AQA A Psychology 3

Notes on Research Methods 




IV- Controlled by researcher

DV- Change in DV caused by the IV

Extraneous variables- Can cause change in DV. If they are not controlled the results will be meaningless. Therefore researcher may not have tested what was intended to test. Therefore control is vital.�Researcher�should control as many EV's as possible.


Aim of a study is to get information on how people behave in real life. If study is too�artificial�participants will not act as they would normally.

Mundane realism refers to how experiment mirrors the real world.

1 of 48



Point of realism in research is to be able to generalise results to everyday life.

A study can have high realism but still lack generalisability.


Validity can be�separated�into internal and external validity.


Concerns what goes on inside study eg:

  • Whether IV produced change in DV or something else.
  • Whether researcher tested what he intended to test
  • Whether study�possessed�or lacked mundane realism.

To get high internal validity you must design research carefully, control EV's and test what is intended to test.

2 of 48



Is affected by internal validity- you cant generalise results of study that was low in internal validity because results are�meaningless.�

External validity concerns the ability to generalise.

3 of 48

Ethical issues

1. Informed consent- �Means revealing true aims of study. Might cause participants to change behaviour. From participants point of view they should be told what they will be doing so they can make informed decision. This is basic human right. Another problem is requirement for researcher to point out likely risks which are not always easy to predict.

2.Deception- Can be necessary to�deceive�about true aims otherwise participants may alter their behaviour. Distinction should be made between withholding some details for research (reasonably acceptable) and�deliberately�giving false info (less acceptable). Deception is unethical. Prevents participant given informed consent. Can also lead people to seeing psychologists as untrustworthy.

3.Right to withdraw- important for participants especially if they feel�uncomfortable�or distressed. However leaving study will bias results as people who remain are more likely to be more�obedient�etc. Sometimes they may feel unable to withdraw if they have been paid.

4 of 48

Ethical issues

4.Protection from harm- Studying some of the more important issues may involve a degree of distress. Its also difficult to predict outcome of certain procedures and therefore hard to guarantee�protection from harm. From participants point of view nothing should happen to them that causes harm. Acceptable if�risk�is no greater than ordinary life. Can be physical eg smoking or psychological eg�embarrassed. They should be in same state after as they were before.

5.Confidentiality- May be hard to protect as researcher may want to publish results. They can guarantee anonymity but can still be possible to work out who it is. Data Protection Act makes confidentiality a legal right.

6. Privacy- May be difficult to avoid invasion of privacy when studying someone without them knowing. However, people do not expect to be observed in certain situations eg in their own home, while they may expect it when sitting in a park.

5 of 48

How to deal with ethical issues


Tell�psychologists�what behaviours are not acceptable and how to deal with ethical issues. The rules are rather general as its impossible to cover every situation.


Committee must approve of a study before it begins. They weigh up benefits or research�against�cost to participants.�Problem�with this is that how are costs and benefits quantified? How much does personal distress cost?


If psychologists behaves in unethical manner they may decided to bar them from practising. Its not a legal matter so they wont be sent to jail.

6 of 48

How to deal with ethical issues

1. Informed consent:

How to deal with it- Ask them to agree based on comprehensive info and nature of study.

Limitations- If participant is given info reagarding nature of study they may try to change their behaviour which will bias results.

How to deal with it- Offer right to withdraw

Limitations- Will bias results�

7 of 48

How to deal with ethical issues

2. Deception

How to deal with it- Debrief after study and offer right to withdraw

Limitations- Cant turn back the clock. They may still feel�embarrassed

3. Right to withdraw

How to deal with it- Inform at�beginning�of right to withdraw

Limitations- May feel they�shouldn't�leave as it will spoil study.

Limitations- May feel they cant leave as they have been paid.

4. Protection from harm

How to deal with it- Avoid risks greater than everyday life

Limitations- Not always able to predict risks of taking part

8 of 48

How to deal with ethical issues

5. Confidentiality

How to deal with it- Should not record names, use numbers instead

Limitations- may still be able tyo work out who they are based on other info.


How to deal with it- Dont observe anyone without informed consent unless in public

Limitations- No universal agreement on what�constitutive�a public place.

9 of 48

Experiments and hypothesis


A hypothesis states what you believe is true. Its precise and testable statement of relationship between 2 variables.


Turning a hypothesis into a testable statement. Eg Turning "People remember more when they study in short bursts" to "Peope get more questions right on a test of recall wha they study in short burts (10 minutes at a time) than when studying for longer sessions (1 30 minute session)


Directional hypothesis states the expected direction of results eg people remember MORE. Used when there is previous research.

A non directional hypothesis predicts there will be a difference between 2 conditions. Used when there is no previous research.

