Research Methods 2 Flashcards
Features of a science
Theory construction
hypothesis testing
empirical method
paradigms
replicability
objectivity
falsifiability
what is a science
knowledge and evidence
using scientific method
scientific method
make an observation
develop an explanation
test empirically
do findings fit theory
if not then develop new explanation
paradigms
set of assumptions, theories, methods, terminology shared by psychologists e.g. approaches
Paradigm shift
When an established paradigm is challenged to the point that a new one takes its place
Thomas Kuhn - normal science- theory is dominant
-scientific revolution - caused by disconfirming evidence for normal
-paradigm shift - normal is overthrown - theories and methods change
social sciences - paradigms
Kuhn - science has single paradigm
social sciences lack universally accepted paradigms due to too many conflicting approaches so psychology is deemed pre-science
Falsifiability
theory cannot be considered scientific unless possible for it to be proven untrue
Karl Popper
all swans are white - no matter how many instances of white swans we observe, this does not justify conclusion that all swans are white
should seek disproof rather than examples that confirm theory
null hypothesis
e.g. not all swans are white
researchers should aim to reject null e.g. if no black swans are sighted this strengthens theory
pseudoscience
Popper - cannot be falsified e.g. freud - unscientific
Theory construction
set of general laws or principles that explain a specific behaviour
Hypothesis testing
testable expectations
if scientist fails to support hypothesis then the theory requires modification
empirical method
refers to any methods that provide evidence based on direct experience rather than unfounded belief
reports on world how it really is
well controlled and objective
empirical method in psychology
the variables that are measured in psych can be difficult to directly observe e.g. happiness/intelligence levels
Objectivity
removal of any bias - results collected in a way that is independent of researcher
why objectivity is important feature of science
builds confidence that findings represent real effect rather than views of investigator
helps ensure methods used were well controlled and high internal validity
identify scientific fraud
Replicability
repeatability to determine validity and reliability
why replicability is important feature of science
ensures results are reliable and builds confidence that finding is trustworthy and represents real effect
ensures methods are standardised improving validity
Abstract
overview of entire investigation
allows reader to decide whether or not to read rest of report
Introduction
review of previous research so reader knows what research has already been done
should follow logical flow of ideas to persuade about reasons for carrying out the study
also state aims and hypothesis
Method
design, participants, apparatus, procedure, ethics
should be enough detail for someone to replicate the study
Results
descriptive stats and inferential stats
How to write results section
1) always be very clear on precisely what the findings show with fully operationalised names to conditions and measurements
2) give full information including all numbers and all details that led you to choose critical value
Discussion
1)summary of results - statistically significant?
2)relationship to previous research
3)consider methodology and improve suggestions
4)implications for psychological theory
5)suggestions for practical applications
6)suggestions for future research
Selling Peanuts (to) Monkeys In Pretty Fleeces
References
full details of any books, journals or websites mentioned in report
Case studies
investigation of single individual, group, institution or event
Case study - advantages
+provide rich detailed data
complex interaction of many variables - helps identify what may have been overlooked
+allows behaviour that would be unethical to deliberately manipulate to be studied
e.g. clive wearing
Case study - disadvantages
-may lack validity
population validity due to focus on small sample
prone to social desirability bias
-ethical issues
informed consent e.g. not able to fully comprehend or too young e.g. clive wearing and little hans
confidentiality e.g. easily identifiable due to unique characteristics
Content analysis
observational study where people are observed indirectly - focuses on communications people have produced such as speeches, diaries, films, books etc
coding system is used to convert qualitative to quantitative
Sampling method for content analysis
Selected randomly or identify characteristics
Every page or every nth page
Time sampling or event
Coding
Placing into clearly operationalised categories and use tallying (quantitative) or describing examples in each category (qualitative)
How to conduct content analysis
1) gather materials to be analysed using sampling method
2) read and reread to familiarise
3) form key categories and operationalise
4) tally up number of times each one occurs
5) draw conclusions
Content analysis advantages
+easy to replicate
+high ecological validity
Content analysis disadvantages
-suffers observer bias
-likely to be ethnocentric - culture influences interpretation
Thematic analysis
Identifying recurring themes that emerge from the data and organising them
Allows data to be summarised - remains qualitative
How to conduct thematic analysis
1) familiarise
2) break data into meaningful units - small bits of text that convey meaning equivalent to sentences or phrases
3) assign label to each unit
4) combine labels into broader themes
5) check by collecting new set of data and applying themes
6) write up report with themes as headings
Thematic analysis advantages
+maintains much of detail from original material
+high ecological validity
Thematic analysis disadvantages
-Time consuming
-suffers observer bias
Assessing validity - face
Whether it looks like it is measuring what the researcher intends to
Assessing validity - concurrent
Whether findings match those from a more recognised and established test of same topic
Must correlate strongly +0.8
Improving validity
Questions removed, revised or rewritten
Assessing reliability - test-retest
Same test or measurement given to same participants on two separate occasions
0.8 correlation
Assessing reliability - inter-observer reliability
The extent to which there is agreement between two or more observers
0.8 correlation
Improving reliability
Inter observer - behavioural categories weren’t operationalised clearly enough
Test-retest - questions were ambiguous or too complex
Nominal data
Named categories where frequencies are counted in each category
Discrete
Mode
Bar chart
+basic and simple
-less precise
Ordinal data
Data placed in an order with no fixed intervals between units
Subjective
E.g. ranking, rating, test score
Median
Histograms and line graphs
+more precise than nominal
-less precise than interval
Interval data
Data based on scale that has fixed intervals between each unit
E.g. time in seconds, distance in metres
Mean
Histograms and line graphs
+More objective and precise
-not basic and simple
Probability
Numerical measure of the chance that certain events will occur
Number of particular outcomes divided by number of possible outcomes
Significance
A result that is strong enough for us to be confident it represents a real effect
Stringent level
P < 0.01
99%
Medicine
Standard level
P< 0.05
95%
Psychology
Lenient level
P < 0.1
90%
Pilot study
Type 1 error
Optimistic - false positive
Accepting a hypothesis that is false
Likelihood increased when lenient significance level
Type 2 error
Pessimistic - false negative
Rejecting a hypothesis that is true
Higher likelihood when stringent significance level
Choosing a stats test
Difference Relationship
Unrelated Related Related
Nominal: Chi squared Sign test Chi squared
Ordinal : Mann- Whitney Wilcoxon Spearman’s
Interval: Unrelated T Related T Pearson’s
Chi squared - association + unrelated
Sign test
Number of pluses =
Number of minuses =
Lower number = calculated s value
S value must be less than critical to be significant
Chi squared - degrees of freedom
Rows x columns
Calculated value must be more than critical value to be significant
One tailed hypothesis
Directional
Two tailed hypothesis
Non directional
Template stats answer
The result of the statistical test is/isn’t significant
Therefore the null is rejected/accepted
And the hypothesis is accepted/rejected
This is because the calculated value (X) is greater than the critical value (X)
At the p=0.05 level of significance when N/Df = (X) for a one/two tailed test
This means there is/isn’t a significant ————— between X and Y in terms of their Z
Chances of making type one error
The given level of significance (e.g. P=0.05) tells us the chance (e.g. 5%)
Chances of making a type 2 error
If results aren’t significant look across table to whether significant at a more lenient level
Brief
Read to participants before hand
Asking for consent
Debrief
Read after
Thank them
Reveal true aim
Ethical considerations
Any questions?
