Research methods

• experimental methods concern the manipulation of an IV to have

  an effect on the DV, which is measured and stated in results. These

  experiments can be: field, laboratory, quasi or natural.
• An aim is a general statement made by the researcher which tells us what they plan on investigating, the purpose of their study. Aims are developed from theories and develop from reading about other similar research.
• A hypothesis is a precise statement which clearly states the relationship between the variables being investigated. The hypothesis can either be non-directional or directional.

  • A directional hypothesis states the direction of the relationship that will be shown between the variables whilst a non-directional hypothesis does not. (directional hyp tends to be used when there has already been a range of research carried out which relates to the aim of the researcher’s investigation. The data from this previous research would suggest a particular outcome)
• The independent variable refers to the aspect of the experiment which has been manipulated by the researcher or simply changes naturally to have an effect on the DV which is then measured.

  The dependent variable is the aspect of the study which is measured by the researcher and has been caused by a change to the IV. All other variables that could affect the DV should be carefully controlled so that the researcher is able to confidently conclude that the effect on the DV was caused by only the IV.
• Operationalisation refers to the act of a researcher clearly defining the variables in terms of how they are being measured. This means the variables should be defined and measurable. The hypotheses states should also show this operationalisation
• In an experiment, the only aspect that should affect the DV is the IV. Any other variables that may interfere with the IV or the DV should be removed from the experiment or well controlled. Such variables can be confounding or extraneous. An extraneous variable refers to any other

  variable which is not the IV that affects the DV and does not vary systematically with the IV, they are essentially nuisance variables. Examples are the lighting in the lab or the age of participants - these variables do not confound the results of a study but just make them harder to detect.
• A confounding variable = a variable other than the IV which has an effect on the DV - they change systematically with the IV = it becomes difficult for the researcher to be sure of the origin of the impact of the DV as the confounding variable (not the IV) could have been the cause.
• Demand characteristics = any cue the researcher or the research situation may give which makes the participant feel like they can guess the aim of the investigation - can cause the participant to act differently within the research situation from how they would usually

  act. They may change their behaviour to fit the situation rather than acting naturally.

  ‘Please-U effect’ - may act in a way they think researcher wants them to act

  'Screw-U effect’ - may intentionally underperform to sabotage the study’s results

  affects validity of results
• Participant reactivity may also lead to investigator effects which refers to any unwanted influence from the researcher’s behaviour, either conscious or unconscious, on the DV measured (the research’s results). This includes a variety of factors :- the design of the study, the selection of participants and the interaction with each participant during the research investigation.
• Randomisation is the use of a chance to reduce the effects of bias from investigator effects - minimise the effects of extraneous or confounding variables. This can be done for the design of materials, deciding the order of conditions, the selection of participants e.t.c.
• Standardisation describes using the exact same formalised procedures and instructions for every single participant involved in the research process. This allows there to eliminate non-standardised instructions as being possible extraneous variables
• Lab experiment: Laboratory An experiment that takes place in an unfamiliar, controlled environment whereby different variables can be

  carefully controlled.

  + High degree of control- experimenters control all variables,the IV has been precisely replicated, leading to greater accuracy.

  + Replication-researchers can repeat experiments and check results.

  -Experimenter’s bias- this bias can affect results and participants may be influenced by these expectations.

  -Low ecological validity-high degree of control makes the situation artificial, unlike real life.
• Field An experiment conducted in a more natural environment,

  not in a lab but with variables still being well controlled.

  + Naturalistic -so more natural behaviours hence high ecological validity.

  +Controlled IV

  -Ethical considerations- invasion of privacy and likely to have been no

  informed consent.

  -Loss of control-over extraneous variables hence precise replication not possible.
• Quasi: An experiment whereby the IV has not been determined by the researcher, instead it naturally exists e.g gender difference studies.

  + Controlled conditions- hance replicable, likely to have high internal validity.

