PaperlayerView exact revision on arXiv

Source-linked AI summary

Stuck in the Permissions With You: Developer & End-User Perspectives on App Permissions & Their Privacy Ramifications

Mohammad Tahaei, Ruba Abu-Salma, Awais Rashid

arXiv:2301.06534v2cs.HCcs.CRcs.SE

TL;DR

Empirical evidence on developers’ permission choices and understanding, and how their views compare with end-users’, has been limited. The study combined interviews with 19 developers and a survey of 309 Android and iOS end-users. Both groups connected permissions with app functionality, while developers sometimes included multiple permissions because of scope confusion or third-party libraries.

  • Problem

    The study addresses the lack of empirical research on why developers request permissions, how they understand them, and how their perspectives compare with end-users’.

    Source: §1 · p. 1

  • Method

    The researchers interviewed 19 developers and used those findings to design a survey of 309 Android and iOS end-users.

    Source: §1 · p. 1

  • Results

    Both groups associated permissions with app functionality, while developers sometimes included multiple permissions because of scope confusion or third-party library requirements.

    Source: §1 · p. 2, §6 · p. 15

  • Takeaways & Limitations

    The findings offer insights for improving the permission ecosystem for developers and end-users.

    Source: §6 · p. 15

    The samples may not represent all developers or end-users; developer findings are not generalizable, and the end-user survey targeted the U.K.

    Source: §3.3 · p. 4

Abstract

from arXiv · show

While the literature on permissions from the end-user perspective is rich, there is a lack of empirical research on why developers request permissions, their conceptualization of permissions, and how their perspectives compare with end-users' perspectives. Our study aims to address these gaps using a mixed-methods approach. Through interviews with 19 app developers and a survey of 309 Android and iOS end-users, we found that both groups shared similar concerns about unnecessary permissions breaking trust, damaging the app's reputation, and potentially allowing access to sensitive data. We also found that developer participants sometimes requested multiple permissions due to confusion about the scope of certain permissions or third-party library requirements. Additionally, most end-user participants believed they were responsible for granting a permission request, and it was their choice to do so, a belief shared by many developer participants. Our findings have implications for improving the permission ecosystem for both developers and end-users.

1 INTRODUCTION

The study addresses the lack of empirical evidence about how developers understand and choose app permissions, and how their views compare with end-users’. Interviews with developers and a follow-up end-user survey examine these perspectives and the challenges of integrating permissions.

Background

  • Permissions protect smartphone data and resources, while developers choose which permissions to include and end-users decide whether to grant them.Source: §1 · p. 1

Study aim and approach

  • The study fills a gap in empirical research on developers’ permission practices and compares their perspectives with end-users’ through interviews and a follow-up survey.Source: §1 · p. 1

Findings

  • Developers and end-users shared concerns that unnecessary permissions can harm trust and app reputation, while developers sometimes added multiple permissions because of scope confusion or third-party library requirements.Some developers added permissions to avoid crashes or unexpected consequences; libraries sometimes led apps to request permissions that developers did not expect.Source: §1 · p. 2
  • Developers primarily included permissions for app features, but rarely removed them; they also recognized that requesting too many could harm end-user experience and trust.Source: §1 · p. 2
  • End-users primarily granted permissions for features, and some did so because they trusted the app; 31.4% saw no harm in permissions, while others worried about unintended data use.57.9% had never removed granted permissions, citing no reason, not knowing how, or not having considered it.Source: §1 · p. 2

2 RELATED WORK

Prior work documents privacy concerns and unnecessary permissions, but offers little empirical evidence about developers’ permission perspectives. This study contributes interview-based insights and compares them with end-user views.

2.1 App Permissions

  • When permissions match expected app functionality, end-users are likely to grant them, yet unnecessary permissions can frustrate users and carry privacy ramifications.Prior work links unused or excessive permissions to development-stage changes, copied code, limited documentation, and confusion about permission names or scope.Source: §2.1 · p. 2

2.2 Empirical Privacy Studies With Developers

  • Earlier developer research focused on iOS permission descriptions in 2014, while this study adds in-depth interviews about developers’ understanding and decision-making and compares those findings with end-user perspectives.Related work also identifies permission management complexity and confusion about third-party libraries’ data practices and permission requests.Source: §2.1 · p. 2, §2.2 · pp. 2–3

3 METHOD

The study paired interviews with 19 developers and a survey of 309 smartphone end-users. Interview findings informed the survey, enabling comparison of the two stakeholder groups’ perspectives on permissions.

