AFI 2024 : METHODOLOGY (APP STORE)

I.

Introduction and Definition of App Freedom

App Freedom measures the extent to which users in a given region can freely access, download, and use applications from the App Store without undue government interference, censorship, or restrictive platform policies. The App Freedom Index (AFI) evaluates and ranks countries or territories based on their mobile app environments, focusing on both the availability of a curated set of sensitive apps and the broader landscape of app availability.

This index is designed to move beyond a narrow focus on censorship and app removals, offering a more holistic understanding of digital freedoms. Rather than concentrating exclusively on specific cases of censorship, it evaluates how accessible a broad spectrum of apps can be— especially those that uphold fundamental rights such as freedom of expression, privacy, and access to information. In doing so, the App Freedom Index provides a more nuanced picture of the mobile app environment, reflecting the true depth and breadth of users’ digital freedoms.

The index also aims to highlight how government interventions, platform governance decisions, and other factors influence users’ digital freedoms. By incorporating data not only on specific sets of ‘sensitive’ or otherwise significant app categories, but also on the broader app availability landscape and government takedown requests, the AFI provides a comprehensive picture of app freedom worldwide.

II.

Scope and Objectives

ASM-based Data Published Online: Click Here

The AFI for 2024 focuses on two main components. The first component is the App Score, derived from a curated set of 1,000 iOS applications. The second component is the Availability Score, drawn from a broad measure of app unavailability rates to ensure that the index does not rely solely on the curated set. By combining these two measures, the AFI captures both a targeted and a general snapshot of a region’s app environment. In addition, data on government takedown requests, sourced from Apple’s Transparency Reports, provide a third, external dimension to the analysis, directly reflecting the impact of state interventions on app availability.

Objectives:

  • Evaluate a curated set of 1,000 apps organized into ten sensitive or thematically significant categories. These categories were chosen because apps in these domains are more frequently targeted for restrictions in repressive environments, or resemble other apps—or even websites—that have been censored or blocked in specific regions.
  • Incorporate data on broader app unavailability rates for each region, as provided by AppleCensorship.com. This complements the focused assessment of the 1,000 curated apps with a more holistic metric.
  • Integrate data from Apple’s Transparency Reports, which detail how many apps have been removed due to government requests. This reveals the extent to which state authorities shape the app market.
  • Produce a normalized, 0–100 scale score for each region, enabling straightforward comparison and classification into tiers of app freedom.

Data Sources:

III.

Measuring App Availability with the App Store Monitor (ASM)

The first key component of the AFI is quantifying how many of the 1,000 selected apps are available in each region’s App Store. Since directly assessing all two million apps worldwide is not feasible, the methodology focuses on a curated sample spanning ten thematic categories. These categories include content and functionalities often affected by censorship or restrictions because of their importance in public discourse, cultural and religious diversity, freedom of expression, and protection of personal freedoms and privacy.

III.I) App Categories

The 1,000 selected applications are divided evenly into ten categories. Although not every app is necessarily sensitive or vital, the categories were chosen because they represent areas that are either prone to censorship or crucial for fundamental freedoms. For example, news and media apps can be censored due to political content, VPN and proxy apps help users bypass restrictions, and religious or cultural apps may face restrictions in regions hostile to certain minorities.

Categories

Category Description
News, Media & Information Key sources for press freedom and information dissemination
Religion and Culture Apps promoting diverse religious and cultural practices
LGBTQIA+ & Dating Apps supporting diverse relationships and social inclusivity
Social Media Platforms central to public discourse and political activism
VPN and Proxy Services Tools for bypassing restrictions and ensuring private internet access
Digital Security & Privacy Privacy, encryption, and security-focused apps
Communication Messaging, VoIP, phone, and email services critical for free expression
Human Rights, Intl. Orgs & Civil Society Apps advancing civic awareness and human rights
Education & Public Health Tools providing educational resources and public health information
Miscellaneous Other relevant apps that highlight broader censorship trends

III.II) Scoring of App Availability

For each of the 1,000 apps, its availability in the region’s App Store is assessed.

