Guokao Xingce - independent foundations common mistakes
Guokao Xingce - independent foundations · Understand the traps and how to avoid them.
Inferring a unique cause from a trend
Several simultaneous changes may influence the same outcome, and chronology does not distinguish their contributions.
What to do instead
Specify the mechanism and compare alternatives before making an attribution claim.
Treating an output as a user outcome
An organization can complete activities while users continue to face the original barrier.
What to do instead
Measure the relevant user experience as well as the implementation count.
Changing the denominator between groups
A comparison may use incompatible populations or periods despite apparently similar percentage labels.
What to do instead
Record each numerator, denominator and period explicitly before comparing rates.
Inventing a budget or legal power
A recommendation can appear precise while relying on authority or resources absent from the supplied scenario.
What to do instead
State missing dependencies and verify authority rather than silently assuming them.
Removing minority experiences from an average
An improved mean can coexist with worse access for a disadvantaged group.
What to do instead
Examine distribution and relevant user subgroups alongside the overall statistic.
Confusing a condition with its converse
If A entails B, observing B does not exclude other ways in which B can arise.
What to do instead
Write the original condition and contrapositive, then test a counterexample to the converse.
Compressing evidence into a stronger claim
Editing sometimes changes some to all or an association into certainty.
What to do instead
Compare each final statement with the exact scope of its supporting evidence.
Repeating a recommendation as proof
A desirable objective does not demonstrate that the proposed mechanism will achieve it.
What to do instead
Explain the causal link, dependencies, indicators and evidence that could disconfirm the proposal.