Investigation Design + methods + techniques

Investigation Design+ methods + techniques

HideShow resource information


There are several methods of research that a physcologist can employ to collect data regarding event/ set of behaviours. These include experimental and non experimental methods

1 of 48


Experimental methods:

  • laboratory experiments
  • field experiments
  • natural experiments

Non-experimental methods:

  • interviews
  • surveys and questionnaires
  • observations
  • correlational studies
  • case studies
2 of 48


The 1st step in designing a research investigation involves; identifying a topic or an issue to study and review the background literature on that area, e.g. what research has been done before?

Once this is done the research in a position to identify the aim and the hypothesis of the investigation.

AIM- intend to find out

HYPOTHESIS- what to expect to find out

apparatus, method/ procedure

3 of 48

Types of hypothesis

A hypothesis is a testable statement which may include a prediction, it is NOT a theory!


Directional hypothesis one tailed- More precise to non-directional and specifically states the direction of the results.

Non-directional hypothesis*two-tailed* - Direction of results not predicted, results is specified but could go either way.

Null hypothesis- This tends to state that there will be no relationship between the variable being investigated.It is used because it make a very precise prediction (nothing will happen!) that can be disproved, thereby supporting the alternative hypothesis.

4 of 48


Theories are complex sets of statements which when put together offer a way of explaining an observed behaviour/ phenomenon.

5 of 48

variable overview

Variables are anything that may change/ alter in anyway. The researcher will control/ manipulate/ observe variables which are central to the research.

Psychologists need to be able to define variables successfully if their research is to be treated as scientifically worthwhile.

Once the definition has been decided you must than compare ideas with other members of team.

With such variables as stress/ concentration levels, these are even more difficult to define. it could be visible signs of the effect on p's or measuring the effects on a aspect of behaviour.

6 of 48

Operationalizing Variables

The variable has to be refined and narrowed down to be used as a research focus.

Once null and alternative hypotheses have been wrote, the researcher has to operationalise it. What this means is that this means is that the exact nature and means of measuring/  observing the variables to which the hypothesis refers to must be defined clearly and objectively, the researcher is said to have now produced opernantionalized definitions.

7 of 48

Relability and validity

Reliability- 2 or more measurements/ observations of some psychological event will be consistent.

Validity- You are actually measuring what you claim to be. 

INTERNAL- a research study/ experiment has internal validity if the outcome is the result of the variables that are being manipulated.

EXTERNAL- to extent to which findings can be generalised to setting other than where researched; includes population validity ( if findings can be generalised to other people) ecological validity (other setting)


face validity- most basic. involves 'eye ball ' does test look as if it is measuring what it should be!

concurrent validity- comparing results from old to new test.

Predictive validity- Ability of test to predict performance on future tests, if can GOOD!!

8 of 48

Ethical issues



Right to withdraw

Prevention of physical and psychological harm



9 of 48


a. protect the p's and their welfare in the present and future.

b. protect the field of psychological study and not bring to dispute

c.protect the researcher so that does cause harm to p's

10 of 48


The British Psychological Society:

Guidelines for research with human p's was developed after concerns were raised about the lack of consent etc.there are 9 different aspects of ethic that relate to research with humans, p's and they all come under the heading ETHICAL PRINCIPLES FOR CONDUCTION RESEARCH WITH HUMAN P'S

11 of 48


Consent- researchers should indicate the aims of the research v.clearly and all aspects should be disclosed. p's should given written consent and if under 16 from parents/ guardians.

Deception- No information can be withheld from the p's nor should they be mislead also to avoid them to withdraw alternatives should be considered.

Debriefing- During the investigation, p's should always be informed with what the research is about.

Withdrawal from investigation- P's have right to withdraw at any time, they should be informed of this prior to investigation, should a p withdraw at a later stage/ after study been conducted, this would mean that the researcher must destroy any data/ information collected from the p.

Confidentiality- P's have right to this, if it cannot be assured than the p must know prior to investigation. The data protection act requires you to maintain confidentiality of the information collected from p's.

12 of 48


Protection of p's- Psychologists have to ensure the p's have not been put through physical/ psychological harm due to his research includes stress. The risk must not  be greater than p's experience daily.

Observational research- Observational studies have to protect the privacy and psychological well being of the p's; p's must not be observed in places they usually wouldn't.

Giving advice- If a problem arises during the research either psychological or physical to the p than consult professional advice.

Colleagues- Responsible for ethical conduct of research + colleagues. if you fell a colleague is acting unethically should raise concern with them!

13 of 48

How to deal with deception and young


Debrief- after a research the true aim is fully told so p leaves research study as came in.

Retrospective informed consent- Get consent before research so that they known they will not be debriefed till after.