  • Each available app: 1 point
  • Each unavailable app: 0 points

The total points for each category (out of 100 apps) are summed and then converted into an average availability percentage for that category. After computing the average scores for all ten categories, these category scores are averaged to yield the Overall App Score on a 0–100 scale.

App Availability Scoring

Data Points:

- App availability (Available = 1 point, Unavailable = 0 points)

Steps to Calculate the App Score:

  1. Calculate Category Scores:

    • Sum the availability points for the 100 apps in each category.
    • Convert this sum into a percentage (e.g., 80 available out of 100 = 80%).

  2. Calculate Overall App Score:

    • Compute the average of the ten category percentages.
    • The Overall App Score will be between 0 and 100.

Example (Country A):

  • Suppose Country A’s “News, Media & Information” category has 85 available apps out of 100: Category Score = 85%
  • After computing all ten category scores, Country A’s overall average might be 90%
  • Thus, Country A’s Overall App Score = 90.

Formula:

Overall App Score = (Sum of the 10 Category Percentages) ÷ 10

III.III) Unavailability Rate and the Availability Score from AppleCensorship.com

Beyond the curated set of 1,000 apps, the AFI also considers a broader indicator: the Unavailability Rate from AppleCensorship.com. This rate is the proportion of tested apps that are unavailable in a given Store.

To transform this into a positive measure, we compute:

Availability Score = 100 – Unavailability Rate

For example, if 30% of tested apps are unavailable, the Availability Score is 70.

This Availability Score provides a more general sense of the app ecosystem’s openness. Even if a region scores well on the 1,000 curated apps, a high overall unavailability rate indicates broader restrictions.

Availability Score Calculation

Data Points:

  • Unavailability Rate: Percentage of tested apps unavailable in a given Store.

Steps to Calculate the Availability Score:

  1. Obtain the Unavailability Rate from AppleCensorship.com

  2. Convert it to an Availability Score:


    Availability Score = 100 – Unavailability Rate


Example (Country A):

  • Suppose Country A has a 20% Unavailability Rate.
  • Availability Score = 100 – 20 = 80

III.IV) Combining the App Score and the Availability Score

Next, we combine the App Score (focused measure) and the Availability Score (broad measure) into a Preliminary Score. The App Score is weighted at 70%, reflecting its importance in identifying targeted censorship, while the Availability Score is weighted at 30%.

Preliminary Score = (0.7 × App Score) + (0.3 × Availability Score)


Combining App and Availability Scores

Weights:

  • App Score: 70%
  • Availability Score: 30%

Formula:

Preliminary Score = (0.7 × App Score) + (0.3 × Availability Score)


Example (Country A):

  • From previous steps:
    • App Score (Country A) = 90
    • Availability Score (Country A) = 80

Preliminary Score = (0.7 × 90) + (0.3 × 80) = 63 + 24 = 87

IV.

Incorporating Apple’s Transparency Reports

A critical dimension of the AFI involves evaluating government interventions through Apple’s Transparency Reports, which detail how many apps have been removed due to official requests. By averaging these requests over about 5.5 years (H2 2018–2023), an annualized figure is obtained.

A penalty is applied to the Preliminary Score based on these annualized takedown requests, up to a maximum of 10 points. This penalty is proportional to the number of requests and ensures that government actions cannot boost a region’s score, only maintain or reduce it.


Penalty Calculation:

The 1,000 selected applications are divided evenly into ten categories. Although not every app is necessarily sensitive or vital, the categories were chosen because they represent areas that are either prone to censorship or crucial for fundamental freedoms. For example, news and media apps can be censored due to political content, VPN and proxy apps help users bypass restrictions, and religious or cultural apps may face restrictions in regions hostile to certain minorities.

