john Gill technology header image

An Investigation into Audio Device Icon Preference

Conducted as part of the MediVoice Project (a Co-operative Project, funded by the European Commission under FP6)

by Antoinette Fennell
January 2007

 


Introduction and Methods

The aim of this study was to investigate icon preference for labelling the buttons on an audio device. The study explored if conventional icons on audio devices are recognised and preferred by partially sighted readers, when provided with alternative icon options. In some cases, the alternative options were simply variations of the conventional icon. In other cases, the conventional icon was provided alongside very different alternatives.

A sample of n = 93 partially sighted test subjects were presented with thirteen different sets of icons printed in black and white on paper. All icons represented button labels on an audio device. Each set represented a different button function (see Table 1).

The images were presented as black icons on a white background, with a black surround. A previous study into colour contrast preference identified this option as favoured over three other black and white combination options. With respect to buttons on a keypad, this colour option would be presented as a black icon, printed on a white button, with a black surrounding area. See Fig 1. for an example of this colour option on the buttons of an audio device.

Figure 1. A simple 3-Dimensional sketch of an audio device with the preferred colour contrast option of black icons on white buttons, with a black surrounding area.

A simple sketch of a cassette player with black icons on white buttons and a black surrounding area.

The test subjects were asked to choose their favourite in each set, basing their choice on how clear the icons appear as well as how appropriate the symbols are. It should, therefore, be noted that this study did not explore the effectiveness of the symbols using task analysis, but rather was an investigation into personal preference.

The thirteen icon sets are presented in Fig. 2. In the printed questionnaire, the icons were approximately 15mm x 17mm (measurement represents the white square in which the black icon was framed).

In four of the icons sets, a word is provided as an alternative option. Therefore, it should be noted that for the purposes of this report the term "icon" is used to describe any of the button label options depicted in Fig. 2. In the Results and Discussion section, where it is neccessary to distinguish between the two, the term "word option" is used to describe an option where the word, such as 'On' or 'Off', is spelt out as a potential label.

Table 1: A list of the button functions. For each function, between three and six icon options were provided in the printed questionnaire.

Button Name

Button Function

Play

Play

Stop

Stop

Rewind

Rewind

Fast forward

Fast forward

Skip

Skip to next track

On

On

Off

Off

On/Off all

One button, which carries out the two functions on and off.

Information

Information

Help

Help

Pause

Pause

On/Off icon

 

One button, which carries out the two functions on and off, with only icon options provided.

Volume

One button, which carries out the two functions volume up and volume down (i.e. sequential pressing of the button gradually increases the volume to the maximum level, when the button is pressed again it returns the volume to the lowest level.)

 

Figure 2. The icon options provided for each of the button functions listed in Table 1.

Figure showing the icon options provided for each of the button functions listed in Table 1

 

The test included subjects with a wide variety of eye conditions. These subjects were recruited from the RNIB Scientific Research Unit database. Visual acuity was not tested during the course of this study; information on eye condition was provided by the subjects themselves. The reported eye conditions in this group included Retinitis Pigmentosa (n = 22), Macular Degeneration (n = 13), Nystagmus (n = 11), detached or damaged retina (n = 6), vision impairment resulting from injury or illness (n = 5), Glaucoma (n = 4), hereditary (n = 3), Cataract (n = 3), Optic Atrophy (n = 3), brain tumour (n = 3), myopia (n = 3), blocked optic nerves (n = 1), more than one eye condition (n = 14). Two test subjects did not provide this information. 

To avoid bias the order in which the icon pairs were presented was randomised. The icon sets were presented in four randomly chosen orders; n = 27, 22, 24 and 20 test subjects saw the icon pairs in one of the four orders.

The results were analysed (in SPSS 14.0) using Chi-Square Test and McNemar Test for 2 related samples.

 

Results and Discussion

Chi-Square analysis revealed significant differences within all of the thirteen icon sets (see Table 2).

Table 2. The results of the Chi-Square analysis for the thirteen icon sets.

Button Function

N

Chi-Square

df

p-value

Play

89

125.9

5

< 0.001

Stop

91

28.5

3

< 0.001

Rewind

92

38.5

3

< 0.001

Fast-forward

92

58.2

3

< 0.001

Skip

89

27.9

3

< 0.001

On

92

85.7

4

< 0.001

Off

92

101.2

4

< 0.001

On Off all

91

71.6

4

< 0.001

Info

91

84.8

5

< 0.001

Help

92

20.3

4

< 0.001

Pause

92

10.0

3

0.019

On Off icon

80

26.8

2

< 0.001

Volume

91

65.4

4

< 0.001

 

The remainder of the results will be presented by button function. For the purposes of this report, the icons are numbered as they appear in Fig. 1. The farthest icon on the left is numbered as 'Icon 1', the second icon from the left is 'Icon 2', and so on.

1. Play
Icon 4 (see Fig. 3) was chosen a significantly higher number of times than the five other icon options (McNemar Test: Chi-Square values ranged from 30.7 - 43.9, n = 89, all p < 0.001). The preferred 'Play' icon was the boldest of the options provided. The shape of the icon (an equilateral triangle) means that the thinning out at the points of the triangle was kept to a minimum, this is in contrast to the other five options which have thinner points and relatively more white-space.

Figure 3. The preferred 'Play' icon

figure showing the preferred icon for the play button

2. Stop
Icon 3 (see Fig. 4) was chosen the highest number of times (39%) and Icon 4 the second highest (34%). Icons 3 and 4 were very similar in shape. Icons 2 and 1 were chosen 25% and 2% of the time, respectively. The pairwise comparisons, for the identification of the most popular icon, were not significant.