3 METHOD

  • The researchers first interviewed 19 developers, then used interview findings to design a survey of 309 smartphone end-users about permissions and developer-end-user comparisons.Source: §3 · p. 3

3.1.1 Developer Participant Recruitment.

  • Developer participants were recruited through professional and crowdsourcing platforms, screened for programming knowledge and at least two years of software-development experience, and 19 of 67 eligible invitees participated.Source: §3.1.1 · p. 3

3.1.2 Developer Participant Demographics.

  • The 19 developers had varied platform experience and app domains, and reported collecting data and using libraries according to app needs.Participants had an average of eight years of software-development experience; their apps included healthcare, utility, commerce, education, social media, gaming, and finance.Source: §3.1.2 · p. 3

3.1.3 Study Procedure.

  • Interviews explored developers’ backgrounds, permission understanding and decisions, permission updates, challenges, and views of common permissions, sensitive data, privacy, and laws.Interviews averaged 44 minutes, were conducted in English, and continued until the researchers judged data saturation had been reached.Source: §3.1.3 · p. 3

3.1.4 Data Analysis.

  • Two authors inductively coded all 19 interview transcripts, iteratively refined a codebook until saturation, and resolved coding disagreements.Average Cohen’s kappa across codes was .71; qualitative findings use cautious frequency qualifiers rather than reported occurrence counts.Source: §3.1.4 · pp. 3–4

3.2 Survey Study With End-Users

The end-user survey examined participants’ understanding, concerns, opinions, and interactions with smartphone permissions. Its design drew on the developer interviews and compared users’ views with developers’ assumptions and perspectives.

3.2 Survey Study With End-Users

  • The survey recruited U.K.-based Android and iOS users through Prolific, with a gender-balanced sample and English-language questions.The U.K. setting reflected sampling, privacy-context, resource, and translation considerations.Source: §3.2 · p. 4, §3.2.1 · p. 4

3.2.1 End-User Participant Recruitment.

  • The end-user sample was located in the U.K.; 51.1% used Android and 48.9% used iOS, and the average age was 37.Source: §3.2.2 · p. 4

3.2.2 End-User Participant Demographics.

  • Survey questions covered users’ understanding of permissions, perceived harms, opinions about common Android and iOS permissions, and interactions with permission requests.Questions were inspired by the interviews and, where relevant, adapted from prior surveys.Source: §3.2.3 · p. 4

3.2.3 Study Procedure.

  • The researchers descriptively analyzed closed-ended responses and collaboratively grouped open-ended responses by similarity.Source: §3.2.4 · p. 4

3.3 Limitations

The findings are bounded by nonrepresentative samples and a developer study that did not include a survey. The study also did not examine developers’ broader privacy-compliance processes.

3.3 Limitations

  • Developer participants were mostly male and from Asia, while the end-user survey targeted the U.K.; findings are not generalizable to all developers.The authors also note that neither sample necessarily represents all developers and end-users.Source: §3.3 · p. 4
  • The study did not survey developers or collect data on their apps and development modes, limiting quantitative comparison and contextualization.The authors suggest future surveys and collection of app and development-mode data.Source: §3.3 · p. 5
  • Privacy-law compliance was outside the study’s scope, which asked only whether developers changed permission requests to comply with privacy laws.All interviewed developers said they had made no such changes.Source: §3.3 · p. 5

4 FINDINGS

The findings examine how developers and end-users understand permissions as controls on resource access and navigate permission decisions involving privacy, trust, and user choice. They cover developers’ reasons for adding or adapting permissions and challenges integrating them, as well as end-users’ decisions to grant, keep, or remove permissions.

