Authors:
(1) Nowshin Nawar Arony;
(2) Ze Shi Li;
(3) Bowen Xu;
(4) Daniela Damian.
Inclusiveness Concerns in Different Types of Apps
Inclusiveness Across Different Sources of User Feedback
Automated Identification of Inclusiveness User Feedback
To answer our second research question, How does inclusiveness related user feedback differ for different types of apps?, we analyzed the distribution of the six major categories in our taxonomy across the fifty apps in our dataset, which consisted of five types of software: business, entertainment, financial, e-commerce, and social media (with ten apps in each type of software). Table 2 illustrates the total number of inclusiveness related user feedback in the five types of apps from across the three sources. We found that 633 out of 1,211 of the inclusiveness related user feedback emerged from social media apps. This is followed by financial apps (212) and entertainment apps (196). Whereas e-commerce (99) and business (71) software contain the least number of inclusiveness posts.
The distribution of the inclusiveness categories is shown in Figure 5.Fairness-related concerns represent the most frequently discussed (355 out of 1,211) category and is more prominent for social media, financial, and e-commerce apps. On the contrary, technology is the more dominant category in business and entertainment software and a close second for social media apps.
Social media consists of a significant number of user feedback from the 6 categories, particularly for fairness, technology, and usability. These three from social media surpass even the most popular categories for the other app types. However, we note that the other three categories, privacy, demography, and other human values, are still more frequently occurring than in the other apps.
We also observed varying levels of popularity for each category depending on the type of app. For example, 45 out of 121 inclusiveness concerns for financial apps are about privacy, which is to be expected as financial apps obtain confidential information from users, and it is their responsibility to adequately safeguard this data. Similarly, we find demography to be a popular category for entertainment type of software. Demography being popular in this context appears reasonable, given that location, language, and socioeconomic status may impact the content provided or suggested in entertainment apps. For business apps, we found technology as the most popular. Since productivity is critical for business apps, technological restrictions that prevent users from completing their tasks and work would definitely be a high area of concern.
This paper is available on arxiv under CC 4.0 license.