Wide-field searches for transiting extrasolar giant planets face the difficult challenge of separating true transit events from the numerous false positives caused by isolated or blended eclipsing binary systems. We describe here the investigation of GSC 01944-02289, a very promising candidate for a transiting brown dwarf detected by the Trans-Atlantic Exoplanet Survey (TrES) network. The photometry and radial velocity observations suggested that the candidate was an object of substellar mass in orbit around an F star. However, careful analysis of the spectral line shapes revealed a pattern of variations consistent with the presence of another star whose motion produced the asymmetries observed in the spectral lines of the brightest star. Detailed simulations of blend models composed of an eclipsing binary plus a third star diluting the eclipses were compared with the observed light curve and used to derive the properties of the three components. Using the predicted stellar parameters, we were able to identify a second set of spectral lines corresponding to the primary of the eclipsing binary and derive its spectroscopic orbit. Our photometric and spectroscopic observations are fully consistent with a blend model of a hierarchical triple system composed of an eclipsing binary with G0 V and M3 V components in orbit around a slightly evolved F5 dwarf. The rotational broadening of the spectral lines of the F5 primary (vsini~34 km s-1) and its brightness relative to the eclipsing binary (~89% of the total light) made the discovery of the true nature of the system particularly difficult. We believe that this investigation will be helpful to other groups pursuing wide-field transit searches as this type of false detection could be more common than true transiting planets and difficult to identify.