LABORATORY EXPERIMENTS - the researcher manipulates the IV to see what effect it has on the DV.
- Cause and effect can be concluded with confidence due to the high level of control over extraneous variables
- Lab experiments are easy to replicate in comparison to other methods because they are controlled under stict conditions
- Often have low ecological validity because of the artidicial and unrealistic nature of the tasks. Therefore findings and conclusions cannot be generalised to real life situations
- Lab experiments have ethical issues such as the difficulty to withdraw due the power role of the experimenter.
Field Experiments- the researcher manipulates the IV to see what effect it has on the DV. This experiment is conducted in the participants natural environment.
- Greater ecological valitdity due to the natural setting of the study. hence findings and conclusions can be generalised with confidence to real life situations.
- Reduction of demand characteristics beacause the particpants are unaware they are taking part in the experiment.
- Lacks control over extraneous variables therefore cause and effect from the results cannot be drawn with confidence.
- harder to replicate due to the lower levels of control,
- Ethical issues such as: invasion of privacy,right to withdraw and defbriefing.
Natural/Quasi- IV is changed ny natural occurance and the researcher records the effect on the DV. (participants are not allocated to conditions)
- demand characteristcs and experimenter effects are minimised as participants are not aware they are being studied.
- high level of external validity because they are based on real life events, therefore people behave naturally.
- not a real experiment- little control over extraneous variables and the IV therefore difficult to establish cause and effect.
- ethics: no informed consent,deception and invasion of privacy.
A type of analysis used to investigate the realtionship between 2 or more variables.
Correlation coeffficeint; a number that indicates the strength of the realtionship between two variables. The higher the number the stronger the relationship.
- provides valuable information about the strength of relationship between variables
- good for exploratory investigations, which can lead to experimental investigations.
- often used for measuring the relaibility of tests.
- no cause & effect can be established. A third factor maybe responsible for the relatonship
- direction of causuality is uncertain. You cannot be sure which variable is causing the other variable to increase or decrease.
A research method used to investigate behaviour by observing what happens in certain situations.
NATURALISTIC- observing behaviour in a particpants natural environment usually without intrusion.
- high level of ecological validity
- less risk of demand characteristics
- no control over ectraneous variables
- open to researcher bias
CONTROLLED- observation is carried out in controlled environment.
- high internal validity
- high risk of demand characteristics.
A research method that involves direct verbal questioning of participants by the researcher. This can be structured, semi-structured or unstructured.
- private aspects of behaviour can be investigated
- can provide detailed qualitative data about sensitive issues
- misinterpretation of data
- demand characteristics
- interviewer bias and open to investigator effects.
An indepth and detailed study of an individual or specific group. This may take place over a long period of time and are often applied to unusual examples of behavior or experience.
- case studies provide valuable insights into certain types of behaviour such as child deprevation and maybe the the only appropriate method to investigate these types of behaviours,
- provide an insight into persons natural environment or behaviour and therefore have good ecological validity.
- case studies lack population validity as they provide insights into only one person or a small group of people. These findings and conclusions therefore cannot be generalised to the wider population,
A technique that involves using a structured set of questions,for asking a large sample of people about thier views,attiudes and behavioures etc...
2 types: closed- responses are fixed by 'yes' ot 'no' answers or to choose from a number of fixed alternatives.
- easy to analyse
- no freedom of expression. (produces low qualitative data)
open- respondants can answer using thier own words. They can express thier opinions and views of thier own behaviour.
- a greater depth of qualitative data is produced
- can be difficult to quantify and analyse.
- Can be used to gather data on a wide variety of topics
- once they are constucted they are easy to administer
- large amounts of data can be gatherd quickly and relatively cheaply
- closed questionnaires are easy to analyse. A comparison betweenn groups is easy
- both quantatative and qualitative data can be produced
- you cannot be sure that the respondant will answer questions truthfully (social desirability bais)
- researcher effects- if the researcher administers the questionnaire, respondants may be influenced by such things such as the researchers age,ethinc origin,gender or attractiveness
- respondants may interpret questions differentleyif the wording is slighlty ambigious
- demand characteristics are a problem.. The partcipants may answer in a way that they think the researcher wants.
Hypothesis - is a clear and testable statement that allows the researcher to predict the outcome of the research.
