**3. Results**

The 26 observers participating in Study A generated a total of 115 unique terms to describe the behavioural expression of the sheep viewed in the term generation session, with an average of 17 ± 6 (min: 6, max: 31) terms per observer. The 20 observers participating in study B generated a total of 85 unique terms, with an average of 12 ± 4 (min: 6, max: 21) terms per observer. There was significant consensus between observer scores for each of the six treatment comparisons; Procrustes Statistics are shown in Table 3. Terms with the strongest correlation with each of the GPA dimensions for each treatment comparison are shown in Table 3 (max 10 terms); the first two terms are used as labels for each GPA dimension (Figures 5–7).

We recorded significant effects of vehicle driver (included as a random factor in all our analyses) on GPA dimension scores for all except Study B3 (Table 3).

#### *3.1. Study A: Comparison of 1. Trailer Crate Design, 2. Sheep Breed and 3. Point of Origin*

#### 3.1.1. Study A1 Trailer Crate Design: Standard Crate vs. Convertible Crate

The first three GPA dimensions explained a total of 66.5% of the variation between animals (Table 3). The positions of individual sheep on the first two GPA dimensions are shown in Figure 5a. There was a significant treatment effect for GPA dimension 1 (*p* < 0.001) with sheep transported in a standard crate scored as more *calm*/*relaxed* than sheep transported in a convertible crate, which were scored as more *agitated*/*anxious*. There were no significant treatment effects on GPA dimensions 2 or 3. Of the transport trips assessed, 19 of them stocked adult sheep. One transport journey that stocked lambs; this transport journey is labelled 'C2' in Figure 5a and stood out as an outlier, with sheep in this clip scored as more agitated/anxious than any other clip.

#### 3.1.2. Study A2 Sheep Breed: Merino vs. Fat-Tail Sheep

The first three GPA dimensions explained a total of 65.0% of the variation between animals (Table 3). The positions of individual sheep on the first two GPA dimensions are shown in Figure 5b. There was a significant treatment effect on GPA dimensions 1 (*p* = 0.006) and 2 (*p* = 0.011) with fat-tail sheep scored as more *agitated*/*distressed* (GPA1) and *curious*/*alert* (GPA2) than merino sheep, which were scored as more *calm*/*relaxed* and *tired*/*content*. There were no significant treatment effects on GPA dimension 3.

#### 3.1.3. Study A3 Point of Origin: Feedlot vs. Saleyard

The first three GPA dimensions explained a total of 63.1% of the variation between animals (Table 3). There was a significant treatment effect on GPA dimension 3 (*p* = 0.014), with sheep transported from the feedlot scored as more *comfortable*/*relaxed* than sheep transported from the saleyard, which were more *nervous*/*curious*. However this dimension only explained a small proportion (5.8%) of the variability in data. There were no significant differences in GPA dimension 1 and 2 scores.

#### *3.2. Study B: Comparison of the Deck Level within a Crate*

3.2.1. Study B1 Deck Level: Standard Crate Upper vs. Lower Deck

The first three GPA dimensions explained a total of 67.6% of the variation between animals (Table 3). The positions of individual sheep on the first two GPA dimensions are shown in Figure 6a. There was a significant treatment effect for GPA dimension 2 (*p* < 0.001) and 3 (*p* < 0.001), with sheep transported in the upper deck scored as significantly more *alert*/*curious* and *stressed*/*alert* than sheep

transported in the lower deck, which were scored as more *miserable*/*unsure* and *nervous*/*agitated*. There were no significant treatment effects on GPA dimension 1.

#### 3.2.2. Study B2 Deck Level: Convertible Crate Upper vs. Lower Deck

The first three GPA dimensions explained a total of 72.4% of the variation between animals (Table 3). The positions of individual sheep on the three GPA dimensions are shown in Figure 7a,b. There were significant treatment effects for all three GPA dimensions, with sheep on the upper deck scored as significantly more *calm*/*relaxed*, *alert*/*curious*, and *agitated*, *cramped* than sheep on the lower deck, which were scored as more *agitated*/*stressed*, *dejected*/*weary* and *nervous*/*worried*.

#### 3.2.3. Study B3 Standard versus Convertible Crate: Upper and Lower Deck

The first three GPA dimensions explained a total of 71.1% of the variation between animals (Table 3). The positions of individual sheep on the first two GPA dimensions are shown in in Figure 6b. There was no significant effect of crate on the GPA dimension scores. There was a significant effect of deck for GPA dimension 2 (*p* = 0.008) with sheep on the upper decks during road transport scored as more *curious*/*alert* than sheep on the lower decks, which were scored as more *distressed*/*stressed*. The crate x deck interaction terms were not significant.
