Figure 2.
The legend identifies specific, unspecific, and nonsensical. Panel a compares word counts by sample type. In time-based data, 1 to 24 has about 19 per cent specific, 54 per cent unspecific, and 82 per cent nonsensical. For 25 to 49, values are about 35, 26, and 18 per cent. For 50 to 74, values are about 16, 11, and 0 per cent. For 75 to 99, values are about 10, 5, and 0 per cent. For 100 or more, values are about 20, 4, and 0 per cent. In event based data, 1 to 24 has about 12 per cent specific, 44 per cent unspecific, and 100 per cent nonsensical. For 25 to 49, values are about 39, 36, and 0 per cent. For 50 to 74, values are about 26, 10, and 0 per cent. For 75 to 99, values are about 17, 6, and 0 per cent. For 100 or more, values are about 7, 4, and 0 per cent. Panel b compares duration in minutes. In time based data, specific peaks at 2 to 5 minutes with about 46 per cent, unspecific peaks at 1 to 2 minutes with about 38 per cent, and nonsensical peaks at 1 to 2 minutes with about 44 per cent. In event-based data, specific peaks at 2 to 5 minutes with about 63 per cent, unspecific peaks at 2 to 5 minutes with about 40 per cent, and nonsensical peaks at 1 to 2 minutes with about 43 per cent.Reflection quality by (a) word counts and (b) response durations per sample
Source: Authors’ own work