HCD.8:6 - Bias-Annotation
Lenses: Gov, Arch, Onto/Epist, Prag, Did. The table concerns obtaining or building support for a particular development use.
| Recurring bias | Likely drift | Repair |
|---|---|---|
| train-everyone bias | Every provider is sent through a common programme whether a contribution is missing or not. | Start from an established gap. Use a qualified available contribution that satisfies the dependency. |
| asset-readiness bias | Materials, credentials, allocated hours, or installed software are called operational. | Use evidence of representative operation under the required service conditions. Obtain a further observation if that claim remains open and can be tested. |
| internal-build bias | Local capacity is preferred without comparing delay, evidence, risk, and continuity. | Compare obtain and build routes on the current contribution and window. |
| title-competence bias | Reputation, occupation, or contract is taken as contribution-specific capability. | Establish the provider’s fit and representative operation for the required task. |
| generic-enablement bias | A diagram arrow hides the relation or direct result needed from another owner. | Identify the missing result and which provider work depends on it. Use a qualified source or the relevant practice to establish it. |
| infinite-recursion bias | Every support dependency triggers another full build. | Stop at available qualified contributions, optional branches, or a truthful unsupported result. |
| same-example bias | Repeating the training example is treated as operating discrimination. | Use representative evidence of the required variation or judgement. Use another case if the evidence is missing and a further test is appropriate and available. Otherwise limit the claim to the performance observed. |
| service-effect bias | Successful provider operation is reported as learner capability or programme effectiveness. | Keep provider operation, learner receipt, learner attempt, transfer, and retention separate. |