3 ms·
So am I right in concluding that the design was simple enough to be mostly forking-paths-free, and there weren't any oddly specific choices made in the analysis
by currymj 8y ago
So am I right in concluding that the design was simple enough to be mostly forking-paths-free, and there weren't any oddly specific choices made in the analysis (to dig a "significant" result out of the data)?
The fact that it could be replicated from details in the paper without excessive expense or difficulty is also reassuring, even before knowing if such a replication has taken place.
- 9wzYQbTYsAIc 8y agoAbsolutely, this was a straightforward research design with a straightforward statistical analysis. The statistical significance was strong. To give you a better mental model: they essentially came up with two sets of 6 words where one set was paired to representing empathy and the other to personal distress. The authors then grouped doctors to receive either biological or psychosocial stimuli and asked them to rate themselves on each word after being presented the stimuli. Statistical analysis was simply comparing the results. Another noteworthy item is that they also included another set of six words to be rated, which they did not use in the statistical analysis, but which they claimed were used as fillers (words like angry, resentful, joyous, etc. - implying that they took care to not make it readily apparent what specific variable they were measuring). Study integrity aside, they found that schizophrenia vignettes elicited almost the same magnitude of empathy for either bio or psychosocial explanations and depression showed the widest gap.
- 9wzYQbTYsAIc 8y agoIn all honesty, I’d recommend that you spend an evening perusing the actual research article as it is a good example of academic social science research, in my opinion.