Table 3. The interrater reliabilities and percentages of occurrence of each relationship and context theme in each conversation description: Their last conversation, a conversation in which they felt heard, and a conversation in which they felt not heard.

Code	Kappa [95%CI]	Last (N = 206)	Felt heard (N = 103)	Felt not heard (N = 80)
Relationship type
Service provider(s)—one-off	.81 [.72, .90]	6.80%a	12.62%a	12.50%a
Service provider(s)—repeated	.82 [.74, .90]	13.59%a	11.65%a	17.50%a
Colleague(s)	.92 [.88, .96]	32.04%a	24.27%a	31.25%a
Supervisor(s)	.84 [.75, .93]	5.83%a	9.71%a, b	13.75%b
Friend(s)	.87 [.79, .96]	7.77%a	10.68%a	5.00%a
Family	.95 [.91, .99]	20.39%a	14.56%a	5.00%b
Rest	.81 [.71, .91]	6.31%a	8.74%a	12.50%a
General context
Business	.81 [.75, .87]	50.97%a	55.34%a, b	67.50%b
Private	.83 [.77, .88]	41.75%a	37.86%a, b	28.75%b

Note. The uncodeable conversation descriptions were excluded. Percentages do not add up as the categories were non-exclusive. The percentages are fractions of the total number of described conversations in each category, so 206 for Last, 103 for Felt heard and 80 for Felt not heard conversations. Percentages in the same row that do not share the same superscript significantly differ at p < .05, based on a 2-sample z-test for equality of proportions.