Objective: Investigate the psychometric properties of the Severity of Dependence Scale (SDS) for codeine and its association with aberrant codeine related behaviors. Design: A voluntary and uncompensated cross-sectional online survey. Setting: Online population (≥18 years). Respondents: Two hundred and eighty-six respondents (66% women) who had used codeine containing medicines in the last 3 months and were living in the UK. Results: Of the respondents (mean age = 35.4 years, SD = 12.5), more than half were employed. Only 3.5% respondents reported no income. The majority of respondents (45.1%) primarily obtained prescription-only codeine from a consultation with a health professional, whilst 40.9% mainly purchased "over-the-counter" codeine containing medicines in a pharmacy without a medical prescription. Principal component analysis indicated a single factor solution accounting for 75% of the variance. Factor loadings ranged from 0.83 to 0.89. Cronbach's Alpha was high (α = 0.92). Several behaviors relating to codeine use were found to significantly predict probable codeine dependence. These included: daily codeine use in the last 3 months (OR = 66.89, 95% CI = 15.8-283.18); tolerance to codeine (OR = 32.14, 95% CI = 13.82-74.75); problems with role responsibility due to intoxication (OR = 9.89, 95% CI = 4.95-19.78); having sought advice on the internet to manage codeine use (OR = 9.56, 95% CI = 4.5-20.31); history of alcohol or drug treatment (OR = 3.73, 95% CI = 1.88-7.43). Conclusions: The SDS was acceptable and feasible to use to assess probable psychological codeine dependence in an online sample of people using codeine containing medicines. SDS scores were associated with behaviors known to be indicators of codeine dependence. Studies are needed in well-defined populations of people who use codeine to test the different aspects of psychometry of the scale compared against "gold standard" criterion [a diagnosis according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-5)].