Informed consent- YOUNG CHILDREN - ADULTS and SPECIAL NEEDS as do not fully understand what they are participating in this impacts their ability to give informed consent.

YOUNG ETC- Prior general consent- before research get consent that p is fine than assume for rest research fine to be debriefed

Presumptive consent- random sample of population introduce to research if fine with being deceived assume rest would.

14 of 48

How to deal with protection of p's

remind right to withdraw

researcher responsible to end research if too much harm.


15 of 48

Experimental Research

true experiment has 3 key features: manipulate the IV to see if it would change the DV, all other variables that might influence the results should try be kept constant / eliminated.


  • can control all other variables and so easier to notice difference with IV and DV
  • can create conditions wanted; whenever, wherever so not to wait for it to naturally occur.
  • generates quantitative data (number measures of DV) can be analysed using inferential statistics which allows researcher to see how likely results by chance or by manipulated IV
  • p's usually know taking part so have demand characteristics
  • only if method of sampling brings representative sample can it be generalised.
16 of 48

Experimental Designs- independent

The independent groups design

what is it? the p's either take part in control / experimental condition.

When to use it? this design is not effected by the no. of p's, although in a small sample there's a risk that any differences between conditions could be due to individual differences of p's

  • No order effects
  • Reduced chance of demand characteristics
  • can use the same stimulus materials
  • Least effective design for controlling p variables
  • more p's required.

Controls - Absolutely crucial that p's are randomly allocated to the condition

17 of 48

Experimental Designs- Repeated

Repeated measures design

what is it? every p takes part in both conditions of the iv ( so they the control)

When to use it? when there is only a small no. of p's- the same p's can be used for both conditions

  • p's variables are eliminated
  • uses fewer p's; it is sometimes difficult to get people to participate in a research
  • order effects
  • increased chance of demand characteristics
  • cannot use the same stimulus materials

Controls -  Order effects can be controlled through counterbalancing ( the method used to balance order effects. 1/2 the p's would complete the experiment in 1 sequence, and other other way ABBA.) Demand characteristics - use single blind technique.

18 of 48

Experimental Designs- matched

Matched p's design

what is it? each p in 1 experimental conditions is matched as close as posb. to a p i other condition; could be on age, gender, iq, personality traits. when matching then randomly allocated 1 in 1 condition other in other.

When to use it? when time is not of essence and have money etc.

  • no order effects
  • good attempt at controlling p variables
  • hard to exactly match
  • more p's required

Controls -  Mono-zygotic twins (identical) provide researchers with a v.close match for p variables

19 of 48

Choosing a target population

A target population is a group of people who share a given set of characteristics about which a researcher wishes to draw conclusion.

Use sample - if larger sample more likely to provide an accurate estimate about nature of population and if small it is practical & save time + money.

A sample is a part of the population selected that is considered to represent the population as a whole

20 of 48

Types of sample - volunteer


What is it? -> The p's self-select, they volunteer to be part of the research.

How to do it? --> Advertise the research, and whoever responds is part of the research. 

Population validity --> research shows there is a particular type of people who volunteers so LOW population validity.

21 of 48

Types of sample - random


What is it? -> A sample in which every member of the target population has an equal chance of being selected.

How to do it? --> Every member of the target population is identified & random sampling technique is used to select sample. e.g. names are numbered & picked out of a hat.

Population validity -->HIGH representative sample + population validity.

22 of 48

Types of sample - opportunity


What is it? -> Whoever is available to participate.

How to do it? --> Approach people & ask to take part, take advantage of whoever is willing + available to take part.

Population validity --> HIGH chance sample will be biased!! so LOW population validity.

23 of 48

Labortary Experiment

Lab experiments provide psychologists with the highest levels of control over variables & are widely used in psychology, this is where technical equipment can be used + accurate measurements can be made.

E.g --> The Stroop effect, Milgram's study of obedience, Harlow's study of attachment in monkeys.

Ethical Issues

  • Consent- a fully informed consent is needed from all p's. All p's should have the right to withdraw at any point of the experiment should they feel uncomfortable.
  • Deception- If deception is needed to be used. e.g avoid demand characteristics the p's should offer to be fully debriefed afterwards.
  • Use of animals- as can't obtain consent the animals must NOT be subjected to unnecessary suffering.
24 of 48

Labortary Experiment -evaluation



  • Replicability of procedures - if experiment is well carried & clearly reported procedures than it can easily be repeated is similar results are gained than confidence in results are raised.
  • Control over variables - easier to control potential confounding variable in lab so high levels of precision, also if all variables other than I.V controlled / eliminated successfully than cause & effect can easier be established.
25 of 48

Labortary Experiment -evaluation part 2


  • Less valid- the high levels of control can make the experiment artificial, it is less ecologically valid as the findings often can't be generalised to the public.


