Reconexp: Improving ESM Vassilis-Javed Khan, Panos Markopoulos, Berry Eggen, Boris de Ruyter,...

Reconexp: Improving ESM

Vassilis-Javed Khan, Panos Markopoulos, Berry Eggen, Boris de Ruyter, Wijnand IJsselsteijn

Mobile HCI, Amsterdam, 04 september 2008

INDEX

• What is Experience Sampling Method?• Our Application and Study• How does it improve ESM?

Experience Sampling Method

• A quasi-naturalistic method that involves signaling questions at subjects at random times throughout the day.

ESM is gaining popularity in HCI

• Hudson et al. (2002) have used the ESM to explore attitudes about availability of managers at IBM Research

• Consolvo and Walker (2003) have used the ESM for evaluating an Intel Research system called Personal Server

• Froehlich et al. (2007) used ESM to investigate the relationship between explicit place ratings and implicit aspects of travel such as visit frequency

HOWEVER…

• Reported loss of data: at least 20%

HOWEVER…

• Reported loss of data: at least 20%

• How significant is the data lost?

GOAL OF OUR STUDY

• Find out what information are “busy parents” willing to automatically share among themselves

+ under which context

OVERVIEW

1. Insert info to personalize the ESM

2. Execute ESM for a week

3. Foreach day review answers & fill out the

missing points

OVERVIEW OF THE STUDY

• INITIAL TASK (WEBSITE): – Name Places (during a typical working day)

Name Places

OVERVIEW OF THE STUDY

• INITIAL TASK (WEBSITE): – Name Places & activities (during a typical working day)

Link Activities to places

OVERVIEW OF THE STUDY

• INITIAL TASK (WEBSITE): – Name Places & activities (during a typical working day)– Link information to context (place & activity)

Link info to context

Link info to context – name other info

OVERVIEW OF THE STUDY

• INITIAL TASK (WEBSITE): – Name Places & activities (during a typical working day)– Link information to context (place & activity)

• A TYPICAL WORKING WEEK (WEBSITE & PDA):– Ask several times about context (at that time) & info willing to automatically share,

MOBILE DEVICE

REVIEW & UPDATE ON WEBSITE

REVIEW & UPDATE ON WEBSITE

REVIEW & UPDATE ON WEBSITE

OVERVIEW OF THE STUDY

• INITIAL TASK (WEBSITE): – Places and activities (during a typical working day)– Link information to context (place & activity)

• A TYPICAL WORKING WEEK (WEBSITE & PDA):– Ask several times about context (at that time) & info willing to automatically share,

• SEMI-STRUCUTRED INTERVIEW.

11 PARTICIPANTS

• Mean Age: 38 (max: 44, min: 28, σ = 5.72)

• Mean Number of children: 1.91 (max: 4, min: 1, σ = 0.79)

• Mean Age of children: 5.47 (max: 8.5, min: 0.7, σ = 2.57)

• Mean Years of marriage: 10.86 (max: 20, min: 2, σ = 5.22)

• Mean Hours of work per week: 28.18 (max: 40, min: 20, σ = 6.63)

LOG

1. Log into the system2. Link information statements to previously not answered question3. Name Activity which was not answered4. Name Activity which was not answered using existing value5. Name Location which was not answered6. Name Location which was not answered using existing value7. Name “Other” Activity8. Name “Other” Activity using existing value9. Name “Other” Information10. Name “Other” Information using existing value11. Name “Other” Location12. Name “Other” Location using existing value

RESULTS

• Mean number of actions performed (logins not counted in this number) 55.64

• Mean logins (in 5 days) 2.91

• Mean percentage of activities (2nd question) not answered: 48.81%

• Mean percentage of activities recovered (with the website use): 60.13%

• Overall improvement of the website to the method is 29.35%

CONCLUSIONS

• V.j.khan@tue.nl• HTTP://WWW.AWARENESS.ID.TUE.NL

• Improves on the data loss of ESM

• Participants & Researcher can have an overview of the data during the execution of the study

• Answers are personalized • Answers can be annotated at a later point in time