developmental sciences and research integrity Flashcards
what methods are used, QRPs, research practices
1
Q
what are the types of reproducibility
A
- are we working with when we’re trying to determine whether an effect might be reproducible
- are we working with same or different dataset when we’re trying to determine an effect is there
- are we doing doing the same or different analysis to previous study say
2
Q
what are the modes of reproducibility (ANOVA)
A
- particular analysis on this dataset and then do the same analysis again
- if you can repeat the same result then you would consider your results reproducible
3
Q
what are the modes of reproducibility (mixed effect modelling)
A
- using same data but really applying a different analytic technique to it
- find similar results
- doing same analysis looking did we get the same results
4
Q
what are the modes of reproducibility (collecting new data)
A
- collect new data from a new set of participants and do the same analysis or do a different analysis
- look at a new measure instead of a whole new study
- result is generalisable if the data is consistent
5
Q
why do babies make good science harder
A
- protected population
- higher recruitment and testing costs
- fickle population
- a lot of data to fully anonymise
- focus on single studies
little consideration of power
6
Q
what is the WOW factor in developmental science
A
- new discoveries about infants especially when go younger draw a lot of attention from the general public
- could also be finding false positives
- inflated effect size
7
Q
what are the methodological considerations in infancy research
A
- differences observed in consented and non participating infants
- base rate for consent is 50%
- systematic differences in who decides to participate
- is it the parental characteristics in the child that changes their experience
- hard to get multiple measures
8
Q
what are the methodological considerations (single outcome measures)
A
- rather than doing a study based on one singular measure they will measure multiple different responses
9
Q
what are the methodological considerations (validity)
A
- rich conclusions from infants responses to simplified or artificial stimuli
- interplay between measurements and ecological validity
- use multiple measures or stimuli with infants
- need to have tighter controls in studies
10
Q
what are the QRPs in developmental
A
- whistleblower papers
- protocol flexibility
- fluid experimental process - eyeballing the data before finishing data collection - conflating pilot and experiment data
- no stopping rules
- modifying methods on the fly
- Ad Hoc theorising - HARKING
11
Q
what are piloting practices
A
- piloting in order to determine whether the data confirms your hypothesis - whether the data is interpretable
12
Q
what are the condition assignments, binding and exclusions
A
- blinding presenters, online coders and offline coders - accidental violations of blinding not identified - parental interference policies
- exclusion of data when parent become unblinded
13
Q
what are the exclusions criteria
A
- preset = min number of habitation rituals
- procedural deviations
- decision makers awareness of how the infant responded prior to exclusion
- equipment failures
- fussiness/inattentiveness
14
Q
what are the exemplary research practices
A
- infants can be a tricky population to get access to
- lead to small sample sizes
- issues with potential replicability
- susceptible to QRPs
15
Q
what is ManyBabies consortium
A
- collaborative approach to infant research
- promotes reproducibility, best practice & theory building
- empirical and methodological data
- test generalisability