  • Demand characteristics- the p's may try work out the aim in the experiment and act accordingly so findings less valid. Another problem is evaluation apprehension; this is where a p would get concerned over what the researcher may find out about them. Also social desirability can be a problem this is where a p changes hows they act daily to be perceived more favourable.
26 of 48

Field experiment

Field experiments are carried out in natural environments. The I.V is manipulated by the researcher to produce the change in D.V. A field experiment is used in situation where it is considered necessary.

E.g --> Mating success of long-tailed widow birds, bonding between mother & infant.

Ethical Issues

  • Consent- a fully informed consent is needed from all p's. All p's should have the right to withdraw at any point of the experiment should they feel uncomfortable.
  • Confidentiality- Identify of any organisation & p's involved must be protected, could just remove name & use descriptive characteristics, where confidentiality is not secure p's must be warned.
  • Use of animals- Where animals used their natural environment is being altered due to the very nature this leaves great responsibility for researcher to minimize the disruptive effect of any manipulation.
27 of 48

Field Experiment -evaluation



  • Improved ecological validity - this means that by avoiding the artificiality of the lab the researcher can eliminate the common criticism of lab experiments & can generate findings to similar studies.
  • Reduction of demand characteristics-  p's may be unaware they are taking part in research, so act naturally minimizing any demand characteristics.
28 of 48

Field Experiment -evaluation - part 2


  • Establishing controls- difficult to not only control with I.V and D.V but also confounding variables. e.g. p's can intrude through conversation / simple presence.
  • Generalising to other situations- although realism is high, results cannot be generalised to other real-life situations that differ from the one which the experiment took place.
29 of 48

Natural expirment

In natural experiments the researcher exploits naturally occurring differences in the I.V; the researcher does NOT directly control the I.V. This could be seen as a quasi-experiment ( not true experiment). Purists could regard it as non-experimental

E.g --> A twin study, adoption study.

Ethical Issues

  • Consent- a fully informed consent is needed from all p's. All p's should have the right to withdraw at any point of the experiment should they feel uncomfortable.
  • Protection of p's- making sure leaving p's not in stress.
  • Confidentiality- right to expect the information they provide will be treated confidential & if published not identifiable that is theirs, if cannot guarantee it should be warned prior research.
30 of 48

Natural Experiment -evaluation



  • Lack of direct intervention - experimenter does not intervene directly although his/her presence may affect p's behaviour.
  • Reduction of demand characteristics- unaware in experiment so demand characteristics are avoided.


  • Loss of control- IV not directly controlled by researcher so reduces likelihood of cause & effect being established.
  • Likelihood of the desired behaviour being displayed- the naturally occurring situation that the researcher wishes to study may occur only rarely, reducing the available opportunities for research.
31 of 48

case study

The gathering of detailed information about individual/ group. Focuses on single case so its ideographic in its nature. Typically involved production of case history, relevant details. Other information can be gained through interviews, surveys and observations by nature very individualistic. Method generally descriptive & focuses on research

E.g --> Baby P, Maddie Mcann

How information is found

  • Retrospective information- not reliable, lacks detail, poor recall, could lie, confused recall
32 of 48

case study -evaluation



  • Produces rich, meaningful data
  • High levels of ecological validity so realism is high
  • Can challenge established thinking & lead to new psychological insights


  • Difficult to replicate so hard to establish the reliability of data
  • As its ideographic nature hard to generalise results beyond the study so population validity
  • Possibility of researcher bias, which further calls into question scientific credibility.
33 of 48

correlation study

Psychologists sometimes want to know/find out a relationship between 2 variables

E.g --> Smoking & cancer, Aggression in kids & violent tv.

Positive & Negative Correlation

The term correlation means to co-relate & refers to the relationship between 2 variables:

  • Positive- as 1 variable increases so does other
  • Negative- as 1 variable increased other decreases.
34 of 48

Recognising + measuring correlation

Recognising + measuring correlation

Correlational relationships are identified and shown through use of statistical techniques. So variable have to be operationalised (numeric value), The reason correlation studies rely on quantitative data as they actually measure strength + direction of relationship.

  • Scatter-gram- produces visual presentations, it is descriptive statistic & illustrates correlation
  • Correlation coefficient-numeric representation of strength and direction of relationship between 2 variables
35 of 48

correlation study- evaluation



  • Can establish relationship between 2 variables
  • Allow researchers to statistically analyse situation that could not be manipulated experimentally for ethical/ practical reasons


  • Does not establish cause & effect only establishes that there is a relationship, could be 3rd unidentified variable that affects the relationship
  • Only identifies linear relationships not curvilinear. E.g. aggression & temperature is curvilinear ( only positive to point than decrease, inverted U relationship not detected by correlation.)
36 of 48

Surverys and Interviews

Surveys involved asking a large sample of people for information on a specific topic at a particular point. Through asking questions the researcher usually wants to establish what people do & think in relation to the topic. Places great emphasis on selecting a representative sample (in order  to generalise findings).When considering how to design questionnaire consider their type, easy to understand, amount of information asked to minimum, ethical and pilot study first!