NULL hypothesis : predicts no difference, relationship or association between variables. Any difference will be due to chance.
ALTERNATIVE/EXPERIMENTAL/RESEARCH hypothesis: predicts a significant difference/relationship/association between variables.
EXPERIMENTAL hypothesis: predicts a signifant difference between conditions. Differences in the IV will cause a difference in the DV.
CORRELATIONAL hypothesis: predicts a significant relationship between two variables. this can be a positive or negative relationship or it can just predict a relationship without stating a direction.
ONE TAILED / TWO TAILED HYPOTHESES
ONE TAILED: a particular direction is predicted e.g.girls will perform significantly higher in thier sceince exam in comaprison to boys.
TWO TAILED: a correlation or difference is stated but no direction is stated e.g. there will be a significant difference in the UMS scores achieved in the psychology exam for males and females.
IV- is the variable that the researcher manipulates and is assumed to affect the dependent variable,
DV- is the variable that is measured and is assumed to be affected by changes in the independent variable,
Situational variable- any aspect of the researcher situation that might influence the particpants behaviour e.g. physical environment,levels of tiredness amd difficulty of task.
These variables are controlled by:
- a standardised procedure,
- counterbalancing to reduce order effects
- random allocation of particpants to groups in order to control participant variables.
EV- is any variable other than the IV effecting the DV e.g.room temp, noise and exhaustion.
EXTRANEOUS VARIABLES (EV)
INVESTIGATOR EFFECTS- this is where some aspect of the researcher's behaviour or characteristics causes the participant to change thoer behaviour e.g. gender, age or attractiveness. the researcher may also be biased when when recording thier data.
This can be overcome by using a double blind procedure where both the researcher and the participant are unaware of the aims and hypothesis. Also the researcher could use INTER-OBSERVER RELIABILTY to overcome investigator effects.
DEMAND CHARACTERISTICS- this is where participants pick up on cues/clues in the research situation about the aims and hypothesis. They may then behave in a way they feel as though they are expected to.
This can be overcome by using a single-blind procedure or use whilst using questionnaires use distracting questions in order to disguise the true aim of the research.
EXTRANEOUS VARIABLES (EV)
Evaluation apprehension and social desirability bias- this is where the particpants may feel concerned that they are being judged or evaluated by the researcher.This can lead to particpants hiding certain aspects of thier natural behaviour which they believe will be judged negatively. Hence participants may give socially desirable answers.
- This can be overcome by ensuring particpants are aware thier answers will remain confedential and emphasising the importance of thier truthful answers for the research.
- Ensuring particpants complete tasks under the least amount of pressure and away from the view of other participants can help to reduce thier evaluation apprehension and thier need to protray themselves in a socially desirable way.
experimental design :INDEPENDENT SAMPLES
INDEPENDENT SAMPLES: Two separate groups of participants. one group takes part in one condition, the other group takes part in the other condition.
- no order effects since participants take part in only one condition
- no participants are lost from the study between trials beacuase the only take part in one condition.
- participant variables are not controlled and this can act as a confounding variable.
- need many participants, twice as many in a repeated measures design
- Random allocation of participants to conditions. Participant have an equal chance of bieng in thier condition. Participants have an equal chance of being in either condition so participant variables are spread between groups,making the groups fairly similar.
experimental design: REPEATED MEASURES
REPEATED MEASURES: One group of participants who take part in both condtions.
- participant variables are controlled (same people with the same abilities in each condition)
- need fewer participants than in an independent sample
- order effects may influence performance in the second condition. This could be due to,practice,fatigue or boredom.
- paticipants may drop out or may not be avaliable for the second test.
Counterbalacning. Half the particpants do condition 1 followed by condition 2 , half do condition 2 followed by 1. This will 'cannel out' any order effects that occur.
experimental design: MATCHED PAIRS
MATCHED PAIRS: One group of participants take part in one condition the other groupin the other condition. Participants are matched on variables considerd important for the experiment e.g. intelligence.
- No order effects since particpants take part in one condtion only.
- Atleast some controlb of participant variables
- some subject variables present,as it is difficult to get an exact match.
- time consuming to gather an exact match.
- time consuming to gather a sample of participants and therefore expensive.
Randomly allocate matched pairs to random conditions.
A random sample is when every member of the target population have an equal chance of being selected.
