Table 3 Analysis and interpretation of transition matrices
From: Using hidden Markov modelling to reveal in-session stages in text-based counselling
Hidden state | Transition matrix from more satisfying sessions | Transition matrix from less satisfying sessions |
|---|---|---|
Stage 1: Rapport-building | Has a probability of 5.3% of remaining in Stage 1, indicating a relatively more efficient start. | Has a probability of 21.4% of staying in Stage 1 and 57.0% transiting to Stage 2, indicating a relatively less efficient start. |
Stage 2: Problem-identification | Has a high likelihood of progressing to Stage 4 (45.5%), demonstrating a tendency of quickly engaging in solution development after identifying the problem. | Often remains in the same stage (49.0%) or moves to Stage 3 (43.6%), indicating a thorough, albeit perhaps prolonged problem identification and exploration. |
Stage 3: Problem-exploration | Exhibits a strong orientation towards solution development with a probability of 52.1% moving to Stage 4, reflecting efficient problem exploration. | Displays a quite high probability of reverting to Stage 2 (52.4%) and staying in Stage 3 (39.9%), pointing to possible unclarities of the problem and difficulties in conceiving any actionable solutions. |
Stage 4: Problem-solving | Demonstrates a focus on developing solutions with a high likelihood (47.5%) of staying in Stage 4 and a need to frequently reassess the problem with a probability of 31.7% revisiting Stage 3. | Has a probability of 46.2% of staying in Stage 4 and a notable chance of revisiting Stage 2 (37.6%), suggesting efforts to maintain solution-focused discussion and a need to repeatedly clarify the problem. |
Stage 5: Wrap-up | Always stays, marking the end of sessions | |