• Tidak ada hasil yang ditemukan

Chapter 5. Study 4] Future Mobile Display Devices: Rollable Display

5.3. Results

statistical analyses. Significance was concluded when p < 0.05.

Table 5.2 Effects of Grip Type, Device Thickness, and Hand Length on Grip Widths and Grip Comfort Hand Measures Statistics

Grip type (Grip)

Device thickness

(Device)

Hand length (Hand)

Grip

× Device

Grip

× Hand

Device

× Hand

Grip

× Device

× Hand

Left

Horizontal grip width

p <0.0001 0.93 0.94 0.55 0.73 0.084 0.60

F-ratio F2, 54 = 123.64 F2, 54 = 0.068 F2, 27 = 0.057 F4, 108 = 0.76 F4, 54 = 0.51 F4, 54 = 2.17 F8, 108 = 0.80

Partial η2 0.821 0.003 0.004 0.027 0.037 0.139 0.056

Vertical grip width

p <0.0001 0.33 0.90 0.27 0.85 0.92 0.39

F-ratio F2, 54 = 14.36 F2, 54 = 1.13 F2, 27 = 0.10 F4, 108 = 1.30 F4, 54 = 0.34 F4, 54 = 0.24 F8, 108 = 1.07

Partial η2 0.347 0.040 0.008 0.046 0.024 0.017 0.073

Grip comfort

p <0.0001 0.15 0.99 0.022 0.38 0.27 0.44

F-ratio F2, 54 = 63.79 F2, 54 = 1.99 F2, 27 = 0.011 F4, 108 = 2.98 F4, 54 = 1.07 F4, 54 = 1.34 F8, 108 = 1.01

Partial η2 0.703 0.069 0.001 0.100 0.074 0.091 0.069

Right

Horizontal grip width

p <0.0001 0.072 0.55 0.55 0.99 0.55 0.79

F-ratio F2, 54 = 90.99 F2, 54 = 2.76 F2, 27 = 0.61 F4, 108 = 0.77 F4, 54 = 0.99 F4, 54 = 0.77 F8, 108 = 0.58

Partial η2 0.771 0.093 0.044 0.028 0.006 0.054 0.041

Vertical grip width

p <0.0001 0.56 0.81 0.32 0.78 0.27 0.52

F-ratio F2, 54 = 20.14 F2, 54 = 0.60 F2, 27 = 0.22 F4, 108 = 1.18 F4, 54 = 0.44 F4, 54 = 1.33 F8, 108 = 0.90

Partial η2 0.427 0.022 0.016 0.042 0.032 0.090 0.063

Grip comfort

p <0.0001 0.0094 0.55 0.014 0.14 0.97 0.52

F-ratio F2, 54 = 72.48 F2, 54 = 5.09 F2, 27 = 0.60 F4, 108 = 3.30 F4, 54 = 1.79 F4, 54 = 0.13 F8, 108 = 0.90

Partial η2 0.729 0.159 0.043 0.109 0.117 0.010 0.063

Note. Values of p less than 0.05 are underlined.

Figure 5.2 Interaction effects of grip type × device thickness on grip comfort for each hand (A and Aʹ: high grip comfort groups for the left and right hands according to Tukey HSD testing;

error bars indicate SEs; SE ranges = 1.7–2.8).

5.3.2. Grip type effects

The effects of grip type on both the horizontal and vertical grip widths for the left hand were significant (p < 0.0001; Table 5.2; Figure 5.3). Post-hoc analysis of the horizontal grip width for the left hand showed that the grip type levels were statistically split into two groups (GripFF-GripFP

and GripMM). The mean (SD) horizontal grip width of GripFF, 15.7 (4.6), was the widest, and that of GripMM, 8.9 (3.0), was the narrowest. Regarding the vertical grip width for the left hand, the grip type levels were statistically split into two groups (GripFF-GripFP and GripMM). The mean (SD) vertical grip width of GripFF, 93.5 (20.2), was the widest, and that of GripMM, 83.9 (16.9), was the narrowest.

The effects of grip type on both the horizontal and vertical grip widths for the right hand were significant (p < 0.0001; Table 5.2; Figure 5.3). Regarding the horizontal grip width of the right hand, the grip type levels were statistically split into two groups (GripFF and GripFP-GripMM). The mean (SD) horizontal grip width of GripFF, 15.8 (4.6), was the widest, and that of GripMM, 8.8 (2.7), was the narrowest. Regarding the vertical grip width for the right hand, the grip type levels

were statistically split into three groups (GripFF, GripMM, and GripFP). The mean (SD) vertical grip widths corresponding to GripFF, GripMM, and GripFP were 91.9 (20.8), 82.2 (17.7), and 77.3 (11.3), respectively.

