psychological research Flashcards
identify key points in the history of psychological science
earliest record was metaphysical/supernatural where human experiences and behaviour were attributed to nonphysical forces. ancient greeks only used logic and intuition. scholars over time became dissatisfied with the absence of evidence. mid-late 1800s Germany - started tackling q’s in cognitive P - initial resistance but now empirical evidence in all domains
why was there initial resistance to psychology as an empirical scientific field?
introducing science to assess human experience was seen as an insult to religion
what are the 4 key features of psychological science?
determinism, parsimony, systematic empiricalism, testability
describe determinism
assumptions are in order with no randomness or chaos. the events have systematic meaningful causes
describe parsimony
aims to have the simplest explanation
prefers simplicity over complexity
explanations and causes should not include unnecessary factors of processes
occam’s razor concept
describe systematic empiricism
a structured, organised approach to gathering data/observations in order to answer questions
what is testability?
when ideas should be confirmed or discomfirmed using available research techniques
what is falsifiability?
concept linked to testability. the openess of researchers to let their idea be disconfirmed (shown to be wrong)
what are the benefits of taking a scientific approach to studying psychological q’s?
enables systematic collection of empirical data that would be hard to obtain via intuition/logic/causal observation
provides evidence across multiple studies to support particular answers to a Q
builds valid evidence informed theory
can rule out answers that aren’t supported by evidence
can produce recommendations for policy and practiced based approaches
what are the 7 steps in the research process?
generate a research question, consult theory, state hypotheses, set up study, collect data, analyse data and interpret results, answer the research question
define theory
a general principle or set of principles that explain a phenomenon or event
what is a hypothesis?
a prediction to be tested in a research study
define a theoretical/conceptual framework
a set of principles presented in a systematic or structured way
describe the role theory plays in research processes
can be used to develop a new research question, some principles may only have a few supported studies so can be strengthened by replication
there may be competing explanations that require clarification
can be used to answer a research question
can refer to theory developed to explain other phenomena
results from research studies inform theory by (dis)confirming existing principles, adding new principles with new factors/processes
(theory is continually being updated depending on what research shows)
why is theory important to the research process?
theories offer systematic explanations for the causes and consequences of perceptions, beliefs, emotions, actions
(cant rely on intuition or popular beliefs)
theory provides a systematic recipe for doing empirical research - more structure
provides a rationale for development of research questions
theories inform recommendations for practice/application
explain why every research study needs a rationale
every study needs a rationale (justification) as an argument for why this study will make a novel contribution to existing knowledge. Conducting a study has many potential costs (money, time, resources) why should the study use these up?
In a world of limited time and resources, we need a good reason for conducting and reporting any research study
Reported research also becomes part of the literature of the topic
what is a rationale?
good enough reason for a research study to be carried out
what are the 3 types of rationales for conducting a research study?
new contributions to theory or conceptual understanding
new contributions to methodology
new contributions to practice
describe different types of novel contributions to research
the first test of a new principle, testing competing explanations, replicating a result that hasn’t been widely tested, extension of research
what are the 2 different types of research questions?
differences between groups
relationship between variables
what are the 3 different types of hypotheses that can be generated from a research question?
null hypothesis
experimental/alternative
directional
what is the difference between systematic and unsystematic variation?
systematic = variation due to the effect that is being investigated unsystematic = variation due to other reasons than the effect being investigated
give examples of unsystematic variation
individual differences, context, mood, different researchers, characteristics of study
what is the aim of research?
to investigate how much of the overall variation in a study is systematic or unsystematic
explain the aim and logic of null hypothesis testing
there will never be systematic variation only
always be unsystematic variation (variation due to other effects)
unsystematic V only = null = no effect (all due to other reasons)
systematic and unsystematic = experimental
we focus on null hypothesis because its the simplest to investigate
if we are right about the H1 we need to discredit the opposing H0 of no systematic effect
testing H0 emphasises disconfirmation/falsifiability
if we find no evidence for H0 then we can conclude that there is support for H1
what is probability?