- a random sample is representative of thr target population because every participant has an equal chance of being chosen; therefore it is an unbiased sample.
- however sometimes it can be very difficult or impossible to gather a random sample of the target population because the participants might not be avaliable or willing to take part.
This method is potentially unbiased because all members of the target population have an equal chance of bieng able to take part. It is free from researcher bias as the researcher does not choose the participants themselves and also free from volunteer bias as participants do not volunteer themselves. However the sample could still end up biased as some people may refuse to take part or the sample could be all one particular sex when trying to represent a mixed target population and therefore the results will not be representative of all sexes.
An opportunity sample is when the researcher asks whoever is avaliable and willing to take part in thoer research. Most university studies involve an opportunity sample of undergraduate students.
- an opportunity sample is that it is the eaisest sample to gather because the researcher uses the first participants avaliable and willing to take part; therefore it is the quickest sample and gather the least expensive sample to use.
- however it is likely to be biased and therefore not representative of the target population.
The sample is biased because it is drawn from a small part of the target population. It is likely that this sample is also likley to be subject to researcher bias as the researchers may only appraoch people who think they look friendly and less imtimidating.
SELF-SELECTED VOLUNTEER SAMPLE
A self selected volunteer sample is when the participants put themselves forward to take part in the research.Alternatively, in a field study may unconciously self select to take part.
- using this sample is is possible to gather a range of people with particular requirements for the study.
- however this type of sample is likely to be biased because research shows that particular types of people voulnteer, such as thoes motivated to take part or having time to take part, therefore self-selecting samples are likely to have low population validity.
this sample is subject to volunteer bias beacause the particpants who volunteer are likely to be highly motivated. This makes the generalisations of the findings to the non-volunteering population questionable.
ISSUES WITH RELIABLITY
Relaibilty- refers to how consistent results are over time.
External reliability- refers to whether results are consistent overtime or situations
Internal relaibility- is a measure of the extent to which something is consistent within its self e.g.all the questions on an IQ test should be measuring the same thing.
reliability can be tested and improved in 3 ways.These techinques include:
- TEST-RETEST RELIABILITY
- SPLIT-HALF RELAIBILITY
- INTER-RATER RELIABILITY ASSES: Used in mainly observational studies because two different researchers may observe the same behaviour but rate it differently. Inter-rater relaibility compares two or more researchers to identify thier level of agreement. A high correlation between obsever ratings suggests the behaviour was observed reliably.
IMPROVE: If the observations do not establish a high correlation the researcher would need to produce clearer behavioural/operationalised definitions and would need to use these definitions to train all observers.
TEST-RETEST and SPLIT HALF
TEST-RESTEST ASSESS: Participants are given a test/questionnaire to complete and are then given the same task on a separate occasion e.g. one week later. The participant's responses are then correlated to identify if they have given similar responses on both occasions.
IF A CORRELATION OF 0.8 IS ESTABLISHED BETWEEN TASKS IT IS CONSIDERD A RELAIBLE MEASURE !!!
IMPROVE: If the correlation between two tests is LOWER than 0.8 the researcher would need to review the measures and then carry out another test-retest on the new test.
SPLIT HALF ASSESS: This technique assesses if the items are consistent with other items on the test. particpants complete a task that is compromised of several numbers of items. This task is then split randomly. the scores on the first half of the test are correlated with thoes in the second half, IF THE TEST IS SEEN TO BE RELAIBLE IT HAS TO ACHIEVE A CORRELATION OF 0.8 OR HIGHER
IMPROVE: A correlation lower than 0.8 suggests that the test should be re-examined /re-constucted to resolve the problems. It would then be tested using SPLIT- HALF.
ISSUES WITH VALIDITY
THE TERM VALIDITY RELATES TO THE EXTENT TO WHICH THE TEST IS MEASURING WHAT IT CLAIMS TO BE MEASURING.
Internal Validity: this refers to how far the research is measuring is measuring what it is intending to measure. This reduced by issues such as low control and demand charactersitcs.
Low control: Is when the researcher has low levels of control, then it reduces the internal validity as other factors could be the cause of the research findings gained.
CONDUCT IN A CONTROLLED ENVIRONMENT.
Demand characteristics: when particpants alter thier behaviour to be in line with what they belive the researcher is looking for then the research is not measuring what it intends to.
SINGLE BLIND PROCEDURE.