37 of 48

Types of questions used in surveys

Open question- allows respondent to write their own answer in words (QUALITATIVE)

  • Rich, detailed data
  • Allows respondent to express what they really think
  • more realistic
  • cause of qualitative nature hard to analyse.

Closed question (fixed choice)- p chooses their responses from a limited number of fixed responses, predetermined by researcher

  • Quantitative data can be statistically analysed
  • Artificial, questioning ins not realistic
  • not as rich
  • not clear if respondent understood question
38 of 48

survey - evaluation



  • Can be used to question large sample of people relatively quickly
  • Used to collect large amounts of data about what people think as well as what they say and do
  • Efficient in that researcher does not need to be present while p complete questionnaire so also reduces investigator effects and reduction in the influence of interpersonal data.
39 of 48

survey - evaluation part 2


  • Social desirability - person fakes answer to appear more favourable (in particular sensitive data), if cannot rely on authenitcy is major threat to validity of data.
  • Could be distortion in sampling range affect if findings can be generalised.
  • Postal surveys have low response rates reducing representativeness of sample - reduce validity.
  • Hard to phrase question clearly; may obtain different interpretations of questions.
40 of 48


Alternative method to asking questions! Differ from questionnaires as face to face. Useful for gathering more detailed information enabling more natural & flexible approach to questioning. Generally involves conversation exchange. Interviewer usually has certain topic they want to explore, interview can be highly structured or very open.



  • Detailed information obtained, allow interviewer to clarify meaning & significance of information obtained.
  • P can freely express themselves.
  • If unstructured can encourage the p to be more honest.
41 of 48

interviews - evaluation


  • Statistical analysis can be difficult as qualitative data.
  • More time consuming
  • Greater chance of interpersonal variables affecting responses
  • Increases investigator effects.
42 of 48

Observational Methods

Humans are constantly looking round at each other, observing behaviour of people in their environment & either comparing themselves with others or imitating what they see. Psychologists observe in a much more scientific way & will focus precisely on particular categories of behaviours / events. these observations are usually part of a research plan.

Observer Bias- This happens when a observer makes their own particular interpretation of the behaviour they observe, steps & measures must be taken in order to avoid this.

43 of 48

Other factors affecting Observational Methods

Observations if carried out by a single researcher will obviously produce bias, even different psychologists working on same research will possibly place their own interpretation on what they see. Various strategies can be used to increase reliability.

  • Use double blind technique: person completing research should be unaware of hypothesis.
  • Clear definitions of terms & concepts.
  • Pilot study enables observers to practise the use of the definitions & check all observers categorising behaviour in the same way.
  • Ensure the observation is completed by more than 1 observer & compare recordings, if quantitative data gathered than inter-observe reliability can be measured using Correltational techniques, if proven reliable & consistent, expect positive correlation between scores.
44 of 48

Types of observation


  • High levels of ecological validity.
  • P's may be unaware that they being observed so behave more naturally.
  • No control over E.V's
  • If p is unaware that being observed may raise ethical issues.


  • Higher control over E.V's
  • P's may be affected by fact that they know they being observed so not natural lowering ecological validity.
45 of 48

Types of observation part 2


  • V.high ecological validity.
  • Rich qualitative data
  • Easier to understand what observes' behaviour actually means.
  • Relationship based on trust can me made between those being observed & observer.
  • Researcher has to be rely on memory- unreliable
  • Researcher can get emotionally attached so observations can become subjective rather than objective.
  • Presence of observer can change group dynamics.
46 of 48

Types of observation part 3


  • Observes may not realise being observed, so behave naturally.
  • The observer may be more objective as less likely to become emotionally attached.
  • Can record observations as occur so more reliable in terms of memory.
  • The actual meaning of behaviour can be unclear from distance.
  • No relationship formed so observer & observe may have less trust.


  • Reduction in ethical issues as consent gained.
  • Increases trust between researchers & the public.
  • Increase in reactivity as observe may change behaviour as known being observed (demand characteristics)
47 of 48

Types of observation part 4


  • Reduction of reactivity as the p may not realise being observed, so less demand charateristics
  • Ethical issues raised.
  • When observee realises been observed may distrust psychologists in the future.
48 of 48


Manpreet Teji

why doesn't the pdf work? :S

Similar Psychology resources:

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