Oral contraceptives (OCs) are widely used hormonal contraceptives composed of synthetic estrogen (ethinyl-estradiol) and/or synthetic progesterone (progestin). These synthetic hormones bind receptors throughout the body, including in brain regions implicated in higher-order cognition such as the prefrontal cortex. Because of their neuroactive potential, there is sustained interest in whether OCs relate to cognitive functions, including attention and sustained attentional processes.
Prior studies have produced mixed results. An early small-sample study reported increased mind wandering among OC users, suggesting possible attentional detriment. Subsequent larger, well-powered replication attempts that collapsed across OC formulations found little evidence for reliable relations between OC use and increases in inattentiveness. The current study asks whether collapsing across OC brands and formulations may have obscured formulation-specific effects on attention-related traits.
OC brands differ across several pharmacological dimensions: ethinyl-estradiol dose, progestin type and dose, dosing schedule (e.g., monophasic versus triphasic versus progestin-only), and relative estrogenic, progestational, and androgenic activity. Prior cognitive research suggests formulation can moderate behavioural outcomes (for example, differential effects on mental rotation performance), and some formulations have been associated with varying risks for mood outcomes. Therefore, investigating formulation-specific associations with attention-related traits is necessary to determine whether treating OC users as a homogeneous group is appropriate.
The present analysis had two primary objectives: (1) to attempt to replicate previously reported null relations between OC use and trait-level attentional engagement, and (2) to examine whether OC formulation moderates relations between OC use and a set of self-reported attention-related traits.
This study used secondary data from the University of Waterloo Pre-screen and Mass Testing surveys, administered to undergraduate students enrolled in psychology courses. Data were aggregated across five semesters (Fall 2023, Winter 2024, Spring 2024, Fall 2024, Winter 2025) and were collected between September 5, 2023 and April 1, 2025. After applying a priori exclusion criteria and quality checks, the final analytic sample included 2,578 participants: 355 self-reported current OC users and 2,223 non-users.
Participants provided informed consent online and received partial course credit for participation. Importantly, data were collected as part of broader departmental testing batteries, so participants were blind to the specific research question relevant to OC effects.
Exclusions were determined a priori and matched previous procedures used by the authors. Participants were removed if they: had participated in a prior semester (to avoid duplicate records), reported use of a hormonal contraceptive other than OCs, had an IUD, failed to disclose contraceptive status, were pregnant or breastfeeding, or reported current treatment for depression or anxiety (to avoid confounding by mental health treatment and potential psychotropic medication interactions). Participants who failed to respond to key screening items were also excluded.
Standardized data quality checks removed participants with implausible completion patterns: fewer clicks than responses per page, average response speeds faster than 1 second per scale item on more than half of the scales of interest, or completion of fewer than half the items on a given scale. Analyses were conducted with and without age-based exclusions (targeting 18–45 years), and the pattern of results did not meaningfully change; the primary analyses reported exclude participants older than 45 or with missing birth year.
Participants reported the brand name of any current OC in the Pre-screen survey. The authors used brand reports to derive formulation characteristics: ethinyl-estradiol dose, progestin dose, dosing schedule (monophasic, triphasic, or progestin-only), and estimates of estrogenic, progestational, and androgenic activity for each formulation based on referenced compendia. Because only 35 participants provided a clear duration of current OC use, analyses involving length of use were underpowered; exploratory correlations between length of use and attention measures were non-significant (all ps > .062).
Trait-level attention and related constructs were assessed with validated self-report questionnaires. These included separate scales for spontaneous mind wandering (MWS) and deliberate mind wandering (MWD), attention lapses, attention-related errors, boredom proneness, and flow (internal and external). Each mind wandering scale consists of four statements rated on a seven-point scale; higher scores indicate a greater propensity for the construct measured. The authors note that while self-report measures have limitations, they can be better predictors of everyday behaviours than some task-based measures.
The analytic approach mirrored the authors' prior work: comparisons of OC users versus non-users and examination of formulation-level variables derived from reported brands. Key limitations were acknowledged in the source: length of current OC use was infrequently reported, limiting any longitudinal or dose-duration analyses; the dataset relied on self-report measures and cross-sectional secondary data; and participants excluded for current mental health treatment reduce generalizability to those populations. The authors also note that because data were collected as part of larger departmental surveys, the sample composition reflects those recruitment procedures.
Across the derived formulation variables, the authors found no evidence that different OC formulations were related to differences in any of the trait-level attention measures collected (mind wandering, attention lapses, attention-related errors, boredom proneness, or flow). When collapsing across formulations, comparisons of OC users to non-users again revealed no significant differences in trait-level attention in this aggregated undergraduate sample.
These null relations are consistent with the authors' prior larger-sample replications that failed to find reliable associations between OC use and a variety of attention-related constructs, with the notable exception in earlier work where OC use was associated with decreased boredom proneness in one study after controlling for depression, anxiety, and stress; however, that prior effect was not the central finding replicated here after accounting for formulation.
In this secondary, multi-semester undergraduate sample, OC brand-derived formulation variables did not predict differences in trait-level attentional engagement. The authors conclude that, within the scope of these data, OC users can be treated as a relatively homogeneous group when assessing trait-level attention. They emphasize limitations including the cross-sectional and self-report nature of the data and the underpowered assessment of duration of OC use, and they point readers to the anonymized data and analysis scripts available on the Open Science Framework for further inspection. Future work that prospectively tracks formulation, duration of use, and includes behavioral attention measures may be needed to definitively rule out subtle or state-dependent effects.