Figure 4. The preferred 'Stop' icon

figure showing the preferred icon for the stop button

3. Rewind
Icon 2 (see Fig. 5) was chosen a significantly higher number of times than the three other icon options (McNemar Test: Chi-Square values ranged from 11.7 - 25.3, n = 92, all p < 0.01). The conventional 'Rewind' icon was chosen as the preferred icon.

Figure 5. The preferred 'Rewind' icon

figure showing the preferred icon for the rewind button

4. Fast forward
Icon 2 (see Fig. 6) was chosen a significantly higher number of times than the three other icon options (McNemar Test: Chi-Square values ranged from 8.8 - 37.1, n = 92, all p < 0.01). The conventional 'Fast Forward' icon was chosen as the preferred icon.

Figure 6. The preferred 'Fast Forward' icon

figure showing the preferred icon for the fast forward button

5. Skip
Icon 4 (see Fig. 7) was chosen a significantly higher number of times than the three other icon options (McNemar Test: Chi-Square values ranged from 5.6 - 21.8, n = 89, all p < 0.05). The conventional 'Skip' icon was chosen as the preferred icon.

Figure 7. The preferred 'Skip' icon

figure showing the preferred icon for the skip button

6. On
Icon 2 (see Fig. 8) was chosen a significantly higher number of times than the four other icon options (McNemar Test: Chi-Square values ranged from 8.8 - 43.5, n = 92, all p < 0.01).

Figure 8. The preferred 'On' icon

figure showing the preferred icon for the on button

7. Off
Icon 2 (see Fig. 9) was chosen a significantly higher number of times than the four other icon options (McNemar Test: Chi-Square values ranged from 21.1 - 43.3, n = 92, all p < 0.001).

Figure 9. The preferred 'Off' icon

figure showing the preferred icon for the off button

8. On/Off all
Icon 2 (see Fig. 10) was chosen a higher number of times than the four other icon options. Three of these four relationships were significant (McNemar Test: Chi-Square values ranged from 19.4 - 37.5, n = 91, all p < 0.001). The remaining relationship (Icon 2 vs. Icon 3) was only nearly significant (McNemar Test: Chi-Square = 3.5, n = 91, p = 0.06, n.s.)

Despite the preference for this option, it was assumed that such a long label would not always be feasible. Accordingly, the question was asked a second time (see result below in Section 12. On/Off icon) with the word options eliminated.

Figure 10. The preferred 'On/Off all' icon

figure showing the preferred icon for the on/off all button

9. Information
Icon 3 (see Fig. 11) was chosen a significantly higher number of times than the five other icon options (McNemar Test: Chi-Square values ranged from 18.1 - 37.0, n = 91, all p < 0.001). The 'Information' icon options were simply six different font versions of the lowercase letter 'i'. The preferred font for this letter was Bookman Old Style (see Fig. 11). The other fonts used were Arial Black (Icon 1), Tiresias LPfont (Icon 2), Times New Roman (Icon 4), Century Gothic (Icon 5) and Tiresias Keyfont (Icon 6).

Figure 11. The preferred 'Information' icon

figure showing the preferred icon for the information button

10. Help
Icon 4 (see Fig.12) was chosen a higher number of times than the four other icon options. Three of these relationships were significant (McNemar Test: Chi-Square values ranged from 3.8 - 17.3, n = 92, all p < 0.05). One relationship (Icon 4 vs. Icon 1) was nearly significant (McNemar Test: 3/3, n - 92, p = 0.07).

Figure 12. The preferred 'Help' icon

figure showing the preferred icon for the help button

11. Pause
Icon 4 (see Fig. 13) was chosen a significantly higher number of times than the three other icon options (McNemar Test: Chi-Square values ranged from 4.0 - 6.1, n = 92, all p < 0.05).

Figure 13. The preferred 'Pause' icon

figure showing the preferred icon for the pause button

12. On/Off icon
When the word options were removed from the 'On/Off all' options (see Section 8 above), Icon 2 (see Fig. 14) was chosen a significantly higher number of times than the two other icon options (McNemar Test: Chi-Square values 10.7 and 20.4 respectively, n = 80, all p < 0.01).

It is interesting to note that in all icon sets in which a word option was provided (i.e. 'On', 'Off', 'On/Off all' and 'Help'), a word option was always the most popular option. However, with respect to the 'On' and 'Off' options, the second most popular option was the conventional icon (see Fig. 14) and not the other word option that was provided (i.e. Icon 4 for 'On', 'Off' and 'On/Off all' in Fig. 2). Therefore, in cases where an entire word cannot be printed on a button, the icon depicted in Fig. 14 was preferred for the On/Off button.

Figure 14. The preferred 'On/Off icon' option

figure showing the preferred icon for the on/off icon button

13. Volume

Icon 3 (see Fig. 15) was chosen a higher number of times than the four other icon options. Three of these relationships were significant (McNemar Test: Chi-Square values ranged from 9.8 - 37.2, n = 91, all p < 0.01). The remaining relationship was not significant (Icon 3 vs. Icon 4; McNemar Test: Chi-Square = 2.4, n = 91, p = 0.12)

Figure 15. The preferred 'Volume' icon

figure showing the preferred icon for the volume button

 

The author wishes to express her sincerest gratitude to the 93 anonymous volunteers who took part in this study.

 

 



John Gill Technology Limited Footer
John Gill Technology Limited Footer