Module 1 - chapter 1
The empirical cycle
Observation → theory → prediction → Testing → conclusion → evaluation
Falsifiability, parsimony, evidence → Good theory
Research
Basic -- Translational → Applied
Publishing
- For peer review
- The editor’s decision
Module 2 - chapter 2&3
Chapter 2
3 sources of evidence for personal beliefs: experience, intuition,
authority, in comparison with empirical research.
Experience
- In order to draw conclusions, we need all the data - comparison between experimental
condition and control condition; confounding variables need to be controlled, so that
only the chosen independent variable changes
- While personal experience might seem to disprove research, research is
probabilistic - even if personal experience does not align with what is most likely
according to the data, that is because research explains a certain proportion of
the phenomena
- Experience provides one side of the data, leading to biased conclusions
- Experience is liable because it is impossible to control and isolate variables
- Bushman et al 2001 - catharsis is bullshit in alleviating anger
- Practicing aggression only causes more aggression
Intuition
- Human thinking is biased, without the knowledge of the extent of the bias
- Freud supported catharsis because he inspired himself after industrial machines - the
metaphor of “letting off steam” to feel less pressure
- Availability Heuristic - things that pop up easily in our minds guide our thinking
- Overestimation of the likelihood of an event
- We don’t seek out information that we don’t see → present bias
- We only look at evidence that aligns with our previous beliefs; “cherry picking” and
asking for things that confirm what we know→ confirmation bias
, - Bias blind spot = knowing about a possible bias in our judgement and thinking
ourselves unlikely to be biased ourselves
- Empiricism, on the other hand, strives for objectivity to null biases
Authorities
- If the information they provide isn’t empirical, it should not be trusted
- Always check the source of their information and the reliability of the source
- Sources:
- Empirical journal articles
- Review journal articles (meta-analyses) - give the effect size of a relationship
- Chapters in edited books
- Books
- Categorize everything as either argument or evidence
Chapter 3
Claims make statements about variables and the relationship between variables.
Variable varies on different levels (values). A constant is something that could vary but the
researchers chose to keep at one level.
Measured vs. manipulated variables
- measured means the levels of the variable are observed and measured;
- Some variables can only be measured (age, sex, IQ)
- manipulated variable is controlled by the researcher, who changes the levels of it.
- In some cases, it is unethical to manipulate a specific variable (education, drug
use etc).
- Some variables can be both measured and manipulated, depending on the case.
Conceptual variable (construct) - has a conceptual definition; when testing, researchers
create operational definitions → operational variables (the concept becomes measured
or manipulated).
Claim = an argument
- Frequency claims - 1 in 4 people poops
- A rate or degree of a single variable
- In studies with frequency claims, the variables are measured, not manipulated
- Association claims - people who poop are more relaxed
- Variables correlate/covary - are related
- Correlational studies -
- Positive correlation, negative (inverse), and zero associations/correlations
- These can help us make predictions/estimations, except for zero correlation
, - Causal claim - pooping leads to higher rates of relaxation
- One variable is responsible for changing the other
- A positive/negative correlation, where the language suggests that one variable
affects the other (zero correlation - zero causation
- 3 requirements:
1. Establish correlation (≠ 0)
2. The cause variable comes before the outcome variable
3. Must establish that there isnt another explanation
- Tentative language - still causal
Validity = the appropriateness of a conclusion
- construct validity
- In frequency claims, how well a conceptual variable is operationalized, how well
the researchers measured their variables
- In association claims, each variable is evaluated to assess the validity of the claim
- In causal claims - how well were the variables operationalized
- Measuring reliably means that the measure yields similar scores on repeated tests,
different levels correspond to true differences in the variable
- external validity - how generalizable a claim is based on the sample population used in
the study; would the same results be obtained on another population?
- Statistical validity = the accuracy/reason of a conclusion, to what extent do the data
explain the claim?
- In frequency claims - margin of error of the estimate hels describe how well
the sample estimates the true percentage
- In association claims - how strong is the association?
- Type I error - a false positive - there is a correlation when in reality
there isn't
- Type II error - a miss - when the found association is zero but it actually
exists
- Causal claims - asking whether the difference in the DV was statistically
significant → covariance
Causal Claims
Causation has 3 criteria:
1. Covariance = the two variables go together - that’s up to the result of the experiment
2. Temporal precedence = one variable comes first in time, before the other one; the
effect measured on the DV comes after modifying the IV
3. Internal validity = the study’s ability to eliminate alternative explanations for the
association between variables; only experiments can rule out alternative explanations
➢ Causal claims come as results of experiments (the manipulation of one variable (IV) to
measure its effect on another(DV))