Do I Need a Digital Intervention? | Monday Advice

What It Covers

Cal Newport examines a randomized controlled trial published in PNAS Nexus showing that blocking mobile internet apps for 14 days reduces anxiety and depression, raises subjective well-being, and improves sustained attention. Newport presents the study's mechanisms and three practical strategies for maintaining compliance throughout the intervention period.

Key Questions Answered

• •Mobile internet blocking results: A randomized controlled trial of roughly 500 participants found that blocking internet-based apps for 14 days produced measurable jumps in sustained attention, mental health scores, and subjective well-being. Effects partially persisted even after participants resumed normal phone use, suggesting short-term abstinence creates lasting neurological benefit.
• •Screen time reduction mechanism: Blocking internet apps cut average daily screen time from 304 minutes to 161 minutes — roughly a 50% reduction. Researchers identified four mediating factors explaining the benefits: increased time in meaningful offline activities, more social interaction, longer sleep duration, and a heightened sense of personal self-control over daily behavior.
• •Targeted app blocking strategy: Block specifically social media, news, and games — Newport's "SNG" category — rather than all internet access. Leave functional apps like two-factor authentication, parking payment tools, and messaging intact. Broad blocking creates practical friction that causes participants to abandon the intervention entirely, which explains much of the 74.5% noncompliance rate.
• •Compliance friction tools: Only 25.5% of study participants maintained full compliance. To improve odds, Newport recommends physical friction devices like Brick, which requires tapping a separate key fob to disable blocking, or configuring iOS Screen Time so a partner holds the PIN code, making circumvention socially costly and requiring deliberate effort rather than a single tap.
• •AI research quality degradation: A separate PNAS Nexus-adjacent study from the journal Organization Science found that post-ChatGPT, high-AI manuscripts (70%+ AI-generated) face nearly 70% desk rejection rates versus 44% for low-AI papers, and only 4% reach revise-and-resubmit status compared to 12% for low-AI submissions — demonstrating that faster production does not equal higher output quality.