 
* Researcher may lose objectivity

Begin to identify with the ppts and lose sense of direction of the study
* Ppts behaviour may be unnatural if realise their being watched

Non-participant Observation

Researcher remains outside of the group being watched and recorded

Strength

* Researcher is more objective
* Higher internal validity

Limitations

* Unable to establish cause and effect 
* Observational bias 

Will need to use inter-observer reliability 

Recordings may be based on stereotypes
* May lose valuable insight of the study

Observer Bias

When the observers reports are bias to what they expect to see in the study

Solution = inter observer reliability

Inter Observer Reliability/Inter-Rater Reliability

The extent at which there is agreement between two or more observers involved in the observations of behaviour

Many researchers conducting the observational study and reports are compared and a score is calculated

Total #agreements/Total #observations x 100

Unstructured Observational Design

Continuous recording of observations the researcher will write down everything that they observe

Strength

* More detail and richness of recorded observations

Limitation

* Qualitative data

Harder to record and analyse
* Risk to observer bias

Only record behaviours that ‘catch the eye’ of the observer

Structures Observations

Researcher has a predetermined list of behaviours they are observing and the sampling methods they have chosen

Strength

* Systematic

Making the observation and recordings easier
* Quantitative data

Easier to analyse and compare to other data

Less risk of observer bias

Limitations

* Lack of detail
* Difficult to achieve inter observer reliability

Filling predetermined list is subjective

Behavioural Categories

Target behaviour which is being observed is broken into precise observable and measurable components

The categories are operationalised

Time Sampling

Behaviour is recorded only during the preestablished time frame

Strength

* Reduces # of observations
* Less time consuming

Limitations

* Small amount of data collected

May be unrepresentative to the observation

Event Sampling

Counting the # of times a particular behaviour is carried out

Strengths

* Good for infrequent behaviours

Less likely to be missed than when using time sampling

Limitations

* Important details of complex behaviour may be overlooked
* If behaviour is frequent there could be counting errors
* Difficult to judge the beginning and ending of a particular behaviour

Correlation

Mathematical technique used to show the association between two variables

Correlation is not causation

Correlation Coefficients

The value determining the strength and relationship between two variables

* Negative >0 (as one increases the other decreases)
* Positive

Strengths of Correlation

* Assess relationship of co-variables before conduction of experimental study
* Quick
* Cheap
* Uses secondary data making less time consuming
* Precise
* Quantifiable

Limitations of Correlation

* Difficult to establish cause and effect
* May be a third variable affecting the co-variables
* May be misinterpreted or misused

Co-variables

Variables investigated within the correlation (e.g. height and weight)

Qualitative Data

Non-numerical, expressed in words

Strengths

* Depth and richness of response
* Higher external validity

Ppts can expand and develop points/opinions
* Greater insight to ppt views

Limitations

* Difficult to analyse
* Difficult to make comparisons to other data
* Conclusions rely on subjective interpretations of the researcher

Therefore subject to interpretative bias

Quantitative Data

Numerical - Ordinal, Nominal or Interval

Strength

* Statistically analysed
* Can be graphs/charts
* Easy to make comparisons
* Objective
* High internal validity
* High reliability

Limitations

* Lack of detail
* No insight into ppts
* Low external validity
* Cant expand on views/opinions

Primary Data

Obtained first hand by researcher for the investigation

Strength

* Targets exact information the researcher needs
* Data fits aims and objectives

Limitations

* Time consuming
* May be expensive
* Needs planning, prep and access to resources

Secondary Data

Information not collected by the researcher yet used in their study

Strengths

* Inexpensive
* Minimal effort to collect

Limitations

* Data may be outdated or incomplete
* Data may not be reliable
* Challenges the validity of the data

Meta analysis

When researchers combine the results from many different studies and uses the data to form an overall view of what they are investigating

Strength

* Larger sample size
* More generalisability
* High validity

Limitations

* Publication bias

Researcher may only select relevant studies leaving out ones with negative or non-significant results

Central Tendency

Any measure which calculates the average value in a set of data

* Mean
* Mode
* Median

Measures of Dispersion

Any measure calculating the variation in a set of data

* Range
* Standard deviation (SD)

Standard Deviation (SD)

Low SD means data is closer in values therefore a less spread

High SD means a greater dispersion/spread of data

* There may be anonymous results

Types of displayed data

Bar charts

* Frequency of each variable is represented by the height of the bar

Histograms

* Continuous data
* X axis represents the category Y axis represents the frequency

Line Graphs

* IV plotted on the X axis, DV plotted Y axis
* Points connected by lines to show the change of values

Scattergrams

* Show association between co-variables rather than differences

Normal Distribution

Symmetrical pattern of frequency data forming an inverted u shape

Skewed Distribution

Spread of frequency which is not symmetrical instead clusters at one end

* Positive skew - distribution concentrated towards the Y axis
* Negative skew - distribution concentrated towards the X axis

Peer Review

Assessment of scientific work done by experts in the same field to ensure research that is intended for publication is done to a high standard and quality

Main Aims of Peer Review

1) Allocation of funding - to see if giving funding is worthwhile

2) Validate the relevance and quality of research - assess all elements of research for quality and accuracy and to see whether it is fraudulent

3) Suggest amendments or improvements for the study

Limitations of Peer Review

Anonymity

* Reviewers could use anonymity to criticise rival researchers
* May lead to bias
* Some journals may do open reviewing where the reviewers names are made public

Publication Bias

* Negative results or controversial studies/results may not be published
* Creates a false impression on current psychology