chance of an event occurring ranging from 0-1
explain how probability fits into the logic of null hypothesis testing
can never be 100% certain that the decision to discredit ‘unsystematic variation only’ is the correct one as we can never confirm with absolute certainty, so we use probability to evaluate the extent to which we may be right to discredit the null hypothesis and prove there is an effect to the research question
define population
set of individuals with the particular characteristic(s) under investigation in a research Q
define samples
selection of some individuals from the population to take part in a study
used to estimate what happens in a population
explain why samples are needed
don’t have access, time, resources, not all population will want to participate
solution = sample to represent the larger population
explain the role of populations and samples in hypothesis tests
researchers use samples to test hypotheses about what is happening in the population
what is sampling error?
likely discrepancy (difference) between the results found in a particular sample and the results that would of been obtained from the population captures extent to which a sample is representative of population
what 2 factors influence sampling error?
sample size
sample strategy
high level of representativeness =
low level of representativeness =
low sampling error
high sampling error
why is a random sample better?
lower sampling error
less chance of any systematic bias in strategy
what are the 3 different types of sampling strategies?
random - random selection from whole population
stratified - random selection from sub groups
opportunity - advertised, self selected sample
the pro of random and stratified sampling is that it is likely to produce a representative sample. what are the cons of these 2 sampling methods?
higher costs
need lots of info about participants and subgroups
may need incentives for people to participate
define categorical and continuous variables
categorical:
- each individual assigned to a group/category
- no numerical value/meaning
continuous:
- scores can be ordered on a continuum from low to high
- numerical value/meaning
differentiate between within participants and between participants designs
within = every participant is in every condition of the study so each condition has the same group of participants so each participant produces a DV score in every condition
(repeated measures design)
between = each participant is in 1 condition of the study so each condition has a distinct group of Ps so each P provides a DV score in 1 condition
(independent measures design)
what are the 2 different types of experiments conducted from a between participants design?
true experiment and quasi experiment
differentiate between a true and quasi experiment (the different ways participants are allocated to a condition)
true = random assignment of individuals to groups/conditions (equal chance)
quasi = non random assignment of individuals (makes use of pre-existing groups)
what are the pros and cons of within participants design
pros:
can use smaller sample
no ethical issues around assignment to condition
less chance of individual differences between conditions
cons:
more likely to guess aim
order effects - repetition, practice, interference (can address these effects by counterbalancing)
what are the pros and cons of between participants design
pros: less likely to guess aim no order effects cons: need larger sample potential ethical issue if assigned to 1 condition may be individual differences
what is counterbalancing?
systematically varying the order in which participants complete the conditions in a within-ps design
(typically via random allocation to different orders)
what are the pros of a true experiment?
pros:
random = reduces unintended individual differences between conditions/groups
cons:
random assignment may not be possible as participants may already belong to a group being studied, may be ethical issues, practical constraints
what are the pros and cons of a quasi experiment?
pros:
can be used in contexts where random assignment is not possible
cons:
non random = potential unintended individual differences as people in pre existing groups may share other characteristics that may explain differences found
what types of studies are used to investigate relationships between variables?
categorical study and correlational study
differentiate the two different studies that can be chosen when looking at approaches to time
cross sectional = single point in time
correlational = over a period of time
what are the 5 types of measures used in research studies
self report, implicit, biological, test/task performance, behaviour
what is the ideal for choosing an appropriate measure in a research study?
high alignment of measure with variable (direct links between measure and variable being assessed)
low demand characteristics
low level of bias
high potential for bias when choosing a measure is caused by?
socially desirable responses
wanting to help the researcher
demand characteristics of a measure are high when?
participants can guess the aim/purpose of measure
responses to measure are under conscious control and made deliberately e.g. self report
what are the 3 most common research settings used in studies?
