There is a risk that short-term studies either underestimate disturbance effects because of time-lagged responses, including both time-lagged extinctions and colonizations, or overestimate them because of fast recovery. To evaluate the conservation effectiveness of tree group retention (in this case, buffer strips along streams), we studied the bryophyte community once prior to, and twice after logging, comparing one buffer and one clear-cut plot (0.1 ha) in each of 13 riparian sites. We asked whether time-lagged responses or recovery processes had dominated the period between two re-inventories, 2.5 and 10.5 years after logging, focusing both on the whole community and on species of conservation concern. Although there were examples of recovering species in both clear-cuts and buffer strips, the similarity in species composition to predisturbance conditions had decreased in the second re-inventory. Even if the buffer strips displayed more time-lagged colonizations and local extinctions over the later period compared to the clear-cuts, the overall species composition in the buffer strips was still significantly more similar to the prelogging conditions than the clear-cuts. The red-listed species had mostly declined during the first period, and the number of red-list species per plot (mostly species growing on dead wood) was rather stable at <20% of predisturbance levels in clear-cuts and <60% in buffer strips in the last re-inventory. Synthesis and applications. We show that most extinctions of red-list species occurred soon after disturbance and that the conclusions drawn from a study carried out 2.5 years after the disturbance did not change profoundly 8 years later. Although the species composition in the buffer strips continued to change over time, sensitive species survived much better in buffer strips than in clear-cuts, which supports the practice of retaining buffer strips for terrestrial species too. This knowledge should encourage managers to find ways of increasing the efficacy of this practice. One obvious measure could be to retain wider strips or implement other management practices that make the buffer strips less sensitive to wind, which will lead to higher tree retention to support a prelogging species composition.