ISSUES WITH VALIDITY
- External Validity: This relates to how far research can be generalised to other places,people and times.
- External validity is reduced by issues such as an unrepresentative sample or artificial tasks /settings.
- Unrepresentative sample- If the research uses a limited sample e.g. only one gender, it is harder to generalise to the wider population.
- USE A MORE REPRESENTATIVE SAMPLE
Artificial Setting/task- If findings cannot be eaisly generalised beyond the research environment to other settings this reduces the ecological validity of the research findings.
- CONDUCT IN A MORE REAL LIFE SETTING OR USE A MORE REALISTIC TASK.
CONTENT VALIDITY- The extent to wich a task/measure is covering all of the areas of the beahviour it is aiming to study. e.g. if a IQ test measures only english skills and ingnores maths skills has low content validity as it ignoring all components of the IQ test.
MEASURES OF CENTERAL TENDENCY
- using a calculated mean is the most representative measure of centeral tendency beacause it takes into account all the scores in a set of data.
- however using a calculated mean means that it can be distorted by a single extreme score,which would make it unrepresentative of the data set.
- An advantage of using the median is thatit is not affected by extreme scores.
- A disadvantage of using the median is that it is not represenatative of all the data set.
- An advantage of using the mode is that it is easy to calculate.
- A disadvantage is that it is not representative of all the data set
MEASURES OF DISPERSION
- An advantage of using the range is that it is easy to calculate.
- A disadvantage of using the range is that it is distorted by a single extreme score within a data set.
STANDARD DEVIATION- A measure of reliability or dispersion of scores around the mean. The higher the the SD the greater the spread of scores around the mean.
- An advantage of using SD is that it is a more precise measure of dispersion as all the values in the data set are being taken into account.
- A disavantage of using the SD is that it is more complicated to calculate and so often the rage is used to show the dispersion instead.
Content analysis is a method of indirectly investigating the content of something. For example gender sterotypescould be analysed in newspapers,television,radio,books,song and or paintaings. BOTH QUALATATIVE AND QUANTATIVE DATA ANALYSIS CAN BE PRODUCED.
An advantage of content analysis is that it has good ecological validity,because it has good ecological validity, because it is based on real life communication of information. Therefore the content being investigated can be generalised to real life situations.
A disadvantage is that researcher bias and expectations might result in low internal validity, because the content that confirms the hypothesis is more likely to be indentified and recorded than content that contradicts the research hypothesis.
DESIGNING CONTENT ANALYSIS:
- design hypothesis
- collect sample materials
- operationalise variables
- create a coding system
- analyse content
- add up instances of behaviour from all content
DEVELOPMENT OF A CODING SYSTEM
A coding system is a procedure for identyfying specific behaviours which are given a code for ease of recording. A behaviour checklist is an example of a coding system. For example a coding system for measuring aggression in an observation study could be "pushing,kicking or swearing".
ADVANTAGE - using a coding system allows replication of a study especially if variables are clearly operationalised.
A DISADVANTAGE OF- using a coding system is that researhcers can identify variables in a biased way to confirm thier own expectations therefore lowering the interna validity of research.
Discourse analysis is concerned with looking not just at how language occurs within its social context, but also at what it is used for. Research into discourse analysis is concerned with talking and conversation. A researcher collates samples of thier conversation then nalyses what has emerged e.g. interviewing people on thier veiws towards racism.
An advantage of using discourse analysis is that it is a beneficial method in revealing the underlying attitudes which people hold and the stratergies used in conversation which could not be gained using quantitative methods
A disadvantage of using discourse analysis is that it is subjective as it is only the researchers interpretation and different researchers may produce different cocnclusions lowering the internal validity of the research findings.
LEVELS OF MEASUREMENT!!
NOMINAL DATA - Category or frequency data e.g. placing participants in terms of thier gender . (Least precise level of measurement)
ORDINAL LEVEL DATA- data placed in ranks allowing the researcher to analyse and compare the data e.g. positions in a race.
INTERVALL AND RATIO DATA- the main difference between interval level and ratio is that ratio data can use zero as a base line for example temperature is interval rather ratio as 0 degrees does not mean no temperature because there can be minus degrees. Ratio scales are the strongest of the levels of measurement here and can use zero as a base line.
ratio- money found in a womens purses