The effects of grip type on grip comfort were significant for both hands (p < 0.0001; Table 5.2 and Figure 5.4). For the left-hand grip comfort, the grip type levels were statistically split into two groups (GripFF-GripFP and GripMM). GripFF provided the highest mean (SE) grip comfort of 78.6 (1.1), followed by GripFP (75.8 (1.2)) and GripMM (47.7 (1.6)). For the right-hand grip comfort, the grip type levels were statistically split into two groups (GripFF and GripMM-GripFP).

GripFF provided the highest mean (SE) grip comfort of 75.1 (0.8), followed by GripMM (45.3 (1.1)) and GripFP (45.2 (1.1)).

Figure 5.3 Effects of grip type on horizontal and vertical grip widths for each hand (A, A’, a, and a’: high grip comfort groups for the left and right hands according to Tukey HSD testing;

error bars indicate SEs; SE ranges = 0.2–1.3).

Figure 5.4 Effects of grip type on grip comfort for each hand (A and Aʹ: high grip comfort groups for the left and right hands according to Tukey HSD testing; error bars indicate SEs; SE

ranges = 1.1–1.6).

5.3.3. Device thickness effects

The device thickness effects were non-significant (p ≥ 0.072) for all of the dependent variables except for the right-hand grip comfort (p = 0.009; Figure 5.5). The device thickness levels were statistically split into two groups (DeviceThick-DeviceMedium and DeviceThin). DeviceThick provided the highest mean (SE) grip comfort of 57.3 (1.2), followed by DeviceMedium (56.5 (1.1)) and DeviceThin (51.7 (1.2)).

Figure 5.5 Effects of device thickness on grip comfort of each hand (A: high grip comfort group for the right hand according to Tukey HSD testing; error bars indicate SEs; SE ranges =

1.5–1.8).

5.3.4. Hand length effects

The hand length effects were non-significant for all six of the dependent variables considered in this study (p ≥ 0.55; Table 5.2).

5.3.5. Gripped regions and percentile values for grip widths

The bezel regions gripped during screen unrolling with the various grip types, device thicknesses, and hand lengths are summarized in Table 5.3, and the horizontal and vertical grip widths for the mean and 5th, 50th, and 95th percentile values for each grip type are depicted in Figure 5.6. To accommodate 95% of the hand lengths, the width and height of the left (right) bezel should be 20.0 (20.0) mm and 122.6 (123.6) mm for GripFF, 14.0 (13.0) mm and 113.0 (113.6) mm for GripMM, and 20.0 (14.6) mm and 123.0 (95.6) mm for GripFP, respectively.

Table 5.3 Regions gripped during screen unrolling by Grip type, Device thickness, and Hand length

Grip type GripFF GripMM GripFP

Device thickness

Thin Medium Thick Thin Medium Thick Thin Medium Thick

Gripped region (left/right)

L R L R L R L R L R L R L R L R L R

Note. Solid line: short hand, dotted line: medium hand, shaded area: large hand. Each side bezel was 140 mm high and 20 mm wide.

Figure 5.6 Mean and percentile values for the horizontal and vertical grip widths for each grip type (solid lines: horizontal grip widths; dotted lines: vertical grip widths; error bars indicate

SEs; SE ranges = 0.49–3.80).

5.3.6. Bimanual coupling with respect to grip widths and comfort

The bimanual coupling strength was analyzed using bimanual correlations for horizontal grip width, vertical grip width, and grip comfort (Figure 5.7). Overall, when both hands were in identical or similar conditions (identical grip types (GripFF and GripMM) for both hands or identical or similar thicknesses for both device sides), the bimanual correlations for horizontal grip width and grip comfort were high.

In the case of GripFF, the bimanual correlations for the horizontal and vertical grip widths were 0.80–0.86 and 0.82–0.92 (p ≤ 0.0001), respectively, and those for grip comfort were 0.76–0.97 (p

≤ 0.0001), for all device thicknesses. In the case of GripMM, the bimanual correlations for the horizontal and vertical grip widths were 0.60–0.86 and 0.89–0.92 (p ≤ 0.0001), and those for grip comfort were 0.80–0.92 (p ≤ 0.0001), for all device thicknesses. In the case of GripFP, the bimanual correlations for the horizontal and vertical grip widths were relatively low (0.23–0.30 and 0.53–0.57) for all device thicknesses and were all significant (p ≤ 0.045), except for the horizontal grip width for DeviceThick (r = 0.13; p = 0.20). Similarly, in the case of GripFP, the bimanual correlations for grip comfort were relatively low (0.22–0.37) for all device thicknesses, although all were significant (p ≤ 0.005).

Figure 5.7 Bimanual correlations for the horizontal and vertical grip widths and grip comfort according to device thickness and grip type (all of the bimanual correlations were significant, with p ≤ 0.027, except for the horizontal grip width for GripFP × DeviceThick, with p = 0.20).