LESSON 1 Flashcards
for something to qualify as a scientific theory, what did Popper say a theory must be?
falsifability
what is bayesianism?
bayesianism - statistical methods that assign probabilities based previous on experience or data
what is an ‘open science’?
open science - open access to studies to spread knowledge, and allow meta-research. Data must be based off replicable and reproducible experiments
what is reproducibility?
reproducibility - same data = same analysis
what is preregistration?
preregistration - submit plans for research (e.g aims, methods etc)
what are the 3 types of quantitative research?
descriptive
relational
experimental
what is descriptive research?
descriptive research - allows us to describe a behaviour/phenomena
what is relational research?
relational research - allows predictions based off relations with another phenomena
what is experimental research?
experimental research - allows us to determine he cause of behaviour/phenomena
why is triangulation important in research? (2)
1) methodological pluralism - multiple methods to discover diff aspects of a theory
2) convergence of findings in varying studies lends credence to a theory
what is a theory in psychology?
Theories in psychology aim to define and explain human behaviour, they attempt to predict human behaviour e.g lecturing leads to poorer cognitive performance
what are hypothesis?
Hypothesis = a direct or non-direct statement about the relationship between variables. They must be: falsifiable, testable, precisely stated, rational, the simplest explanation possible (parsimonious)
what are constructs in psychology?
Constructs = are the building blocks of theories e.g psychosis (a mental state) – it can’t be directly tested meaning that we have to operationalise it in order to test it
define the difference between theories, constructs and variables in psychology:
theory: schizophrenia is genetically determined
construct: the disintegration of personality expressed as disorder or feeling
variable: a score of 30 or lower on a personality scale
what are the 4 levels of measurement?
nominal, ordinal, interval and ratio
what is nominal data? (3)`
Categorical data
Lowest level of data
Data is assigned a label not a numerical relationship
what is ordinal data? (3)
Data ranked in a continuum
Intervals between ranks are not equal e.g race positions
Likert scales = ordinal data
what is interval data? (2)
Data where intervals between successive values are equal
There is no accurate ‘0’ – meaning one interval can’t be ‘double’ another of another interval e.g if someone has a 14 shoe size, this does not mean their feet are double the size of someone with size 7 shoe
E.g shoe size, temperature
what is ratio data? (2)
Highest level of data
Equal interval and true zero point – this means that you can define a variable to be ‘double’ of a another e.g 10m is double of 5m
e.g height and distance
what are the 4 main types of experimental design in psychology?
independent groups
repeated measures
factorial designs
quasi
what is a ‘within-subjects’ experimental design?
all ppt take part in all conditions of the experiments
what is a ‘between-subjects’ experimental design?
ppt are only exposed to one of the conditions of the experiment
what are the three things psychologists consider when choosing an experimental design?
Between subjects – eliminate order effects
Within subjects – eliminate individua effects
Number of ppt – how many ppt are available/do we need?
what is cross-sectional research?
cross-sectional design: type of research design in which you collect data from many different individuals at a single point in time
what are the benefits of a cross-sectional design? (4)
Simple and Inexpensive
These studies are quick, cheap, and easy to conduct as they do not require any follow-up with subjects and can be done through self-report surveys.
Minimal room for error
Because all of the variables are analyzed at once and data does not need to be collected multiple times, there will likely be fewer mistakes as a higher level of control is obtained.
Multiple variables and outcomes can be researched and compared at once
Researchers are able to look at numerous characteristics (ie: age, gender, ethnicity, education level) in one study.
The data can be a starting point for future research
what are the limitations of a cross-sectional design? (3)
Does not help determine cause and effect
Cross-sectional studies can be influenced by antecedent consequent bias which occurs when it cannot be determined whether exposure preceded disease. (Alexander et al.)
Timing of the snapshot is not always representative
Cross-sectional studies do not provide information from before or after the report was recorded and only offer a single snapshot of a point in time.
Cannot be used to analyze behavior over a period to time
Cross-sectional studies are designed to look at a variable at a particular moment, while longitudinal studies are more beneficial for analyzing relationships over extended periods.
what is a key issue with longitudinal studies? how can this be solved?
can’t counterbalance (can’t test one child when they were six years old and then when they are 2 years old), solution is control group
what is are the main issues of a repeated measures design?
demand characteristics
order effects
what are the two mechanisms which must be employed when using a repeated measures design?
randomisation - eliminate researcher bias
counterbalancing - to reduce order effects/spread extraneous variables evenly over both conditions
what is a factorial experimental design? can you give an example?
experimental designs with 2 or more IVs
they can either be fully independent/rep measures or mixed
e.g impact of alcohol consumption and work shift patterns on productivity
what is a quasi experimental design?
True experimental design – the assignment of ppt to experimental conditions is pre-determined (e.g gender)
what is the main issue of quasi experimental designs? how can this problem be solved?
as the assignment of ppt to conditions is based on fixed characteristics – likely to be difference between the groups other than the variable of interest, due to issues we have to be cautious of inferring cause and effect
Matching groups – to solve this, identification of a variable which ppt can be matched on between the conditions, you can also match individuals (e.g twins or same background)
when an experiment doesn’t allow for counterbalancing, how can researchers ensure that recorded changes in the DV are caused by only the variable of interest?
control group
what are the benefits of using a control group in research? (2)
inclusion of a control group allows us to account for any order effects that might be present
allows statistical control of the difference in the treatment group which may be impacted by order effects