The 12 report functions split into 2 shapes, and the split is what tells them apart.
Data in, object out
These compute and return an object you can read, coerce with
as.data.frame() and often plot.
| Function | Answers |
sframe_codebook() | what the instrument declares, as tables |
sframe_descriptives_results() | the distribution of each variable |
sframe_missing_data_results() | what is missing, and in what pattern |
sframe_quality_results() | which responses the declared checks flagged |
sframe_reliability_results() | alpha and omega per scale |
sframe_item_results() | how each item behaves inside its scale |
sframe_efa_readiness_results() | whether the data suit a factor analysis |
sframe_validity_results() | convergent and discriminant validity |
sframe_assumption_results() | whether a planned test's assumptions hold |
sframe_posthoc_results() | which pairs differ after an omnibus test |
Each returns its primary table through as.data.frame(), with the rest
reachable through the accessors in sframe_accessors. A failure is carried in
the object as an error field, so a report keeps rendering around it.
Object or data in, file out
| Function | Writes |
render_report() | a whole document, computing its sections from the data |
render_results() | a document from analysis results you already have |
Reach for render_results() where run_analysis_plan() has already run,
and render_report() to go from responses to a document in one call.