Spencer v. State, 302 Ga. 133

Spencer was stopped for a nonworking headlight; officer observed slurred speech, odor of alcohol, a bar wristband, and a cup appearing to contain alcohol.
Officer administered the horizontal gaze nystagmus (HGN) test; Spencer exhibited 4 of 6 HGN clues.
At trial the officer testified (over objection and with a continuing objection granted) that, based on his training and experience, 4/6 HGN clues generally indicate a BAC equal to or greater than .08.
Spencer was convicted of DUI (less safe) and open-container; she appealed only the DUI conviction. The Court of Appeals affirmed (Spencer I). The Georgia Supreme Court granted certiorari.
The Supreme Court analyzed admissibility under Harper v. State and related authority, concluding the State failed to establish the scientific foundation to correlate a specific numeric BAC (or a numeric threshold like “≥ .08”) with HGN results.
Because that testimony was admitted without an adequate Harper foundation and the error was not harmless, the Supreme Court reversed Spencer’s DUI (less safe) conviction.

Issue	Plaintiff's Argument	Defendant's Argument	Held
Whether officer may testify that HGN 4/6 generally corresponds to a BAC ≥ .08	Spencer: Such numeric or threshold correlation lacks the scientific foundation required by Harper and is inadmissible.	State: Officer’s testimony described a general correlation (not a specific BAC for Spencer) and therefore was admissible to show impairment.	The Court held the testimony was inadmissible: the State failed to establish scientific validity/reliability to correlate HGN to a numeric BAC or threshold.
Whether HGN is admissible to show impairment (non-numeric)	Spencer: Did not dispute HGN as evidence of impairment generally but challenged numeric correlation.	State: HGN is generally admissible to show impairment; prior cases treat 4/6 as evidence of impairment.	The Court confirmed HGN may be admissible to show impairment generally, but not to establish a specific numeric BAC or threshold absent Harper foundation.

Harper v. State, 249 Ga. 519 (trial court must determine whether a procedure has reached verifiable scientific certainty)
Bravo v. State, 304 Ga. App. 243 (HGN insufficiently established to quantify BAC; admission of numeric estimate was abuse of discretion)
Parker v. State, 307 Ga. App. 61 (HGN 4/6 held evidence of impairment)
Kirkland v. State, 253 Ga. App. 414 (relied on actual breath test results; not a basis to permit HGN-based numeric estimates)
Webb v. State, 277 Ga. App. 355 (discussed limits on using HGN to quantify BAC)
State v. Rose, 86 S.W.3d 90 (Mo. App.) (criticized linguistic devices that present HGN-based numeric inferences without scientific foundation)