FIELD STUDIES (some control, real world natural environment) LAB STUDIES (full control, artificial context) ONLINE/MAIL STUDIES (no control, self report, questionnaires)
what 2 things are important when choosing an appropriate research setting?
high internal validity
high external validity
what is the difference between internal and external validity?
external = study provides a valid accurate representation of the real world phenomenon internal = study provides a valid accurate answer to the research Q
when is internal validity higher in a research setting?
when the researcher has more control over the setting as it reduces sources of variation
describe the historical context in which research ethics have been developed
1800s-early 1900s - research focused on science and the best most rigorous methods
little consideration given to participant experience and feelings
nazi scientists forced prisoners in jails/conc. camps to partake
suffered extreme physical and psychological harm
this brought participant experience to attention
Nuremberg code developed after WW2
code created to protect all humans who participate in medical and scientific research
inspired modern ethical principles
what are research ethics?
use and application of moral principles and practice in the context of research using (human) participants
consideration of risks vs benefits
what are the possible risks of conducting research?
physical and psychological harm to ppts, wasting ppts time for no substantive purpose, misinterpretation/misuse of research outcomes
what are the possible benefits of conducting research?
academic impact (scientific knowledge to build theory) real world impact (insights to improve practice) personal benefit (insights into self and world)
list the 4 general principles of the BPS code of research ethics
respect for autonomy, privacy, dignity of individuals and communities
maximising benefits and minimising harm
scientific integrity
social responsibility
what are the 8 specific ethical considerations in psychology research?
informed consent freedom from coercion confidentiality anonymity debriefing minimise risk of harm contribute knowledge and insights avoid misuse/misinterpretation of results
what 5 specific considerations come under the autonomy, privacy and dignity principle?
informed consent, freedom from coercion, confidentiality, anonymity, debriefing
what specific consideration is linked with the scientific integrity principle?
contribute knowledge and insights
what specific consideration is linked with the social responsibility principle?
avoid misuse/misinterpretation of results
freedom from coercion =
no persuasion using threats or force
individuals are free to decide to not participate, withdraw, leave any task/Q unanswered
confidentiality =
maintenance of a participants privacy
anonymity =
can’t identify individuals from their responses
what is the difference between active and passive deception?
passive = withholding the full truth or key pieces of relevant info
e.g. not mentioning specific variables being measured or describing the full aim of the study
active = intentionally misinforming participants about the true state of affairs in a study
e.g. using confederates, altering test scores to low even if ppt did well to see reaction
what are the ethical problems with deception?
violates principle of fully informed consent,
participants may experience negative harmful responses,
participants may leave the study with misleading info even after being told it was false,
participants may mistrust researchers
explain why deception is still used in psychology
want to study honest real responses to real world events,
telling participants the precise aims/hypotheses/variables could influence their responses, therefore passive deception helps minimise demand characteristics
may not be possible to recreate a real world event while telling the full truth
why are the benefits of active deception?
helps maximise experimental realism
(extent to which a study engages ppts attention so it feels like a genuine activity/experience to them)
helps maximise researcher control
(holds everything constant except the variables of interest)
which is more acceptable, active or passive deception?
passive
how should deception be used in psychology studies?
only use either type of deception if necessary
weigh up the risks vs benefits
always carefully debrief participants
explain why we need to quantify variation and how descriptive stats helps
aim of research = investigate how much of variation is systematic and unsystematic
to differentiate between these 2 types we need to quantify them (else not valid or accurate)
calculations used to quantify are based on info about the scores in the sample
what are the 2 types of descriptive statistics?
measures of central tendency
measures of dispersion
what is the point of central tendency (in a set of scores)?
the score that is the most representative or typical
why is the point of central tendency useful? give 2 reasons
simple description that offers a best guess about the most likely score in a sample and is an essential component of calculating some measures of dispersion
what are the 3 measures of central tendency?
mean, median, mode
what measures of central tendency are appropriate for continuous measures?
mean and median
what measures of central tendency are appropriate for categorical measures?
mode
with continuous measures when is it more appropriate to use mean or median?
mean is used more frequently because it offers more info than median and reflects the actual value of all the scores
median is better to use when there are outliers in the set
define the concept of dispersion
dispersion = distribution of scores
allows us to quantify the variation found in a set of scores
what are the 3 measures of dispersion?
range, variance, standard deviation
define variance
the spread of scores around the mean as a point of reference
describe how you calculate the variance
1) calculate mean
2) calculate difference of each score from the mean
3) square these deviance values to get + no.s
4) add up all squared numbers (total sum of deviance)
5) divide sum by number of scores - 1
what is wrong with calculating just the variance?
