Mixed Methods Research Plan

Pedagogical Friction
in the Age of Generative AI

A qualitative-dominant convergent mixed methods case study in Merriam's interpretive tradition, examining how K-12 educators and school-system leaders make sense of pedagogical friction under conditions of generative AI.

Explore the Study

Study Overview

The Purpose

To investigate how K-12 educators and school-system leaders understand, navigate, and respond to the friction-reducing affordances of GenAI in academic work.

The Stance

Grounded in a Merriam-centered case design and supported by mixed-methods pragmatism, the study treats educator and institutional sensemaking as the conceptually and temporally bounded case.

Primary Research Questions

RQ1

How do K–12 teachers understand and navigate the friction-reducing affordances of generative AI?

RQ2

What institutional conditions enable or constrain friction-preserving pedagogy?

RQ3

How can the Pedagogical Friction Framework inform AI policy development in K–12?

The Evolution of Media Ecology

Extending Walter Ong's framework to account for generative artificial intelligence.

01

Primary Orality

Memory-based, communal, and situational. Knowledge is preserved through repetition and mnemonic devices.

02

Literacy

Writing externalizes memory, enabling analytical detachment, individual authorship, and abstract reasoning.

03

Secondary Orality

Electronic broadcast media (radio/TV) retrieves communal qualities, but operates through one-to-many literate infrastructure.

Proposed Extension
04

Algorithmic Secondary Orality

Social media platforms where humans create symbolic content, but opaque algorithms determine distribution and curation.

Categorical Rupture
05

Tertiary Algorithmicity

Neural networks generate the symbolic content itself, making human authorship optional at scale. The source of content is no longer human.

Tertiary Algorithmicity

Three defining characteristics that threaten the cognitive processes education depends on.

Noetic Displacement

The externalization of cognitive labor moves from storage to generation. The work of synthesizing and making meaning is offloaded to the system rather than occurring in the learner's mind.

Rhetorical Saturation

The flooding of communicative environments with synthetic discourse. The origin of content becomes uncertain, and simulated interlocutors replace genuine contestation.

Existential Abstraction

Symbolic production is severed from lived experience. Text is produced without personal commitment, intellectual risk, or accountability for the claims advanced.

The Threat: Unproductive Success

When generative AI is used to bypass these cognitive processes, the result is what Manu Kapur (2016) calls Unproductive Success. A student generates fluent, correct output without ever engaging in the schema construction required for durable understanding. The artifact of learning is visible, but the process of learning has been bypassed.

The Pedagogical Friction Framework

Four dimensions of resistance necessary for durable learning.

Noetic Dimension

Cognitive struggle. The necessary resistance encountered when wrestling with complex ideas, synthesizing information, and engaging in deliberate practice.

Rhetorical Dimension

Engagement with real audiences. The friction of translating internal thought into a structure that an external audience can understand.

Existential Dimension

Intellectual ownership. The personal stake and authorial stance that connects the learner's identity to the work produced.

Infrastructural Dimension

Policy and institutional conditions that either enable or constrain the preservation of the other three dimensions of friction.

Merriam-Centered Mixed-Methods Design

A qualitative-dominant convergent mixed methods case study in Merriam's interpretive tradition, bounded conceptually and temporally rather than by a single site.

Methodological Alignment Studio

The dissertation's controlling methods language is Merriam-centered: the case is K-12 educator and institutional sensemaking about pedagogical friction under conditions of generative AI. Stake and Yin remain useful comparison lenses, but quantitative, learner-perspective, and AI-artifact strands support the qualitative case rather than redefining it.

Open the Methodology Studio Memoing & AI-use protocol See how the literature review leads here

Phase 1: Concurrent Data Collection

Interviews with classroom-facing educators, building administrators, district/system leaders, and a bounded learner-perspective strand; card-sort and document protocols; survey data; and secondary data compilation.

Phase 2: Independent Analysis

Descriptive and thematic qualitative analysis through the Pedagogical Friction lens, with descriptive analysis, disaggregation, and cautious cross-tabulation of quantitative data.

Phase 3: Integration

Joint displays comparing qualitative themes with quantitative patterns, documenting convergence, expansion, and divergence across educator sensemaking, institutional conditions, learner retrospection, and secondary indicators.

Phase 4: AI Comparison

Administering selected prompts to AI platforms as a bounded artifact comparison between situated human pedagogical reasoning and decontextualized synthetic discourse.

Matched Participant Pairs

Exploring both sides of the friction equation.

Those Who Experience Friction

University Students (N=4)

Students who were in high school (10th-12th grade) during the 2022-2023 academic year when GenAI became publicly available.

  • Provide retrospective accounts of the media transition.
  • Illuminate the Learner Perspective.
  • Address the Secondary Research Question (SRQ).

Those Who Govern Friction

Practitioners & Leaders (N=7-8)

K-12 Teachers, Building Administrators, and District Leaders navigating current practice under "tertiary algorithmicity".

  • Provide accounts of professional reasoning & policy.
  • Illuminate the Institutional Perspective.
  • Address Primary Research Questions (RQ1, RQ2, RQ3).

Supplementary AI Discourse Comparison

Testing theoretical claims about what AI-generated pedagogical reasoning lacks.

discourse-analysis.sh
$ run comparison --models="ChatGPT, Gemini, Claude"
Evaluating human vs AI responses to identical research protocols...

INDICATOR 1: Experiential Specificity
HUMAN: Grounds reasoning in specific incidents and students.
AI: Generates generic pedagogical advice without lived context.

INDICATOR 2: Temporal Authenticity (Student Transition)
HUMAN: Conveys the phenomenological texture of living through a media shift.
AI: Produces a clean before/after narrative lacking messy experiential reality.

INDICATOR 3: Professional Ambivalence
HUMAN: Expresses unresolved tension regarding GenAI integration.
AI: Resolves tensions prematurely into artificially balanced conclusions.

Analysis Complete. AI outputs function as analytical artifacts confirming the Pedagogical Friction Framework's claims.
_

Companion Ecosystem

A digital resource hub expanding on the Qualifying Paper and Dissertation.