Affective and cognitive outcomes of science teaching in schools may be enhanced by science outreach labs (SOLs). Particularly interesting is whether and how to connect SOL visits to in-school science learning. Recent studies--among others, in this journal--have confirmed SOLs' positive affective effects. However, research remains inconclusive regarding cognitive outcomes, the effects of connecting SOLs to classroom teaching, and effects beyond the short term. This study more closely investigates the short- and medium-term effectiveness of SOLs on affective (interest, self-concept) and cognitive variables (conceptual and procedural knowledge), as well as the effects of integrating lab work with school-based preparation and post-lab activities. A quasi-experimental intervention study via a repeated measures design and two SOL treatment groups (TG1/2: with and without integration), and an in-school control group (CG; also with integration of pre- and postlab activities, in accordance with good practice) was undertaken. It took place in a lab work unit on "pressure and buoyancy" for the lower secondary level, with identical instruction and the same instructor across groups, and with several further control measures. The main findings are as follows: (i) SOLs can have substantial learning and affective outcomes [pre-post learning gains: Cohen's d > 1; increases in interest (d >0.8); and self-concept (d > 0.5) immediately after the visit]. (ii) Learning gains were still present at medium term (d > 0.7) but not affective ones. (iii) Integration with classroom teaching was necessary for learning (TG1 vs TG2: d = 0.72), and not harmful for interest. (iv) Learning and affective outcomes of an integrated SOL are as good as those of a well-prepared classroom setting, but not better. (v) No interactions of outcome variables with gender and other covariates were found; the lab units appeared suitable for both sexes and different kinds of learners.