not in standard measurement units so need to square root the value to find out the standard deviation
what is the definition of standard deviation?
the spread of scores around the mean as a point of reference, IN STANDARD MEASUREMENT UNITS
symmetrical, bell shaped curve, mean roughly in middle, scores mostly clustered around the mean.
what type of distribution is this?
normal distribution
skewed to the right, mean slightly to right of curve, bulk of scores at lower end.
what type of distribution is this?
positively skewed distribution
skewed to the left, mean slightly to left of curve, bulk of scores at higher end.
what type of distribution is this?
negatively skewed distribution
explain the 68 - 95 - 99.7 rule
68% = all scores lie within 1SD of the mean 95% = all scores lie within 2SDs of the mean 99.7% = all scores lie within 3SDs of the mean
increasing the number of SDs increases the range of scores included
nearly all scores in a set will lie within 3SDs of the mean
explain the formula used to predict peoples scores in the population
just need to know the mean and SD
formula = mean (+ or -) 1/2/3 SDs
a higher % score of 99.7% has a larger range than 68% score showing distribution
95% confidence =
1.96 SDs
represents 95% of scores as a standardised value
what is the confidence interval formula?
mean +/- 1.96 x SD/square root (n)
why is the confidence interval calculated?
to identify the range of scores that is likely to include the true population mean
explain why increases in sample size reduce the size of confidence intervals
smaller interval = fewer scores in possible range = more precise estimate of population mean
why does a larger sample size give us a more precise estimate?
reduces sampling error and therefore reduces the discrepancy between the sample and the population
define certainty
level of confidence that the true population mean lies within the range estimated from the sample
95% confidence = __% uncertainty
5
5% uncertainty that the mean lies within the interval
higher level of confidence =
smaller interval size =
more certainty
more precise estimate
researchers ideally want what 2 things when looking at confidence intervals?
high certainty (high confidence interval) high precision (smaller interval sizes)
but difficult to have both at same time so solution is a trade off between certainty and precision
trade off = 95% confidence interval
why is the 95% confidence interval a good compromise?
represents an acceptable middle ground of reasonably good certainty and reasonably good precision
explain how researchers come to make decision errors when testing the null hypothesis
samples can offer imperfect estimates of the characteristics due to sampling error and unsystematic variation
if sample statistics don’t match the real population parameters, researchers will make a decision error when making inferences about the population based on the results from the study
so…
the decision to accept or reject the null hypothesis could be wrong
a researcher finds an effect in the sample, when no effect actually exists in the population.
what type of error is this?
type 1 error
a researcher finds no effect in the sample, when an effect does actually exist in the population.
what type of error is this?
type 2 error
false positive =
type 1 error
we think there is an effect when there actually isn’t
false negative =
type 2 error
we think there is no effect when there actually is
if a researcher makes a type 1 error do they accept or reject the null hypothesis?
reject
make the wrong decision
if a researcher makes a type 2 error do they accept or reject the null hypothesis?
accept
make the wrong decision
explain possible reasons for type 1 errors
sampling error
bias in participants responses
effect occurred randomly/by chance
explain possible reasons for type 2 errors
sampling error
unsystematic variation
explain sampling error as a reason for type 1 and type 2 errors
type 1:
effect is not present in overall population but individuals who do show the effect are over-represented in the sample
type 2:
effect is present in overall population but individuals who don’t show the effect are over represented in the sample
incorrectly reject the null hypothesis =
incorrectly accept the null hypothesis =
type 1 error
type 2 error
define p value
probability of finding an effect in a sample if there is no effect in the population
(probability of making a type 1 error)
(probability of incorrectly finding an effect in a study)
what is the p value range?
0-1
0 = certainly won’t occur, 1 = certainly will occur
explain why p values can never be exactly 0 or 1
because we can never be 100% certain about our inferences
what does a small p value mean?
low probability of having made a type 1 error
if the p value is small, do researchers accept or reject the null hypothesis?
reject
because unlikely that a type 1 error (effect in sample but not in population) has been made