10 of 48

A few other things


A pilot study is a small scale trial run of research design before doing the real thing. Done tin order to find out if certain things dont work.


An individual in a study who is not a real participant and has been instructed how to behave by researcher. They may act as the independent variable.

11 of 48

Experiment design


Each participant takes part in every condition.

Weakness- 1 condition may be more difficult than the other.

Dealing with weakness- You can make sure tests are equivalent.

Weakness- When participants do second test they may guess purpose of study.

Dealing with weakness- Use a cover�story�to prevent them guessing what study is about.

Advantage- Avoids problem of participant�variables.

12 of 48

Experiment design


Participants allocated to or more groups representing different experimental conditions.

Weakness- No control of participant variables ie different abilities�of them.�

Dealing with weakness-�Randomly�allocate which theoretically should�distribute�participant�variables evenly.

Advantage- Avoids order effect eg boredom effect.

13 of 48

Experiment design


Pairs are matched in terms of key variables eg IQ or age. One member of pair placed in experimental group other is placed in control group.

Weakness- Time consuming to match on key variables

Dealing with weakness- Restrict matching variables to make it easier.

Advantage- Avoids order effects

14 of 48

Extraneous variables

If EV's not controlled they bias results because change in DV may be due to EV rather than IV.

Things that are act as EV's:

  • Participant variables
  • Participant effects
  • Situation variables


Is any characteristic of individual. �Participant variables only act as EV's in independent groups design. In repeated measures�participant�variables are controlled.

Can be age, IQ, motivation or experience. Gender can also be one. Research suggests women are more compliant than men. So if more women than men in 1 condition, may mask effects of IV.

15 of 48

Extraneous variables


These are different from participant variables. May occur because�participants�seek cues about how to behave.

Hawthorne�effect- the fact that participants performance may change not because of IV but because they are responding to extra attention they get as research participants.

Social desirability bias- they wish to present themselves in best possible way and therefore may not behave as they normally would.

16 of 48

Extraneous variables


1 example of situational variable are order effects eg improved performance may be due to practice effect.

Another type could be time of day, temperature or noise.

Investigator effects- are cues from investigator that encourage certain behaviours. The way in which investigator asks a question may lead participant to the answer the investigator wants.

Demand characteristics- is an aspect of the research situation that triggers a predictable response in�participants, causing mist if not all of them to respond in similar way.�

Described as "the totality if cues which convey an experimental hypothesis to a�participant"�

17 of 48

Quantitative data analysis

Is any means of representing trends from numerical data, such as measures of central tendency.


Inform us about central value for et of data. They are averages. An average can be�calculated�in different ways:

The mean- Add up all numbers and divide by how many there are.

It makes use of all values but it can be�misrepresentation�of the data as a whole if there are extreme values.

The median- is the middle value in ordered list.

Its not affected by extreme scores but its not as sensitive as the mean because all values are not reflected in median.

18 of 48

Quantitative data analysis

The mode- the most common value.

Useful when data is in�categories�(nominal) but its not a useful way of describing data when there are several modes.

Nominal data- data are in�separate�categories�such as grouping people according to favourite footie team.

Ordinal- Data are ordered eg asking people to put a list of footie teams in order of liking.�


Ratio- There is true zero point.

19 of 48

Quantitative data analysis


A set of data can also be described in terms of how spread out the numbers are.

Easiest�measure of dispersion to use is the range.

More precise method of expressing dispersion is called the standard deviation. This is a measure of the spread of the data around the mean.

ADVANTAGES OF RANGE- Easy to calculate.

DISADVANTAGE OF RANGE- Affected by extreme values

ADVANTAGE OF STANDARD DEVIATION- More precise because all values taken into account

DISADVANTAGE OF STANDARD DEVIATION- May hide some of the characteristics of the data set.

20 of 48

Quantitative data analysis


Graphs provide a means of seeing your data.

Tables- numbers you collect�referred��to as raw data- numbers that�haven't�been treated in any way. They can be set out in table

Bar chart

Line Graph

Scattergram- used in�correlational�analysis�

21 of 48

Laboratory and field experiments

All experiments have 1 thing in common they all have and IV and a DV.

Lab experiment- conducted in special environment where variables are carefully controlled.�Participants�are aware they are being studied, though may not know aims of study.

Field experiment- More natural environment. As with lab experiment, the IV is manipulated.�Participants�are often not aware they are being studied.


Lab experiments are contrived:

  • They know they are being studied and this is likely to affect their behaviour.
  • Setting is not like everyday life. Low in mundane realism.

22 of 48

Laboratory and field experiments

Field�experiments�have less control:

  • May be more natural but this means more difficult to control EV's.
  • Major�ethical�issues- if participant�doesn't�know they are being studied.

Lab experiments make control easier but are less natural.

