MobileHCI '05, September 21 2005 1 Mobile Text Entry
 Relationship between Walking Speed and Text Input Task Difficulty David Newton University of Toronto Sachi Mizobuchi Nokia Japan Co., Ltd. Mark Chignell University of Toronto 

MobileHCI '05, September 21 2005 2 Purpose and Research Questions • The analysis of the experiment results focused on two questions. – What is the minimum required size for a soft keyboard? – How is users’ input performance affected by walking? 

MobileHCI '05, September 21 2005 3 Method • Subjects: – 13 U of T students – 6 female, 7 male – 21-33 years old – 1 left handed, 12 right handed • Apparatus: – Testing software – Sample texts – PDA (HP iPaq h2210) – Walking lane with distance 
 measure 

MobileHCI '05, September 21 2005 4 Design • Fully within subject two-factor design (4 key size conditions x 2 posture conditions) • Dependent variables: input speed, accuracy, walking speed, subjective evaluation of ease/difficulty of text input • Subjective ratings were one of: ‘extremely difficult’, ‘difficult’, ‘somewhat difficult’, ‘neither easy nor difficult’, ‘somewhat easy’, ‘easy’, or ‘extremely easy’ 

MobileHCI '05, September 21 2005 5 Testing Program • Written in Java • Presents English sentence for user to reproduce • Distance between bottom of text input box and top of soft keyboard remained constant across size conditions 

MobileHCI '05, September 21 2005 6 Testing Program • Stop / Next button • Appears as ‘Stop’ during text input • Pressing ‘Stop’ ends input, opens dialog for distance walked • Appears as ‘Next’ when text is not being input • Pressing ‘Next’ produces a new sentence for input 

MobileHCI '05, September 21 2005 7 Testing Program • Sample sentence • Chosen randomly from a sample text list 

MobileHCI '05, September 21 2005 8 Testing Program • Editing window • Where users input text using soft keyboard 

MobileHCI '05, September 21 2005 9 Testing Program • Soft keyboard • Four different sizes • The distance between the bottom of the editing window and the top of the keyboard was constant across sizes 

MobileHCI '05, September 21 2005 10 Testing Program • Input time is measured as first key press after ‘Next’ button until last key press before ‘Stop’ button Next button Stop button Input time Text input 

MobileHCI '05, September 21 2005 11 Keyboard Size 20 mm 25 mm 30 mm 50 mm Keyboard Size (w x h) Individual Key Size (w x h) 1 20 x 10 mm 2.0 x 2.5 mm 2 25 x 12.5 mm 2.5 x 3.1 mm 3 30 x 15 mm 3.0 x 3.8 mm 4 50 x 25 mm 5.0 x 6.3 mm 

MobileHCI '05, September 21 2005 12 Procedure • 6 practice sessions: 3 standing, 3 walking • All subjects experienced 4 standing conditions and 2 walking conditions • Subjects input 5 sentences in each combination of conditions • For walking condition, subjects went at own comfortable speed • Experimenter input the distance walked after each sentence 

MobileHCI '05, September 21 2005 13 Results 

MobileHCI '05, September 21 2005 14 Input Speed by Key Size and Posture • Dependent variable is input speed (CPS) • Key size main effect • Significant difference between 2mm and others, 2.5mm and 5mm • No significant difference between standing and walking • No significant difference Between 2.5mm and 3mm Input speed by posture and software keyboard size 1.00 1.20 1.40 1.60 1.80 2.00 2 2.5 3 5 Size (mm) Input speed (CPS) Standing Walking 

MobileHCI '05, September 21 2005 15 Accuracy by Key Size and Posture • Dependent variable = error rate (number of errors / number of characters) • Error rate was higher in walking condition than in standing condition • 2mm keys had significantly higher error rate Error rate 0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00% 3.50% 4.00% 4.50% 5.00% 2 2.5 3 5 key size (mm) Error rate (%) Standing Walking 

MobileHCI '05, September 21 2005 16 Walking Speed • In general subjects walked slowly (mean= 1.77km/h) • Walking speed became slightly faster with larger target size, but difference was not significant • Walking speed differed substantially among subjects Walkingspeed 1 1.2 1.4 1.6 1.8 2 2.2 2 2.5 3 5 size (mm) walking speed (Km/h) 

MobileHCI '05, September 21 2005 17 Subjective Ease • Subjective ease of text input was significantly higher with larger key sizes • The difference between 3mm and 5mm widths in subjective ease seems to be less than the difference between 2.5mm and 3mm 1 2 3 4 5 6 7 2 2.5 3 5 size (mm) score Walking Standing Extremely difficult Extremely easy 

MobileHCI '05, September 21 2005 18 Discussion & Conclusion 

MobileHCI '05, September 21 2005 19 What is the minimum required size for a soft keyboard? • Input speed – The results show significant effect of size and of posture on input speed – Text input speed in 5mm was significantly quicker, and 2mm was significantly slower than the other sizes • Accuracy – Generally few errors made – 2mm wide keys had significantly higher error rate 

MobileHCI '05, September 21 2005 20 What is the minimum required size for a soft keyboard? • Subjective Ease – Higher with larger sizes – Seems to be diminishing returns after 3mm wide • Results suggest that 2.5mm is the minimum required key width for a soft keyboard 

MobileHCI '05, September 21 2005 21 How is users’ input performance affected by walking? • Text input speed decreased when walking • Walking speed was generally slow, with an average speed between 1.5 and 2km/h • Text input seems to require people to slow down when walking, and some people slow down more than others • Once people have slowed to a comfortable rate, walking speed is unaffected by key size 

MobileHCI '05, September 21 2005 22 Recommendation • To avoid slowing people down to much research into key widths >5mm is required • Otherwise, 2.5mm and 3mm widths seem to be a good compromise between making users comfortable and saving screen real estate • If there is sufficient space, 3mm should be used 

MobileHCI '05, September 21 2005 23 Questions & Comments Sachi Mizobuchi - [email protected] Mark Chignell - [email protected] David Newton - [email protected]