  -Cannot randomly allocate participants-to conditions so there may be

  confounding variables presented. This makes it harder to conclude that the IV caused the DV.
• Natural: IV would have happened even if the researcher hadn't been there e.g. if studying reactions to earthquakes (not manipulated)

  + Provides opportunities- for research that would have otherwise been

  impossible due to practical or ethical reasons.

  + High external validity- as you are dealing with real life issues.

  -Natural occurring events- may be rare this means these experiments are not likely to be replicable hence hard to generalise findings.

  -Very difficult to randomise- participants into groups so confounding

  & extraneous variables become a problem.
• sampling: The researcher needs to decide how they select participants to take part in their investigation.The

  population is a group of people from whom the sample is drawn.
• Opportunity sampling: Participants happen to be available at the time

  which the study is being carried out so are recruited conveniently.

  + Easy method of recruitment which is time saving and less costly.

  -Not representative of the whole population hence lacks generalisability.

  -Researcher bias is presented as they control who they want to select.
• Random sampling: when all members of the population have the same equal chances of being the one that is selected. method: each member of the population is assigned a number then either a random number table or a random number generator or the lottery method is used to randomly choose a partner.

  + No researcher bias- researcher has no influence on who is

  picked.

  -Time consuming- need to have a list of members of the population (sampling frame) and then contacting them takes time.

  -Volunteer bias- participants can refuse to take part so can end up with an unrepresentative sample.
• Systematic sampling: A predetermined system is used whereby every nth member is selected from the sampling frame. This numerical selection is applied consistently.

  + Avoids researcher bias and usually fairly representative of population.

  -Not truly unbiased unless you use a random number generator and then start the systematic sample.
• Stratified sampling: With this method the composition of the sample reflects the varying proportions of people in particular subgroups (strata) within the wider population.

  + No researcher bias- the selection within each stratum is done randomly.

  + Produces representative data due to the proportional strata hence generalisation is possible.

  -Time consuming to identify strata and contact people from each.

  -A complete representation of the target population is not possible as the identified strata cannot reflect all the differences between the people of the wider population.
• Stratified sampling method: Firstly you identify strat. Then you

  calculate the required proportion needed for each stratum based on

  the target population. Then select sample at random from each stratum using a random selection method.
• Volunteer sampling: Involves self selection whereby the participant offers to take part either in response to an advert or when asked to.

  + Quick access to willing participants which makes it easy and not time

  consuming.

  + As participants are willing to take part they are more likely to cooperate in the study.

  -Volunteer bias- they study may attract a particular profile of a person. This means generalisability is then affected.

  -Motivations like money could be driving participation so participants may not take study seriously, influencing the results.
• Independent groups design: The participants only perform in one condition of the independent variable (IV).

  + There are no order effects presented.

  + ppts are less likely to guess the aims of the study (demand

  characteristics are eliminated)

  
  

  - No control over participant variables whereby different abilities of

  participants in the various conditions can cause changes to the DV

  -You need more participants than other designs to gather the same

  amount of data.
• solution independent groups design: Random allocation solves the first limitation mentioned. This is as it ensures that each participant has the same chance of being in one condition of the IV as another.
• Repeated measures: The same participants take part in all conditions of the IV.

  + Eliminates participant variables.

  + Fewer participants needed, so not as time consuming finding and using them.

  -Order effects presented e.g. boredom may mean in second condition done participant does not do as well on task.

  solution: Counterbalancing- this is when half of the participants do conditions in one order and the other half do it in an opposite order.
• Matched pairs: Pairs of participants are first matched on some variable that has been found to affect the dependent variable (DV), then one member of each pair does one condition and the other does another.

  + No order effects.

  + Demand characteristics are less of a problem.

  -Time consuming and expensive to match participants.

  -A large pool of potential participants is needed which can be hard to get.