4.1 Commonly-Used Permissions

  • Among end-users, location was the most requested permission: 98.1% reported seeing it in apps they had frequently used in the past year; participants may remember it due to sensitivity or app needs.Camera (82.5%), photos (81.6%), microphone (64.1%), and contacts (54.7%) were next in line for commonly seen permission requests by end-users.Source: §4.1 · p. 5

4.2 Why Do Permissions Exist? What Do They Do?

  • Most developers described permissions as access control, and 87.7% of end-users identified them as granting access to smartphone resources or data.End-users also reported awareness, expected request counts, and understanding of requests, though these self-reports do not establish that they understood permission scope. Their explanations included functionality (35%), informing users (21%), legal compliance (19%), protection (18%), and data collection (13%).Source: §4.2 · pp. 5–7

4.3 Privacy Ramifications of Permissions

  • Developers saw location, camera, microphone, contacts, storage, and Internet permissions as potentially privacy-invasive, while recognizing Internet access as broadly necessary for apps.They described risks including location tracking and behavior inference, recording without notice, contacts sold to third parties, overly broad storage access, and undisclosed data sharing or vulnerabilities; some favored security testing, data minimization, and local processing.Source: §4.3 · p. 7
  • End-users associated permissions with tracking and monitoring (80.6%), unintended data use (77%), and possible harm from granting requests (68.6%).Among those seeing possible harm, 62% cited access to protected resources; others cited security issues (18%), selling data (12%), or uncertainty about developers’ data practices (3%). Some saw no harm because users had a choice (11%) or felt responsible for permission choices (5%).Source: §4.3 · pp. 7–8

4.4 Reasons for Adding, Updating, or Removing Permissions

  • Developers primarily added permissions for app features and requirements, but also included them because of libraries, routine code reuse, platform or store requirements, development changes, and tool errors.Reusing or copying code could leave unnecessary permissions; some developers also used workarounds such as IP-based location or app-list fingerprinting to avoid permission requests.Source: §4.4 · pp. 8–9
  • End-users most often granted permissions for functionality/features (79.6% vs. 68.2%), because apps might not work without them (43.7% vs. 23.8%), or due to trust (28.2% vs. 32.1%).Some also granted requests to make the screen go away (18.4% vs. 10.2%). Most had never or rarely changed settings (34% and 35%).Source: §4.4 · p. 9
  • End-users removed permissions when features were no longer needed (21%) or they felt worried or uncomfortable (16%), while some cited loss of trust or accidental granting (4% each).Other reasons included battery drainage (2%); among those who had never removed permissions, common reasons included no need (27%), confidence in granted permissions (10%), or not knowing how (8%).Source: §4.4 · pp. 9–10

4.5 Considering End-Users & Their Experience

  • Developers said they wrote or adapted clear permission descriptions to explain access, while end-users’ reported attention to descriptions varied: 39.8% read or checked them little or less, 30.4% moderately, and 29.8% a lot or more.Some developers added a separate explanatory screen because built-in dialogs offered little space, which could add clicks for users and work for developers.Source: Developer Participant Perspectives:… · p. 10, End-User Participant Perspectives · p. 10
  • Developers assumed users varied in permission sensitivity and awareness, yet many tried to request permissions only when needed and avoid flooding users with requests.One developer said users might choose an alternative app rather than grant SMS or email access, which they viewed as more sensitive than location access.Source: Developer Participant Perspectives · pp. 10–11
  • 81.2% disagreed that users did not care about permission requests, 79% disagreed that requests were unnecessary, and 87.4% agreed users were responsible for granting them.Many end-users also wanted more precise, transparent permission descriptions.Source: Developer Participant Perspectives · p. 11, Figure 3 · p. 12
  • Developers worried that excessive requests could erode trust, while end-users reported trusting apps with fewer permissions but had mixed views on whether requests signaled safety.One developer described misusing collected contact data, illustrating the potential for access granted in trust to be misused.Source: Developer Participant Perspectives · p. 11

4.6 Developers’ Challenges When Working With Permissions

  • Although most developers were satisfied with permission integration and relied on platform documentation, some wanted clearer guidance, better examples, and help tracking cross-platform changes.Some blindly trusted documentation samples, while a few had difficulty learning whether users granted access; privacy protections could prevent that information from being shared.Source: §4.6 · pp. 11–12
  • Unclear permission scopes led some developers to include similar permissions to avoid missing needed access or causing crashes, while manual updates and late feedback created rework.Unexpected library permissions sometimes surfaced only after submission or testing, and app-store rejection was a primary source of delayed feedback.Source: §4.6 · p. 11

