Objective: The aim of this article was to introduce the Asymmetric Fixed Effects (AFE) model to psychotherapy mechanisms of change researchers as a novel way of studying the effects of improvements and deteriorations in the candidate mechanism(s) separately. Alliance-outcome research was used to illustrate the possibility of estimating separate effects of improvements and deteriorations in the alliance. Method: Two archival data sets were used. One was from community-based primary care services in Sweden using the Clinical Outcomes in Routine Evaluation-Outcome Measure (CORE-OM) and the Working Alliance Inventory-Short form (WAI-S, therapist form) each session with 1,096 patients. The other data set was from a university counseling center in China using the Session Rating Scale (SRS) and the Outcome Rating Scale (ORS) each session with 292 patients. Data were analyzed using the AFE model. Results: The findings indicated that with raw scores, improvements in alliance from one session to the next were followed by lower symptoms/distress scores by the next session, but alliance deteriorations had no effect on next-session symptoms/distress. With alliance deteriorations and improvements defined relative to the samples average linear change over time, improvements, and deteriorations had equal but opposite effects on next session symptom level. Conclusions: Findings confirm the utility of the Asymmetric Fixed Effect model across two cross-national samples in showing that alliance deteriorations and improvements can predict next session symptoms separately at the within-person level. Findings raise new questions regarding the use of detrending in within-patient mechanism of change studies.