Tested 9 hypotheses. Treat lift on small n with caution; we report Wilson 95% lower bound + base rate so significance is self-evident.
Trial start precedes trial completion (sanity baseline)
trial_start → trial_complete · window 1095d
n=662hits=448
P=67.7%
base=16.1%
lift=4.21×
conf(95% low)=64.0%
median lead=263d
NIH grant precedes clinical trial start
nih_grant → trial_start · window 540d
n=1473hits=471
P=32.0%
base=26.2%
lift=1.22×
conf(95% low)=29.6%
median lead=21d
Trial completion precedes SEC 8-K (public sponsor disclosure)
trial_complete → sec_8k · window 120d
n=927hits=0
P=0.0%
base=6.1%
lift=0×
conf(95% low)=0.0%
Federal award precedes contractor SEC 8-K (material event)
fed_award → sec_8k · window 90d
n=0hits=0
lift=—
n=0 < min sample 8
Federal (research) award precedes trial start
fed_award → trial_start · window 365d
n=0hits=0
lift=—
n=0 < min sample 8
Complaint velocity precedes CPSC recall
complaint → recall · window 120d
n=0hits=0
lift=—
n=0 < min sample 8
FDA recall precedes SEC 8-K
fda_recall → sec_8k · window 90d
n=148hits=0
P=0.0%
base=4.6%
lift=0×
conf(95% low)=0.0%
CPSC recall precedes SEC 8-K
recall → sec_8k · window 90d
n=226hits=0
P=0.0%
base=4.6%
lift=0×
conf(95% low)=0.0%
NIH grant precedes sponsor SEC 8-K
nih_grant → sec_8k · window 540d
n=1473hits=0
P=0.0%
base=27.4%
lift=0×
conf(95% low)=0.0%