Gender bias in Health & Fitness Apps

Team Members

Team members: Silke Mulder, Timo den Hartog, Sanne Kalf, Alla Rybina

Part of the project: 'Apps and Their Practices: A comparative issue analysis across app stores and countries'

Facilitated by Esther Weltevrede and Anne Helmond

Introduction

This subproject focused on the gender orientation of the Health & Fitness smartphone applications.

Research Questions

In order to investigate gender bias in the Health & Fitness Applications, the following question have been developed:

How are gender roles represented in health- and fitness applications across the top charts of the Apple App Store and Google Play Store in the top 10 countries with the highest smartphone usage?

Methodology and datasets

The first step was to identify ten countries with the most smartphone phone usage. These countries were: Mexico, Germany, United Kingdome, China, India, United States of America, Russia, Brazil, Indonesia and Japan.

The next step was to identify top-twenty applications in the Health & Fitness category for each of the two app stores as (based on data from January 14th, 2020): Apple App Store and Google Play Store. This resulted in a list of 40 applications per country or 400 applications in total.

The applications were largely repeated from country to country (e.g. "Google Fit: Health and Activity Tracking"-app or "Period Tracker Period Calendar"-app were in top-twenty across many countries), however, there some country-specific applications too (e.g. Japanese "LAVA" app or "Present - Guided Meditation" app in India).

These data were extracted from APP Anie - https://www.appannie.com/.

Several datasets were developed in Excel based on the country. Each initial Excel-sheet contained top-twenty applications per app-store with the app-names and app-icons.

The next step was to identify application IDs for all the selected applications. These were collected from APP Anie as well, a separate Excel sheet with APP IDs for each country was created.

The description of each app in the app store was used as the main source of content for analysis. It became clear during the analysis, that there were no country-specific differences in the apps, that were popular across different countries, meaning that the description of the app remained the same.

The description of each app was content-analysed based on the following categories:

• Gender (coded as "male", "female" or "x" - for gender-neutral apps);
• Subscription price (where available, the monthly subscription was coded in order to investigate whether gender-targeting is reflected in the average subscription price of the app); While initially, only free apps were chosen for the analysis, it became clear that many apps do offer subscription packages or additional features to be paid for.
• Practices (exact functionalities of each app were noted with a question in mind "what does the app offer", for instance, these include a pedometer, recording steps, dietary recommendations and so on.).

Each app-icon was analysed based on the following categories:

• Gender (coded as "male", "female" or "x" - for gender-neutral apps). This variable was coded based on body parts visible on the icon that could be identified as "male" or "female".
• The main colour of the icon.

The next step was to categorise practices and group them based on a smaller and unified set of more general practices that would be feasible to analyse across the apps:

Tracking activity; Tracking exercise; Tracking nutrition; Tracking sleep; Tracking weight; Tracking menstruation; Tracking pregnancy; Tracking water; Guidance diet; Guidance education (consultancy); Guidance meditation; Guidance work-out; Guidance planning; Guidance parenting; Guidance mental health; Guidance sleep; Guidance spiritual; Community; Medical; Beauty.

The data was cleaned in Open Refine, several pivot tables were created in order to proceed further in the analysis.

A new dataset was created based on the list of practice-categories, where each practice was then assigned an app name, gender and country-code, resulting in 1201 rows of data to analyse.

Findings

Visualization of the final dataset was performed in Google Fusion Tables:

1. As we can here from here, the five top practices that were identified in the apps were: Community, Guidance for work-out, guidance in planning, tracking of exercise and tracking of other various activities.
2. UK was found to be the least gender-targeted country when it comes to Health & Fitness applications. While Japan has no mail-only targeted apps at all.
3. Mexico had the most female-targeted app in the Health & Fitness category across two appstores.

Discussion

Male-oriented apps are scarce and were hard to find. Some country specificities were found, that could be of interest for further research, for instance, massage apps in Asian countries, apps aimed at getting taller in some Asian countries and Mexico, meditation and mindfulness apps as opposed to active training and exercise apps. Also, prevalence of certain dietary apps is peculiar to investigate further, for instance the universal character of Keto-diet apps. Also, the globalization trend is evident as there are very few local apps in top-twenty for each country (Japan is particularly interesting to investigate further in this regard).

Limitations

The analysis of smartphones apps was performed through laptops, which may cause bias in data collection. Perhaps, the smartphone interface is better suited for this purpose.

Also, each laptop had different language setting which might have affected the app description, as some descriptions were found in English and some in country-specific languages.

Top-twenty application were collected based on data available on January 14^th, 2020. However, App Anie shows changes top application that are specific to a certain date, therefore the top-twenty may vary from date to date.

Conclusions

Applications are a fruitful ground for the further research and Health & Fitness category is especially forth looking at in order to investigate gender bias. Also, cultural specificities could be could be identified if looking deeper into the apps of this category. Cultural priorities and “beauty stereotypes” could be also identified through this category of apps.

References

• Berners-Lee T (1996) WWW: past, present, and future. Computer 29(10): 69–77. DOI: 10.1109/2.539724.
• Dieter M, Gerlitz C, Helmond A, et al. (2019) Multi-Situated App Studies: Methods and Propositions. Social Media + Society 5(2): 1–15. DOI: 10.1177/2056305119846486.
• Daubs MS and Manzerolle VR (2015) App-centric mobile media and commoditization: Implications for the future of the open Web. Mobile Media & Communication. DOI: 10.1177/2050157915592657.
• Miller PD and Matviyenko S (eds) (2014) The Imaginary App. Cambridge, MA: MIT Press.
• Morris JW and Elkins E (eds) (2015) There’s a History for That: Apps and Mundane Software as Commodity. The Fibreculture Journal (25): 63–88. DOI: 10.15307/fcj.25.181.2015.
• Nieborg DB (2015) Crushing Candy: The Free-to-Play Game in Its Connective Commodity Form. Social Media + Society 1(2). DOI: 10.1177/2056305115621932.
• van Dijck J (2009) Users like you? Theorizing agency in user-generated content. Media, Culture & Society 31(1): 41–58. DOI: 10.1177/0163443708098245.
• Van Dijck J (2013) The Culture of Connectivity: A Critical History of Social Media. New York: Oxford University Press.