1 of 28
Definition
It is used to randomly assign participants to different groups to reduce
bias and is done by the following order:
-Give each participant of one gender a number between x-x (number
of participants)
-Put all the same numbers into a hat
-Take number out one by one without looking
-Assign first half taken out to condition A and remainder to condition B
-Repeat process for other gender
Outline the process of random
Outline the process of randomisation
allocation
Outline the issues of bias in sampling
, Outline and evaluate correlational
analysis
2 of 28
Term
Explain validity including the ways to assess/check it and the types of
validity
If extraneous variables are not controlled in an experiment they interfere with the
cause and effect between IV and DV and potentially ruin the results, an extraneous
variable that is left uncontrolled is then known as a confounding variable
Participant Variables
These are any characteristics of individual participants that could affect the DV such
as age/gender/IQ
Situational Variables
These are features of a research situation that may influence participants behaviour
such as order effects/time of day/noise levels/demand characteristics
Participant Effects
This is when participants actively seek clues about how to behave as they may want
to be helpful such as social desirability
, Validity means whether something is measuring what it is supposed to
measure, valid means true and thus a research method is valid if it is measuring
what it set out to measure
There are two ways to assess the validity of an experiment/research:
Face validity - this is the simplest technique and is where an independent
researcher separate to the experiment looks at the measure being used and
assesses whether or not it will actually measure correctly. The researcher will
see if there are any extraneous variables that could affect the research and
then makes a conclusion
Concurrent validity - this is where the research is compared with a similar
procedure that has been done in the past where validity has already been
established. If the scores correlate as a strong positive then the new test is
likely to be valid
There are two main branches of validity which are internal and external
Internal looks at extraneous variables within the study and how easily a cause
and effect
Reliability relationship
is achieved canresults
when the be drawn whereas
gained external looks
are consistent, at the
reliability extent
just to
means
which we can generalise the results outside the setting.
consistency
Within external validity there is ecological, population and temporal validity.
There are three ways to assess the reliability of an experiment/research:
Test-Retest method - this is where the same test is given to participants on separate
occasions and then compared and if the two sets of scores are consistent it is said
to be reliable. Important to keep large time gap between to reduce order effects
Inter observer method - used to assess reliability in an observation, the extent to
which researchers agree on the behaviours being observed independently. If they
have a consistent number of tallies agreed then it is said to be reliable
Split half method - this is used to see whether the items in a test/questionnaire are
consistent with other items in the test. Items on one test are split into two and then
the scores on each half are compared, if a high positive correlation is seen then the
test is reliable
There are five ways to improve reliability which are:
1) Use standardised procedures
2) Written instructions
3) Pilot studies
, 4) Standardised method for collecting data
5) Training for observers
An interview is a method for asking people questions in a face to face nature which
sometimes may be over the phone/computer and is good for collecting qualitative
information
:) - Detailed quality information can be obtained which is better than quantitative
data which lacks human meaning, and thus we get a better insight into human
behaviour
:) - Participants freely express themselves allowing them to be honest in their
answers
:( - Social desirability can occur where participants lie to present themselves in the
best possible light, this is more likely if the topic is of high sensitivity
:( - Extremely time consuming to conduct hundreds of interviews unlike a
questionnaire that can just be sent to thousands of people
Structured interview - This is where questions are designed in advance with the aim
of structuring participants responses, the interviewer would follow a script.