Needs an expert

* Smith (1999) argues that poor research has been passed because the reviewer hasn’t understood the work

Suppression of Controversial Research

* New and challanging research usually rejected
* Slowing the rate of change in scientific fields

Economic Implications

How will research affect, benefit or devalue national finance

Economical Implications in Psychopathology

Treatment

* CBT
* Rational Emotive Behavioural Therapy
* Drugs

Workers can return to work

Economical Implications in Attachment

Role of the Father

* Field (1978)
* Fathers can take on role as primary caregiver

Mothers can return to work

Flexible working arrangements within families

Maximise income

Economical Implications in Social Influence

Leading to Social Change

* Minority Influence
* Normative Social Influence (NSI)
* Disobedient Models

Union Strikes

Environmental Campaigns

Health Campaigns

Economical Implications in Memory

Eyewitness Testimony’s

* Leading questions and post event discussion may affect the persons testimony

Reduce wrongful convictions

Less people in jail

Reduces money needed for expert resources

Case Studies

A detailed analysis of the life of a person showing the individuals past and present behaviours

Strengths of Case Studies

* Detailed
* Generate hypothesis for future research
* Ethical and Practical
* Study cases that you would not be able to use otherwise

Limitations of Case Studies

* Not generalisable
* Social desirability bias (from individual)
* Interpretative bias (from researcher)
* May rely on memory (often could be inaccurate)
* Time consuming
* Unable to replicate
* Low validity

Internal - unable to replicate

External - unable to generalise

Content Analysis

Observational research where there is indirect analysis of behaviour through communication

E.g. Conversations or Media

Conducting Content Analysis

1) Identify Hypothesis

2) Create coding system - categorise important units e.g. counting a particular phrase

3) Gather data and record in table

4) Analyse data using Thematic Analysis

5) Write scientific report

Thematic Analysis

Qualitative analysis of data where themes are identified

Thematic Analysis Themes

Implicit or explicit ideas that are recurrent

Themes can be developed into broader categories such as stereotyping

Strength of Thematic Analysis

* High external validity
* High mundane realism
* Large set of qualitative and quantitative data
* High replicability
* Data in public domain therefore eliminate ethical issues

Rights of privacy

Confidentiality

Informed consent

Limitations of Thematic Analysis

* Observer Bias

Eliminated by inter-observer reliability
* Investigative Bias

Choice of what data to be analysed
* Interpretative Bias

Researcher may ignore somethings and pay high attention to others
* In public domain so lacks originality

Nominal Data

Categorical data

E.g. Yes, No, Mabey

Central Tendency Measure = Mode

Dispersion Measure = None

Ordinal Data

Ranking data/Ordered data

E.g. 1-10,

Central Tendency Measure = Median

Dispersion Measure = Range

Interval Data

Public scales of measurement where the units are precisely defined

E.g. 1cm = 10mm

Central Tendency Measure = Mean

Dispersion Measure = SD

Abstract

Key details of a report

* Aim
* Hypothesis
* Method
* Results
* Conclusion

150-200 words

Introduction

Links study to previous research on a similar topic

Method

Description of what researcher did so others can replicate

* Design
* Sample (demographics, sample method and target population)
* Materials/Apparatus
* Procedure
* Ethics

Results

All finding from the study presented with inferential and descriptive statistics

Inferential Statistics

Statistical testing and significands

Outlines which hypothesis was accepted or rejected

Descriptive Statistics

Tables, graphs, charts measures of central tendency and dispersion

Discussion

Summery of results in a verbal format

Identify limitations and how future studies can address this

Wider implications of study - i.e. real world app

Referencing

Full details of sources cited in report e.g.

* Author
* Date
* Article Title
* Journal Name
* Volume
* Page Number

Statistical Tests

Determine whether hypothesis should be accepted or rejected

Find if relationships and/or differences of variables are due to chance or are significant \`

Sign Test

Analyse difference in scores between related items

Nominal

1) Difference not association

2) Related experimental design

3) Nominal data

Conducting Sign Tests

1) State hypothesis (alternative and null)

2) Convert to nominal data (-) if data decreased in second condition, (+) if data increased in second condition, if value is the same it will be excluded

3) Add up + or -

4) Less frequent sign is S (S = calculated value)

5) Compare calculated value with the critical value

If S< or equal to the critical value it is significant (5

Choosing Inferential Statistical Tests

1) Study design (related/unrelated)

2) Type of data (nominal, ordinal, interval)

3) Is difference or correlation being measured

Calculated Value

Obtained from the results of the study

Critical Value

Numerical boundary worked out from probability based of degrees of freedom or N value, P value and hypothesis type to find what hypothesis should be accepted or rejected

Rule of R

If R in the name of statistical test then:

Calculated value > or equal to Critical value to be significant

If no R in name:

Calculated value < or equal to Critical value to be significant

Probability

The likelihood of an event occurring

0 = statistically impossible

1 = statistical certainty

Therefore p value is 0.05 (5%) in psychological research

Significance

How sure researchers are about the correlation or difference existing

Null (H0) = no difference or correlation

Alternative (H1) = this is a difference

Type I (optimistic) Error

Null is rejected and alternative hypothesis is accepted when the null is ‘true’ (failure to accept the null that is true)

Finding of significance when there isn’t any (false positive)

Type II (pessimistic) Error

Failure to reject the null that is false

Find no significance when there is on (false negative)

Spearman’s Rho

2) Ordinal

3) Correlation

Conducting Spearman’s Rho

1) Rank each set from lowest to highest (if