  -Difficult to know which variables are appropriate for the participants to be matched.
• A pilot study is a small-scale version of an investigation which is done before the real investigation is undertaken. They are carried out to allow potential problems of the study to be identified and the procedure to be modified to deal with these. This also allows money and time to be saved in the long run.
• Single-blind procedure

  A research method in which the researchers do not tell the participants if they are being given a test treatment or a control treatment. This is done in order to ensure that participants do not bias the results by acting in ways they "think" they should act-avoids demand characteristics.
• Double-blind procedure

  A research procedure in which neither the participants nor the experimenter knows who is receiving a particular treatment. This procedure is utilised to prevent bias in research results.

  Double blind studies are particularly useful for preventing bias due to demand characteristics or the placebo effect. Gives a way to reduce the investigator effects as the investigator is unable to unconsciously give participants clues as to which condition they are in.
• Control group/condition - sets a baseline whereby results from the experimental condition can be compared to results from this one. If there is a significantly greater change in the experimental group compared to the control than the researcher is able to conclude that the cause of effect was the IV.
• observations: Naturalistic- watching and recording behaviour in the setting where it would normally take place.

  + High ecological validity

  + High external validity as done in a natural environment

  —Low ecological validity if participants become aware that the are being watched.

  —Replication can be difficult.

  —Uncontrolled confounding and extraneous variables are presented.
• observations: Controlled- Watching and recording behaviour in a structured environment e.g. lab setting.

  + Researcher is able to focus on a particular aspect of behaviour.

  + There is more control over extraneous and confounding variables

  + Easy replication.

  + More likely to be observing unnatural behaviour as takes place in an unnatural environment.

  —Low mundane realism so low ecological validity.

  —Demand characteristics presented.
• observation: Overt- participants are

  watched and their behaviour is recorded with them knowing they are being watched.

  + Ethically acceptable as informed consent is given.

  + More likely to be recording

  unnatural behaviour as participants know they are being watched.

  —Demand characteristics likely which reduces validity of findings.
• observations: Covert- the participants are unaware that their behaviour is being watched and recorded.

  + Natural behaviour recorded hence high internal validity of results.

  + removes problem of participant reactivity whereby participants try to make sense of the situation they are in, which makes them more likely to guess the aim of the study.

  —Ethical issues presented as no informed consent given.

  —Also could be invading the privacy of the participants.
• observations: Participant- The researcher who is observing is part of the group that is being observed.

  + Can be more insightful which increases the validity of the findings.

  —There's always the possibility that behaviour may change if the participants were to find out they are being watched.

  —Researcher may lose objectivity as may start to identify too strongly with the participants.
• observations: Non-participant- The researcher observes from a distance so is not part of the group being observed.

  + Researcher can be more objective as less likely to identify with participants since watching from outside of the group.

  —Open to observer bias for example of stereotypes the observer is aware of.

  —Researchers may lose some valuable insight.
• observational designs: One problem with carrying out observations is that observer bias is easily presented. This is when an observer's reports are biased by what they expect to see. A solution to this problem is checking the inter observer reliability of the observation. This is done by many researchers conducting the observational study, their reports are then compared and a score calculated using the formula :-

  Total number of agreements / total number of observations x 100 .

  The score that shows high inter observer reliability is any score above 80%.
• observational designs:

  Unstructured- consists of continuous recording where the researcher writes everything they see during the observation

  + More richness and depth of detail.

  -- Produces qualitative data which is more difficult to record & analyse.

  -- Greater risk of observer bias e.g. only record ‘catch the eye’ behaviours.
• observational designs:

  Structured- Here the researcher quantifies what they are observing using predetermined list of behaviours and sampling methods.

  + Easier as is more systematic.

  + Quantitative data is collected which is easy to analyse and compare with other data.

  + There is less risk of observer bias.

  -- Not much depth of detail.

  -- Difficult to achieve high inter-observer reliability as filling the predetermined lists is subjective.
• Whilst conducting structured observations, behavioural categories can be used. This is when a target behaviour which is being observed is broken up into more precise components which are observable and measurable e.g. aggressive behaviour can be broken down to - shouting, punching, swearing etc. When forming a behavioural categories list, they should be clearly operationalised. During structured interviews there are different types of sampling methods: event and time sampling