Research Design and Analysis 3003PSY || with 100% Error-
free Answers.
characteristics of good psychological research correct answers *guided by theory
*standardised procedures
*generalisable
*reliable
*uses mutiple methods
research methodology correct answers the collection of approaches, paradigms and procedures
that give us confidence in data we gather
statistical analyses correct answers algorithms with which we process our (usually quantitative
data) to help arrive at a conclusion about research question
reliability correct answers based on random error
validity correct answers based on coverage of correct underlying construct
3 components of observed scores correct answers *systematic effect of underlying construct
*systematic error (low/high test anxiety perform differently but not cos better/worse stats)
*Chance, random error
What types of RQs suit survey design in general? correct answers if it involves people's beliefs,
attitudes, knowledge, reports or behavior, if interested in measuring rather than manipulating
What components establish causation? correct answers *association (2 things co-occur)
*temporal precedence - cause must precede effect
*being able to excluse other causes
Why can't establish causation with survey design? correct answers can't rule out alternatives as
lack random assignment
Normal distribution correct answers bell shape, unimodal, symmetrical, tails extend indefinitely,
area under curve 100%
Inferential stats correct answers using sample data to make inferences about population
parameters
sampling bias correct answers faulty sampling methods which causes subgroups of population to
be under/over represented
sampling error correct answers no 2 samples will be the same, natural variation
, sampling distributions correct answers the distribution of a statistic that we would expect if we
drew an infinite number of samples. Sampling distribution of mean
Data correct answers peoples observable responses
Threats to data correct answers extrnal validity (from sampling)
internal validity (from poor survey design)
reliability and validity issues
survey correct answers set of (standardised) questions tapping one or more constructs that are
administered to a sample of participants (often probabilistic)
survey modes comparison correct answers Mail
o Lower chance of rapport, control, cost
o Higher chance of nonresponse, misunderstanding
Telephone
o More rapport, less nonresponsive than mail
o Less chance of bias than face to face, but training is crucial
Face to face
o Highest rapport, response, data quality etc
o Downside, most effects of demand, bias etc
Online
o Cost effective recruitment
o Samples just as representative as other modes
Problems with survey data correct answers *misunderstanding questions
*Ps not having access to explicit cognitive rep of item, or hypothetical scanrios not rep of real.
*retrospective recall bias
*social desirability
*demand characteristics
Question types correct answers *open - NOT yes/no, prose, coded later
*closed - common, response scaled, easy, but if too easy = response set
Survey Question Problems correct answers *Bias
*Leading/loaded (should murderers roam streets)
*double barrelled (2 in 1) eg media polls that spread misinformation
T-test and ANOVA good for: correct answers experimental design, but not great for data
collected using surveys
Coefficient of determination correct answers r-squared
quantified relationship between X and Y in a different way, tells us how much of variance of
each variable overlaps
free Answers.
characteristics of good psychological research correct answers *guided by theory
*standardised procedures
*generalisable
*reliable
*uses mutiple methods
research methodology correct answers the collection of approaches, paradigms and procedures
that give us confidence in data we gather
statistical analyses correct answers algorithms with which we process our (usually quantitative
data) to help arrive at a conclusion about research question
reliability correct answers based on random error
validity correct answers based on coverage of correct underlying construct
3 components of observed scores correct answers *systematic effect of underlying construct
*systematic error (low/high test anxiety perform differently but not cos better/worse stats)
*Chance, random error
What types of RQs suit survey design in general? correct answers if it involves people's beliefs,
attitudes, knowledge, reports or behavior, if interested in measuring rather than manipulating
What components establish causation? correct answers *association (2 things co-occur)
*temporal precedence - cause must precede effect
*being able to excluse other causes
Why can't establish causation with survey design? correct answers can't rule out alternatives as
lack random assignment
Normal distribution correct answers bell shape, unimodal, symmetrical, tails extend indefinitely,
area under curve 100%
Inferential stats correct answers using sample data to make inferences about population
parameters
sampling bias correct answers faulty sampling methods which causes subgroups of population to
be under/over represented
sampling error correct answers no 2 samples will be the same, natural variation
, sampling distributions correct answers the distribution of a statistic that we would expect if we
drew an infinite number of samples. Sampling distribution of mean
Data correct answers peoples observable responses
Threats to data correct answers extrnal validity (from sampling)
internal validity (from poor survey design)
reliability and validity issues
survey correct answers set of (standardised) questions tapping one or more constructs that are
administered to a sample of participants (often probabilistic)
survey modes comparison correct answers Mail
o Lower chance of rapport, control, cost
o Higher chance of nonresponse, misunderstanding
Telephone
o More rapport, less nonresponsive than mail
o Less chance of bias than face to face, but training is crucial
Face to face
o Highest rapport, response, data quality etc
o Downside, most effects of demand, bias etc
Online
o Cost effective recruitment
o Samples just as representative as other modes
Problems with survey data correct answers *misunderstanding questions
*Ps not having access to explicit cognitive rep of item, or hypothetical scanrios not rep of real.
*retrospective recall bias
*social desirability
*demand characteristics
Question types correct answers *open - NOT yes/no, prose, coded later
*closed - common, response scaled, easy, but if too easy = response set
Survey Question Problems correct answers *Bias
*Leading/loaded (should murderers roam streets)
*double barrelled (2 in 1) eg media polls that spread misinformation
T-test and ANOVA good for: correct answers experimental design, but not great for data
collected using surveys
Coefficient of determination correct answers r-squared
quantified relationship between X and Y in a different way, tells us how much of variance of
each variable overlaps
