Context: Attention bias to alcohol related cues has been implicated in the development and maintenance in problematic drinking behaviour. Research has demonstrated ‘alcohol attention bias’ (AAB) in subclinical drinking populations as well as problem drinkers, however this is not as firmly established in subclinical drinking populations and a variety of direct and indirect measures of attention are used to detect this phenomena.

Objective: This meta-analysis aimed to evaluate and quantify AAB detection in subclinical drinking populations, investigating whether AAB was present in subclinical drinking populations when attention to alcohol versus neutral stimuli was measured. The review focused on two methodologies (visual probe and eye tracking) commonly used to detect AAB.

Data sources: A systematic review of English language, peer reviewed articles were identified using PsycINFO, Web of Science, and SCOPUS databases (up to and including 1st April 2019). An additional study was identified by searching relevant bibliographies.

Study Selection: Only studies that used visual probe and/or eye tracking methodology to assess cognitive bias between alcohol and neutral stimuli in subclinical drinkers were included. A total of five eye-tracking studies and thirteen visual probe studies met full eligibility criteria.

Data Extraction: Independent extraction of articles by 2 researchers using predefined eligibility criteria. Subclinical drinking group mean reaction times and fixation times were pooled for alcohol and neutral stimuli.

Data Synthesis: Random-effects models were used to assess eye tracking (g = -0.41; 95% Cl (-0.96, 0.15); Z= 1.43;p = .15) and visual probe studies (g = -0.09, 95% Cl (-0.19, 0.01); Z = 0.1 \;p - .92) and revealed no significant attention bias effect in the subclinical drinking populations studied; Considerable heterogeneity was present in eye tracking studies.

Conclusions: No attention bias was present in subclinical drinkers when eye tracking and visual probe methodology is used. Methodological standardisation and measurement issues for future research are discussed.