Table 1 Agnogenic practices and techniques by soft drink manufacturers in the consultation on South Africa’s proposed sugar-sweetened beverages policy

Confounding Referencing

• The misleading use of references which either overstates or gives an entirely false impression of support for a claim or obstructs evidence appraisal.

Cryptic references

• An opaque reference that provides insufficient information to easily locate the original source and which serves to obstruct evidence appraisal.

Faux sources / False authority

• A faux source involves providing an incorrect source for key data. The concept overlaps with an appeal to a false authority, where an alleged authority is used as evidence to support a claim, which, in fact, is not an authority on the facts relevant to the claim.

Out-of-place citations

• References that give a false impression of support for a proposition as a result of being misplaced in the text. These take various forms and can be used to validate illicit generalisations or simply provide a faux source for a key proposition.

Vapid out-of-place citations

• A hybrid confounding reference (combining an out-of-place citation and a faux source) which contains relatively useless contextual information that fails to support, and has no direct relevance, to the claim in the text.

Source laundering

• Provision of a relatively independent source which obscures the use of industry data as the underlying support for the proposition.

.....

.....

Inaccessible source

• The use of a source that is not publicly available.

Misleading Summaries

• Inaccurate reporting of objectives, findings, and conclusions of sources.

Absence of evidence as evidence of absence

• A logical fallacy aimed at representing a relationship that has not been satisfactorily explored as evidence that no relationship exists (usually used in combination with other techniques, such as omission of qualifying information).

False attribution of focus

• Misrepresentation of the focus of studies.

Omission of important qualifying information

• A specific variant of strategic ignorance characterised by precise but inaccurate reporting of study findings in which important qualifying information that significantly changes the implications of the findings is omitted.

Selective quotation

• Reporting extracts either out of context or by omitting qualifying information to give a misleading impression of either the study quoted or the research upon which it is based.

Simple misstatement of key/study findings

• Erroneously and unambiguously claiming that a study has produced a specific finding.

‘The Tweezers Method’

• The practice of picking phrases out of context from peer-reviewed studies with the effect of changing the emphasis and/or intended meaning of the original text.

Acalculiac rounding-up

• Rounding-up estimates without cause or explanation.

Double-counting

• Counting an economic impact (or part of an impact) more than once.

Illicit Generalisation

• A logical fallacy where the underlying evidence is insufficiently developed to support an inductive generalisation.

Evidential Landscaping

• Either promoting alternative evidence (a parallel evidence base) to shift the evidential basis upon which the policy is being discussed and evaluated or purposefully excluding relevant evidence

Data dredging (misuse of raw data)

• Presenting and/or analysing data to depict relationships or trends that either misrepresent actual relationships or obscure other contradictory relationships and/or trends in the data.

Unmodelled data (misuse of raw data)

• Homespun trend analysis summarising patterns across time that ignores key confounding variables or pre-existing/underlying trends. In this latter sense, unmodelled data may involve a faux counterfactual, where the impact of an intervention is not appropriately explored by comparing the world in which the intervention occurred with the world in which it did not.

Observational Selection/Cherry-Picking

• The practice of highlighting individual studies or data to support a pre-determined conclusion, whilst ignoring contradictory (and typically stronger) evidence.

The ‘Hens’ teeth’ technique

• An egregious form of cherry-picking that involves foregrounding obscure, outlying studies.

Passé Source

• Cherry-picking an older source to support an assumption, which although fairly reflecting the state of scientific knowledge when published has since been superseded by developments in the evidence-base.

Strategic ignorance

• The technique of ignoring findings and evidence-backed observations in cited sources that contradict unsupported or weakly supported claims.

Syncopated Estimation

• Missing or failing to fully articulate key steps in economic modelling (including, but not limited to, the failure to: provide a range of estimates to reflect uncertainties in assumptions; fairly review the literature relevant to specifying assumptions; provide a clear and comprehensive assessment of assumptions).

Black-box Computation (information asymmetries)

• Opaque, unverifiable steps in economic modelling.

Inaccessible Data (information asymmetries)

• The reliance on privately held data in economic assessments.