Field experiments are more natural but means there is less control and greater ethical problems.

23 of 48

Natural experiments

Third kind of experiment called natural experiment.

The environment is natural but the change in the IV is also natural.

Not true experiment because no one has deliberatley changed the IV to observe the effect this has on the DV.

Sometimes called quasi�experiments.

Used where you cant change the IV eg in studies of privation.


Drawing valid conclusions from natural experiments is a problem because:

Participants�are not�randomly�allocated�to conditions which means that there may be biases in the different groups of participants.�

Or, sample studied may have unique�characteristics.

24 of 48

Comparing lab, field and natural experiments


Advantage- Well controlled so high internal validity

Disadvantage-�Artificial,�participants�may not behave normally, Low internal validity


Advantage- Less�artificial�so higher realism so higher internal validity

Disadvantage- EV's less easier to control thus reducing internal validity


Advantage- Allows research where IV cant be manipulated eg privation studies

Disadvantage- Can only be used where conditions are very natural.

25 of 48

Observational methods and techniques

Naturalistic observation- behaviour studied in natural situation where everything has been left as it is normally.

Controlled observation- some variables controlled by researcher, reducing naturalness of behaviour being studied. Likely to know they are being observed.�


Naturalistic observation- Behaviour being studied is not interfered with but the researcher may be quite structured in terms of how observations are recorded.

Controlled observation- Strange situation involved structuring the behaviour of participants as well as the observers.

26 of 48

Observational methods and techniques


2 main reasons why observational techniques are hard:

Difficult to know what to record and what not to record.

Difficult to record everything happening even if you do select certain things to record.

Structured observations- Researcher uses various systems to organise observations such as:

  • Behaviour�categories- How to record behaviour your interested in.
  • Sampling procedures- who you are observing and when.

Unstructured observations- Researcher records all relevant behaviour but has no system. Behaviour to be studied is largely unpredictable. Problem with this is there may be too much to record. Also, behaviours recorded most likely to be those that are visible but they may not necessarily be important.

27 of 48

Observational methods and techniques


One of the hardest aspects of�observational�method is deciding how different behaviours should be�categorised.

To conduct systematic�observations, you need to break up the stream of behaviour into different�behaviour�categories. Eg when observing a baby you could have a list including things like smiling, crying and sleeping etc.

The behavioural categories should:

  • Be objective- observer should not have to make interferences about behaviour, but should�just�record actions.
  • Cover all possible component behaviours
  • Be mutually exclusive, meaning you should not mark 2 categories at 1 time.
28 of 48

Observational methods and techniques


When conducting a�continuous�observation,�researcher�should record every�behaviour�is as much detail as possible. However, sometimes continuous observation is not possible as there�would�be too much to record, therefore there must be a systematic method of sampiling observations:

  • Event sampling- counting number of times certain behaviour occurs
  • Time sampling- record behaviours in given time frame.
29 of 48

Designing observational research

In order to make systematic and objective observations, researchers need to develop behavioural categories. This is a coding system method or behaviour checklist.


A list of behaviours to be recorded during an observational study.


A systematic method for recording observations in which individual behaviours are given a code eg number for ease of recording


The observer rates each behaviour or�furniture�etc on 7 point scale

30 of 48



One way mirrors used to observe participants. They dont kow they are being observed. This is called covert.�

Knowing that you are being observed is called overt.


Both involve�naturally�occurring�variables that have not been manipulated by researcher. However, in natural experiment there is an IV and its effects on the DV is observed so we can draw conclusions. In�naturalistic�observations there is no IV.

31 of 48

Evaluating observational research

Like all research methods, observational methods also have advantages and disadvantages. Main�concerns�are related to validity and ethical issues and reliability.


External Validity- Observations likely to have high ecological validity because they involve more natural behaviours. Population validity may be a problem if for example children are only observed in middle class homes.

Internal validity- Observations wont be valid if coding system/checklist is flawed. The validity of observations also affected by observer bias- what someone observes is affected by their expectations.

Improving validity- By conducting experiments in varied settings with varied participants which makes findings more generalisable. Could also use more than 1 observer to reduce observer bias.

32 of 48

Evaluating observational research


Eg people being�observed�with their knowledge involves issues�relating�to informed consent. Some observations can be an invasion of privacy or involve decption

Dealing with ethical issues- ethics committee can be used to approve the study etc.


Refers to�whether�something is consistent. Any tool used to measure something must be reliable. eg a psychological test assessing personality. If the tool is measuring the same thing it should produce the same results every time. If the results�aren't�the same we need to be sure that its the thing that has changed and not our measuring tool.

Reliability of observations- Should be consistent eg 2 observers should produce the same record. The extent to which observers agree is called inter rater reliability.

33 of 48

Validity and reliability

Being reliable is being consistent. Being valid is being on target.

