Statistical reasoning
A sample estimate has uncertainty because a different sample might give a different result.
Loading video…
A sample estimate has uncertainty because a different sample might give a different result. A confidence interval describes a procedure's long-run coverage, not the probability that this fixed interval contains a fixed parameter. A p-value describes how surprising data are under a specified null model; it is not the probability that a hypothesis is true. Effect size, design and possible confounding matter alongside any p-value.
Worked example
A difference of 0.2 units with p=0.03 may be too small to matter biologically. Show the difference with an interval and ask whether 0.2 units changes the scientific decision.