4.7 Developers’ Reasons for (Not) Using Software Development Libraries

  • Developers chose libraries for functionality, routine practice, and endorsement by companies or platform creators, but excessive permissions could prompt removal or replacement.Some libraries facilitated permission management; others’ extra features and permissions made them unsuitable.Source: §4.7 · p. 12
  • Developers reconsidered libraries when opaque third-party data practices raised privacy concerns, and app-store rejection or poor user experience could also lead them to exclude libraries.One app-store review led a developer to remove an ad framework and its dependent analytics library.Source: §4.7 · pp. 12–13

5 DISCUSSION AND FUTURE WORK

The discussion compares developer and end-user perspectives on permissions, highlighting the tension between app functionality, privacy, and responsibility. It identifies scope confusion, third-party libraries, and limited revocation awareness as challenges, and suggests clearer descriptions and shared responsibility as directions for improvement.

5.1.1 Are Permission Requests Fully Grounded in App Functionality

  • 79.6% of end-users tended to grant permissions they considered necessary for app functionality, while developers could add excess permissions through scope confusion, crash concerns, or library requirements.The discussion also raises context-dependent granting and revocation, and notes that requiring developers to write their own libraries could undermine software reuse.Source: §5.1.1 · p. 13

5.1.2 Trust & Reputation as Decision-Making Factors.

  • Developers and end-users both connect fewer or necessary permissions with trust, while developers and most end-users place grant decisions with end-users despite developers’ central role in permission choices.The discussion calls for developer awareness and responsibility, but also argues responsibility should be shared across platforms, developers, and end-users.Source: §5.1.2 · p. 13, §5.1.3 · p. 13

5.1.4 Permission Descriptions and Names Need a Fresh Look.

  • Clearer, more transparent permission descriptions could support informed choices, but current dialogs leave limited space and custom pages risk inconsistency and extra developer work.Similar permission names can confuse developers about scope and lead them to include multiple permissions, potentially exposing more access than intended.Source: §5.1.4 · p. 14

5.1.5 Close the Gap Between End-Users’ Beliefs, Comprehension &

  • End-users’ reported awareness and understanding were not tested with follow-up comprehension questions, so belief may differ from actual knowledge.The discussion contrasts this limitation with a prior survey in which only 6.1% answered all comprehension questions correctly.Source: §5.1.5 · p. 14

5.2 The Usual Suspects: Third-Party Data Practices Worry Developers & End-Users

  • Developers may include libraries that request unneeded permissions, and a library without permission requests may still collect user data.The discussion suggests granular library features and transparent communication of data practices to developers and end-users.Source: §5.2 · p. 14

5.3 Forgotten Permissions: Possibly Rooted in Scope Confusions & Lack of Awareness of Revoking Procedures

  • Developers rarely removed permissions, citing crash or error concerns and scope confusion; 57.9% of end-users said they did not know how to remove permissions.The discussion proposes periodic notices or nudges about unused permissions and suggests examining Android’s automatic checks and the absence of automatic revocation in iOS.Source: §5.3 · pp. 14–15

6 CONCLUSION

The study combines developer interviews and an end-user survey to examine how both groups understand and interact with permissions. Participants connected permissions with app functionality, while developers sometimes included multiple permissions because of scope confusion or third-party library requirements.

6 CONCLUSION

  • Interviews with 19 developers and a survey of 309 Android and iOS end-users examined how both stakeholders understand and interact with permissions.Both groups associated permissions with app functionality; developers sometimes included multiple permissions because of scope confusion or third-party library requirements.Source: §6 · p. 15

A SCREENING SURVEY FOR THE INTERVIEW STUDY WITH DEVELOPERS

The screening survey included programming questions and other participant questions, with some answer options randomized and some questions timed.

A SCREENING SURVEY FOR THE INTERVIEW STUDY WITH DEVELOPERS

B DEMOGRAPHICS FOR THE INTERVIEW STUDY WITH DEVELOPERS

The appendix points readers to a table summarizing developer demographics.

B DEMOGRAPHICS FOR THE INTERVIEW STUDY WITH DEVELOPERS

C INTERVIEW GUIDE FOR THE INTERVIEW STUDY WITH DEVELOPERS

The interview guide asks developers about their work with apps and permissions, their decision-making and challenges, and their views on permissions’ effects on users.

