Problem 7: Psychological research Flashcards
exploratory research
no hypothesis, look at data first, have to be transparent about it
confirmatory research
first hypothesis and analysis method, then data
what can people do to avoid bias and other issues
preregister studies
what kind of research is better according to popper
confirmatory: because it is falsifiable, exploratory is not
HARKing
hypothesising after results are known
what is torturing data
altering your data to fit your hypothesis
pure harking
create hypothesis that is most consistent with results
straw man harking
you already have a hypothesis, but there is the addition of more hypotheses post hoc
suppresses losers hypothesis in harking
there are multiple hypotheses but you only discuss those confirmed by research
post-hoc plausibility and necessity of anticipation (harking)
have a hypothesis ‘formulated before’ that does not align with research
empirical inspiration
hypothesis anticipated as plausible before and after research is conducted
confirmation bias
looking for confirmation
reasons for harking
-confirmation
-requirements set by science
-financial motives
-hindsight bias
-cherry picking the outcomes you need (selective bias of outcome reports)
-subjective bias of the analysis (choosing the test that gives sign results)
-wanting to publish
hindsight bias
in retrospect, past events seem more predictable than in reality
costs of harking
-bad research
-research can’t be falsified
-no replication possible
-no exploration into alternative hypotheses/other possible explanations