A study that lacks reliability will therefore lack validity, but you can havea �study that is reliable but not valid.

34 of 48

Evaluation of observational research


Gives more realistic picture of spontaneous behaviour. Likely to have high ecological validity.

What people say they do is�often�different from what they actually do. Therefore, observations may be more valid than�questionnaires.


Little or no control of EV's which means something�unknown�may account for the behaviour being observed.

If participants�dot�know they are�being�observed there are ethical�problems�eg invasion of privacy.

35 of 48

Selection of participants

If you wanted to do research to find out about schoolchildren in Britain, you couldn't use all school children in England.

Most obvious way of getting sample is to use people around at that time. Called opportunity sample. Most psychologists use a volunteer sample. Ideal sample is called random sample as it's less biased.

36 of 48

Selection of participants


Opportunity sampling- How? Ask people walking by in street.

Advantage- easiest method, takes less time.

Disadvantage- biased as sample taken from small part of target population.

Volunteer sample- How? Advertise in newspaper

Advantage- access to variety of participants eg all those who read paper

Disadvantage- biased as they are more likely to be more motivated (volunteer bias)

37 of 48

Selection of participants

Random sample

Advantage- unbiased, all members of target population have equal chance of being picked

Disadvantage- may end up with biased sample emir boys than girls because sample is too small.

38 of 48

Self report techniques

Observations are non experiment method.

Another type of non experiment method are called self reports. They can be interviews or questionaires.

Quesitonaire is written. An interview is given face to face.


Set of questions desgiend to collect info on specific topic.

2 main strengths are:

1. Can collect same info from large number of people easily.

2. You can access what people think. In observations you guess what they are thinking based on how they behave. In questionnaire you can just ask them. Whether they give you truthful answer is another question.

39 of 48

Self report techniques

Questions can be opened or closed.

Closed questions have range of answers to choose from. It produces quantative data. Answers easier to analyse than open questions.

Open questions are questions that invite participant to give their own answers. Tend to give qualaltive data.


Can be structured or unstructured.

Structured has pre determined questions.

Unstructured has less structure. Questions are developed as you go along. Start out with some general questions and let their answers guide you.

40 of 48

Evaluating self report techniques


External validity- Concerns extent to which answers from questionnaire and interview can be generalised. A major factor to this will be sample used to collect data.

Internal validity- Whether questionnaire or interview measures what intended to measure. Several ways to assess this:

  • Face validity- Does test look as if its measuring what's intended. Eg are questions obviously related to topic being researched.
  • Concurrent validity- compare results to a previous test on same topic. Particpants take both tests and 2 scores are compared.
41 of 48

Evaluating self report techniques


Decption- true aims may be hidden to get truthful data

Harm- May feel distressed by certain questions or sensitive topics

Privacy- Questions may be related to personal issues.

Confidentiality- must be respected, names and details should not be revealed.

42 of 48

Evaluating self report techniques


Internal reliability- Eg questions on IQ should all be assessing IQ.

External reliability- Eg if researcher conducted an interview and did the same interview a week later he should get same results.

43 of 48

Comparing questionnaires and interviews


Advantage- Easily replicated. Info form large numbers. Cheap.

Weakness- Answers many not be truthful. Social desireability


Advantage- Easily replicated because questions are standardised.

Weakness- Affected by interview bias


Advantage- More detailed info

Weakness- Affected by interview bias.

44 of 48

Studies using correlational analysis

Positive correlation- 2 variables increase together

Negative correlation- 1 variable increases but the other decreases

Zero correlation- no relationship


If significant correlation futher research can be done

If no correlation you know not to futher research it as there is no link


Misinterpret to thinkign they have found a cause and effect

May be other unknown variables affecting score

45 of 48

Case studies


He had epilepsy, so had hippocampus removed to reduce effects. No longer able to make new long term memories. Personality and IQ remained and he could remember things from before operation

Ethical issues- if he had no real memory for things that happend to him than how could he give consent to be studied. He did not understand what was being asked of him.

46 of 48

Evaluation of case studies


  • Rich in depth data that could be overlooked using other methods
  • Used to investigate human experiences that are are are and ones that you couldn't normally do eg child being locked in a room.


  • Hard to generalise as everyone is unique
  • Often nessecary to use recolition of past events and such evidence can be unreliable.
  • Important ethical issues


Cant make before and after comparisions

47 of 48

Comparing quantitative and qualitative data


Advantages-Easier to analyse 

Weakness- Oversimplifies reality and human experiences


Advantages-  Gets info you may not get using closed questions

Disadvantages- More difficult to detect patterns

48 of 48


No comments have yet been made

Similar Psychology resources:

See all Psychology resources »See all Research methods and techniques resources »