  • 0 requests/year: No penalty
  • 1 to 10 requests/year: Penalty increases linearly from 0 to 2 points
  • 11 to 50 requests/year: Penalty increases linearly from 2 to 10 points
  • 50 requests/year: Maximum penalty of 10 points

Government App Takedown Penalty

Data Points

  • Annualized number of government requests for app removal (based on Apple Transparency Reports)

Penalty Ranges:

  • 0 requests/year: Penalty = 0
  • 1–10 requests/year: Penalty = up to 2 points
  • 11–50 requests/year: Penalty = 2 to 10 points
  • 50 requests/year: Penalty = 10 points (max)

Example (Country A):

  • Suppose Country A averages 5 government requests/year.
  • At 5 requests/year (midway in the 1–10 range), Penalty ≈ 1 point.

V.

Aggregating the Final AFI Score

After adjusting for government interventions, the final AFI score is computed. The formula initially presented in Section IV (Government App Takedown Penalty) and the combined weighting in Section III.IV both feed into a single cohesive formula:

Final Index Score = (0.7 × App Score) + (0.3 × Availability Score) – Government App Takedown Penalty

This final formula differs slightly in wording from the preliminary calculation, but the essence is the same. The weights (70% for the App Score and 30% for the Availability Score) remain consistent, and the penalty is then applied. Note that this final weighting structure is effectively the same as the preliminary calculation followed by the penalty step.

Final Index Score Calculation


Weights:

  • App Score: 70%
  • Availability Score: 30%
  • Government App Takedown Penalty: 0 to 10 points subtracted

Formula:

Final Index Score = (0.7 × App Score) + (0.3 × Availability Score) – Penalty


Example (Country A):

  • App Score: 90
  • Availability Score: 80
  • Preliminary Score, applying 70%/30% weighting: (0.7× 90) + (0.3 × 80).= 63 + 24 = 87
  • Subtract the penalty (1 point for 5 requests/year):
  • Final Index Score = 87 – 1 = 86

VI.

Ranking Countries and Assigning Tiers

After calculating the Final Index Score for all 175 regions, they are ranked from highest to lowest. The region with the highest score is ranked 1st, and the one with the lowest is ranked 175th. To facilitate comparisons, regions are grouped into tiers based on their final scores:


Tier 1 (High Freedom):
Final Index Score ≥ 98
These regions demonstrate extremely broad app availability, minimal government interference, and a consistently open environment.


Tier 2 (Moderately High Freedom):
Final Index Score 94–97.99
These regions have strong app availability, generally low government intervention, and few restrictive policies, though not quite at the topmost levels.


Tier 3 (Moderate Freedom):
Final Index Score 90–93.99
These regions display moderately good app availability but may face sporadic government requests or subtle platform restrictions.


Tier 4 (Low Freedom):
Final Index Score < 90
These regions experience more significant limitations, higher government interference, or notable platform-imposed restrictions on app availability.


In the event of a tie:

  1. The region with the higher App Score is ranked higher.
  2. If still tied, the region with the fewer Government Takedowns is ranked higher.
  3. If still tied, alphabetical order is used.

Ranking and Tier Assignment


Sorting and Tiering:

  • Sort Final Index Scores from highest to lowest.
  • Assign ranks (1 to 175).
  • Group into tiers:

Tier 1: ≥ 98
Tier 2: 94–97.99
Tier 3: 90–93.99
Tier 4: < 90


Example (Country A):

  1. Fewer Government Takedown Requests
  2. Higher App Score
  3. Alphabetical Order

Example (Country A):

  • Country A’s Final Index Score = 86
  • This places Country A in Tier 4 (Low Freedom).

If Country A ties with another region, compare and App Availability Scores and Transparency Scores. If those are also equal, list alphabetically.

VII.

Interpretation and Limitations

By standardizing data sources and employing a transparent methodology, the AFI offers a practical tool to monitor app freedom trends worldwide. However, it also has limitations:


  • The curated selection of apps introduces bias.
  • The penalty scale is linear and cannot differentiate between legitimate and illegitimate government requests.
  • It partly relies on the accuracy of data from Apple and external monitoring organizations.
  • Continuous refinement and normalization of scores are needed.

Despite these caveats, this initial methodology can serve as a foundational starting point on which we can continue to build and refine. It aims to offer policymakers, researchers, activists, and the public insights into the current state of digital freedoms, potentially helping to track changes over time and identify areas where further intervention or reforms might prove beneficial.