Methods in Human Computer Interaction/Final Report Team Fitts

Introduction

Studies have shown that daily self-monitoring can promote healthy lifestyles and weight loss over an extended period of time. While the act of recording daily activity is in itself beneficial, using mobile applications to monitor fitness makes it easier for people to change their lifestyle habits and adhere to their fitness goals. However, many of these studies also rely on extensive coaching, incentives, and feedback to promote behavioral changes. With the advent of numerous fitness applications available for mobile devices, people are able to collect quantitative data on their day-to-day activity patterns and behaviors. Applications like Fitbit connect a biometric wearable to a mobile device to keep track of daily step count, distance walked, calories burned, and sleep patterns. As the collection of personal data becomes more popular and present in new technology, we seek to study how knowledge of this data would help create positive lifestyle changes. We would also like to further understand features that can help motivate users and increase the effectiveness of the Fitbit application.

Positive motivation is applied in specific ways in the current Fitbit user experience. However, encouragement is known to help with goal attainment and perception (Burke 2012). Perhaps an expanded suite of motivational tactics will help increase the likelihood that positive motivation be beneficial for Fitbit users as they work to attain their fitness goals.

One other common barrier to entry is that people often find it difficult to make time for exercise. A simple daily reminder to exercise could aid in increasing activity. It would be interesting to see if awareness of quantified personal fitness data along with daily reminders and positive reinforcement to walk would increase exercise levels. Are interactions with the Fitbit wearable and mobile application, as well as daily prompts and positive messages, enough to enact real changes in activity level?

Research Question

Do wearables, when paired with a fitness app for monitoring, promote positive behavior change, and can this behavior be sustained over a period of time?

Fitbit Persona

According to studies regarding the Fitbit user base (Sadarangani, 2015), their customer persona is someone who has a desk job, has an expendable income, and is possibly an early adopter of technology. Despite being comfortable with and owning technology they are value conscious. They are willing to pay for a product that would help them improve health habits and want to work out more. The Fitbit consumer owns a smartphone and is active on social media. They are more likely to believe in a product that has good reviews and is from an established company.

User community/sample for this study

The sample of participants that were recruited for this study were selected based on their ownership and usage of a Fitbit wearable tracker and application. Having a smartphone was also a requirement for this study, as it would allow a medium by which to set reminders for their daily exercise. Participants should have a fairly stable baseline activity level. We also recruited participants who demonstrated a desire to add more activity in their lives, regardless of their availability to actually do so.

Literature review

Methods

We used a number of different methodologies to collect our data: pre-screened surveys, self-reported step counts before and during study, pre-study survey, and post-study survey. Participants were screened via an online qualifying survey using Qualtrics as our primary survey data collection platform. We recruited 12 smartphone users from Des Moines, IA, Los Angeles, CA, San Francisco, CA and Seattle, WA to pre-qualify for our study. The data from this survey was used to determine if the potential study participants already had the Fitbit fitness tracker and mobile app, to understand their current usage of mobile technologies, and to understand their fitness goals and aspirations. 9 out of 12 participants were selected to join the study.

The participants were sent a pre-study survey so that we could measure their current perceptions regarding self-image, motivation to exercise, their goals and aspirations for participating in our study, and what they’ve found to be challenging with regards to their exercise goals. Since we were working with human participants, we provided each person with an informed consent form that outlined the purpose of the study. Participants received detailed instructions on how to report their daily steps and how to set up notifications and alerts on their smartphone calendars.

In order to establish a threshold of behavior and perception change, we designed a short baseline period by which the step measurements from the first two days of the study would act as a baseline activity level. For this baseline period, participants reported their step counts from their Fitbit mobile application, but did not receive any reply from their study contacts.

For the next eight-day period, participants were then asked to send us their daily step count as well as to add daily timed alerts on their smartphone calendar to remind them to take a walk every day for at least 15 minutes. During this eight day study phase, participants were provided positive reinforcement in the form of an encouraging text or email message soon after they report their daily step count. A predetermined list of encouraging statements were used for all participants to keep messages consistent. An example of positive reinforcement messages used for the study were, “Good job!”, “Excellent!” and “You’re making good progress!” along with their daily step count number.

At the conclusion of the study phase, participants were given a final online survey that mirrored the pre-study survey to measure changes in impressions before and after the study as well as new questions about their feelings on positive reinforcement and their feelings of goal attainment.

Discussion

Participants achieved their ideal exercise amount with daily Fitbit interaction, reminders, and positive reinforcement. The average participant increased their step count over baseline by 19%. Weekend step counts varied much more than weekday step counts, due to changes in routines and schedules that greatly affected activity (With weekends omitted, the average participant increased their step count over baseline by 28%). The participants had an overall positive opinion on step tracking and the survey results also reported that participants’ awareness of their daily step count was motivating to them. Participants found positive reinforcement motivating as well, they also reported that they felt less self-conscious about others viewing their step count data after the study. The participants reported that the biggest source of difficulty in adhering to a new exercise regimen was variability in daily schedules. A program with a remote coach sending encouraging statements based on step count such as the one piloted in this study, was recommended by the participants as a possibility for helping people keep to an exercise routine.

Conclusions

While our study was not long enough to measure statistically significant changes in behavior, our participants responded to reminders to walk, interaction with the Fitbit wearable/app, and positive reinforcements to increase their activity. The study interventions supported our participants’ ambitions to achieve their exercise goals. Most participants seem to be motivated to continue their new, daily exercise routine.

Limitations

The 15 minutes goal of extra walking may not have been enough to see an average difference considering the length of time for this study. Participants who experienced variations in daily routines and changes in schedules may have found it difficult to find time to exercise. Other unexpected daily challenges (work, school, family etc.) could also affect step count measurements, especially if their baseline was high. A larger sample size could have produced different results, but because of the study time lines it was not possible to gather a larger sample size to discover other relationships in the data. Another limitation to this study was that we did not having a sample of participants who represented a population that reported very low activity levels so that a relationship could have been gathered based on prompts to exercise.

Participants preferences to workout outdoors or indoors may have also affected the overall results. The environmental conditions, for example weather, was also not taken into account for this study. Participants who prefer outdoor workouts without having an indoor alternative and had bad weather during the study may have had their step counts adversely affected just due to circumstance which cannot be accounted for in the results. Self reported data can also be biased and since a post survey was conducted for this study these reports may be providing data that is bias. While participants report that they intend to continue with their new exercise regimen, the study must be expanded and contain a follow up period to see if people made actual long term lifestyle changes. This study, as was conducted, was not long enough to enact real behavior modifications.

Emergent Methods

Biometrics

1. Burke LE, Styn MA, Sereika SM, et al. Using mHealth Technology to Enhance Self-Monitoring for Weight Loss A Randomized Trial. American journal of preventive medicine. 2012;43(1):20-26. doi:10.1016/j.amepre.2012.03.016.

2. Weinberg, R. S., Bruya, L., Garland, H., & Jackson, A. (1990). Effect of goal difficulty and positive reinforcement on endurance performance. Journal of Sport & Exercise Psychology, 12(2

3. http://www.slideshare.net/pranaysadarangani/fit-bit-gtm-spring-2014