C INTERVIEW GUIDE FOR THE INTERVIEW STUDY WITH DEVELOPERS

  • The guide asks about app and job backgrounds, including app types, user age groups, and data collected.It also asks whether developers share data between apps and why.Source: C INTERVIEW GUIDE FOR THE INTERVIEW… · p. 18
  • Developers were asked how they choose which permissions to include, whether they update or remove them, and what makes permissions confusing.The guide also asks what sources developers consult when resolving confusion.Source: C INTERVIEW GUIDE FOR THE INTERVIEW… · pp. 18–19
  • Developers were asked how permissions affect users’ app choices and decisions to accept or reject requests.The guide also asks about permission descriptions, privacy regulations, and ways to improve permission management and models.Source: C INTERVIEW GUIDE FOR THE INTERVIEW… · p. 19
  • Developers were asked how third-party libraries affect permission choices and what functionality those SDKs support.Prompts ask how developers learn which permissions libraries need and whether permissions have led them to exclude a library.Source: C INTERVIEW GUIDE FOR THE INTERVIEW… · p. 19
  • The guide probes developers’ understanding of permission behavior and their views on possible privacy consequences and sensitive user data.Participants reviewed a list of commonly used permissions and discussed what each does and whether it may affect users’ privacy.Source: C INTERVIEW GUIDE FOR THE INTERVIEW… · p. 19

D CODEBOOK FOR THE INTERVIEW STUDY WITH DEVELOPERS

The developer interview codebook is presented with occurrences counted as the number of unique participants mentioning each theme.

D CODEBOOK FOR THE INTERVIEW STUDY WITH DEVELOPERS

E QUESTIONNAIRE FOR THE SURVEY STUDY WITH END-USERS

The end-user questionnaire covers smartphone permission knowledge and experience, perceived privacy implications, and attitudes toward granting and managing permissions.

E QUESTIONNAIRE FOR THE SURVEY STUDY WITH END-USERS

  • The questionnaire asks how users understand permissions, which requests they have seen, and why they grant them.It also asks about awareness, expectations, and understanding of permissions in frequently used apps.Source: E QUESTIONNAIRE FOR THE SURVEY STUDY… · pp. 20–21
  • The survey asks about permission-related concerns, trust, choice, responsibility, and whether users have removed granted permissions.It also asks whether permission requests have led users not to install an app and whether current requests seem broad or generic.Source: E QUESTIONNAIRE FOR THE SURVEY STUDY… · p. 21
  • Users rate privacy implications for twelve permission categories, including photos, location, camera, contacts, and storage.The survey also asks how often they check permissions before installing apps and read permission explanations.Source: E QUESTIONNAIRE FOR THE SURVEY STUDY… · p. 21
  • The questionnaire asks how often users contact developers or change smartphone settings to manage permissions.It also includes a statement about whether users would be bothered by permission requests after installing an app.Source: E QUESTIONNAIRE FOR THE SURVEY STUDY… · p. 21
  • Users report concerns about monitoring, tracking, unwanted access, and unauthorized use or sharing of personal information.These items are framed as statements about possible consequences of accepting app permissions.Source: E QUESTIONNAIRE FOR THE SURVEY STUDY… · p. 21

F ADDITIONAL PLOTS FOR THE SURVEY STUDY WITH END-USERS

The appendix figures show end-user responses about privacy implications, permission-related privacy ramifications, developer-study topics, frequently seen requests, and reasons for granting permissions.

F ADDITIONAL PLOTS FOR THE SURVEY STUDY WITH END-USERS

  • Figure 4 presents end-user responses about the extent to which granting permissions has privacy implications.The prompt asks about privacy implications of granting permissions on a device.Source: Figure 4 · p. 22
  • Figure 5 presents end-user answers to questions about permission privacy ramifications.The caption does not specify the individual questions or response values.Source: Figure 5 · p. 22
  • Figure 6 presents end-user answers to questions rooted in the developer study.The caption does not specify the individual questions or response values.Source: Figure 6 · p. 23
  • Figure 7 asks which five permission requests end-users frequently saw during the previous year.The prompt concerns perceived frequency of permission requests on participants’ smartphones.Source: Figure 7 · p. 23
  • Figure 8 presents end-users’ selected reasons for granting smartphone app permissions.The question allows participants to select all applicable reasons.Source: Figure 8 · p. 24

Related papers

Loading 2301.06534v2…