:) - Less likely to deviate from the topic and thus easier to analyse the results
:( - No follow up questions if a new line of enquiry appears
Unstructured interview - May contain a topic area for discussion but no set
questions. Provides highly detailed data but is hard to analyse
:) - Effective for when the interview can adapt the questions based on the answers
:( - More affected by interviewer bias as the interviewer will be coming up with
questions on the spot which may be less objective/leading
3 of 28
Definition
Independent Measures Design
This is where different participants take part in each condition of the IV
ie. one group drinks redbull one group doesn't
:) - There will be no order effects
Definition
It is used to randomly assign participants to different groups to reduce
bias and is done by the following order:
-Give each participant of one gender a number between x-x (number
of participants)
-Put all the same numbers into a hat
-Take number out one by one without looking
-Assign first half taken out to condition A and remainder to condition B
-Repeat process for other gender
Outline the process of random
Outline the process of randomisation
allocation
Outline the issues of bias in sampling
, Outline and evaluate correlational
analysis
2 of 28
Term
Explain validity including the ways to assess/check it and the types of
validity
If extraneous variables are not controlled in an experiment they interfere with the
cause and effect between IV and DV and potentially ruin the results, an extraneous
variable that is left uncontrolled is then known as a confounding variable
Participant Variables
These are any characteristics of individual participants that could affect the DV such
as age/gender/IQ
Situational Variables
These are features of a research situation that may influence participants behaviour
such as order effects/time of day/noise levels/demand characteristics
Participant Effects
This is when participants actively seek clues about how to behave as they may want
to be helpful such as social desirability
, Validity means whether something is measuring what it is supposed to
measure, valid means true and thus a research method is valid if it is measuring
what it set out to measure
There are two ways to assess the validity of an experiment/research:
Face validity - this is the simplest technique and is where an independent
researcher separate to the experiment looks at the measure being used and
assesses whether or not it will actually measure correctly. The researcher will
see if there are any extraneous variables that could affect the research and
then makes a conclusion
Concurrent validity - this is where the research is compared with a similar
procedure that has been done in the past where validity has already been
established. If the scores correlate as a strong positive then the new test is
likely to be valid
There are two main branches of validity which are internal and external
Internal looks at extraneous variables within the study and how easily a cause
and effect
Reliability relationship
is achieved canresults
when the be drawn whereas
gained external looks
are consistent, at the
reliability extent
just to
means
which we can generalise the results outside the setting.
consistency
Within external validity there is ecological, population and temporal validity.
There are three ways to assess the reliability of an experiment/research:
Test-Retest method - this is where the same test is given to participants on separate
occasions and then compared and if the two sets of scores are consistent it is said
to be reliable. Important to keep large time gap between to reduce order effects
Inter observer method - used to assess reliability in an observation, the extent to
which researchers agree on the behaviours being observed independently. If they
have a consistent number of tallies agreed then it is said to be reliable
Split half method - this is used to see whether the items in a test/questionnaire are
consistent with other items in the test. Items on one test are split into two and then
the scores on each half are compared, if a high positive correlation is seen then the
test is reliable
There are five ways to improve reliability which are:
1) Use standardised procedures
2) Written instructions
3) Pilot studies
, 4) Standardised method for collecting data
5) Training for observers
An interview is a method for asking people questions in a face to face nature which
sometimes may be over the phone/computer and is good for collecting qualitative
information
:) - Detailed quality information can be obtained which is better than quantitative
data which lacks human meaning, and thus we get a better insight into human
behaviour
:) - Participants freely express themselves allowing them to be honest in their
answers
:( - Social desirability can occur where participants lie to present themselves in the
best possible light, this is more likely if the topic is of high sensitivity
:( - Extremely time consuming to conduct hundreds of interviews unlike a
questionnaire that can just be sent to thousands of people
Structured interview - This is where questions are designed in advance with the aim
of structuring participants responses, the interviewer would follow a script.
:) - Less likely to deviate from the topic and thus easier to analyse the results
:( - No follow up questions if a new line of enquiry appears
Unstructured interview - May contain a topic area for discussion but no set
questions. Provides highly detailed data but is hard to analyse
:) - Effective for when the interview can adapt the questions based on the answers
:( - More affected by interviewer bias as the interviewer will be coming up with
questions on the spot which may be less objective/leading
3 of 28
Definition
Independent Measures Design
This is where different participants take part in each condition of the IV
ie. one group drinks redbull one group doesn't
:) - There will be no order effects
