With Visual Integrity and Care: A Framework for Mixed Methods Research on Visual Social Data
A strong best-paper contribution that reframes visual social data research as a methodological and ethical problem, not just a tooling gap. Its main value is the integration of visual grammars, human interpretation, computation, and care into a coherent reusable framework.
Axes Lens
Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.
Contribution shape
- Knowledge form
- method knowledge typical · 29/268
- Novelty type
- framework typical · 59/268
- Abstraction level
- practice typical · 85/268
- Generalization target
- field argument typical · 55/268
- Validation mode
- mixed methods typical · 136/268
Evidence profile
- Evidence strength
- strong typical · 158/268
- Claim alignment
- strong typical · 231/268
- Overclaim risk
- low typical · 53/268
Review Summary
This paper makes a substantial methodological contribution by treating visual social data research as something that cannot simply inherit text-centric social computing methods or be handed off to automated vision systems. Instead, it proposes a grounded, interpretivist framework that organizes inquiry around three linked components: visual grammars, human analysis, and computationally supported analysis, all under an explicit commitment to care. That combination is the paper’s real contribution. It is not merely saying that images matter; it is specifying how researchers can systematically study them while preserving visual integrity and protecting teams exposed to harmful material. The framework is also unusually well supported for a methodological paper because the authors show repeated application across three distinct case studies rather than a single illustrative example. At the same time, the paper is careful about scope. It does not claim universal metrics or broad statistical generalizability, and it acknowledges that its interpretivist orientation makes it unsuitable for some research questions. It also notes contextual limits around Global North settings, English-speaking research teams, and the need for further work on video and other analysis modes. Those limitations make the contribution more credible, not less. Overall, this is a field-shaping CHI paper because it gives HCI a concrete way to study increasingly visual online environments without collapsing visual meaning into text proxies or ignoring the labor and harm involved in doing this work well.
What Changed
Canon before
Social computing research and methodological training have heavily relied on textual methods, with a dominant assumption that textual analysis suffices for social media research despite the increasing prominence of visual content.
Departure from common sense
The paper argues against the default assumption that textual methods are enough for studying social media, claiming that visual social data requires a dedicated mixed-methods framework centered on visual analysis, human interpretation, and researcher care.
Actual novelty
The contribution is a methodological framework for grounded, interpretivist, computationally supported mixed-method research on visual social data. Its novelty lies in explicitly combining visual grammars, human analysis, and computationally supported analysis under a commitment to care and showing the framework across three case studies.
Evidence
Evidence comes from the introduction and methods framing that identify a gap in text-dominant social media research, the framework section that specifies the three components and care commitment, the case studies section that documents repeated application across three empirical projects, and the limitations section that explicitly bounds generalizability and applicability.
“In this section, we detail our research framework, which supports a grounded, interpretivist approach to analyzing “big” visual social data that deeply integrates mixed (qualitative and quantitative) analyses. This approach incorporates multiple, often iterative, analyses that build upon one another (although we present them in a more linear style here)”
actual novelty · 3 Components and commitment of our framework · confidence 0.97
“ However, social media research has largely relied upon textual analysis [132], leaving a methodological gap in equipping researchers to rigorously study the visual content shaping our on and offline lives”
departure from common sense · 1 Introduction · confidence 0.95
“Our framework does not have quantifiable or fully generalizable metrics. We follow traditions of qualitative and interpretivist research focusing on transferability over generalizability [38, 54] but acknowledge this could make our framework inappropriate for some research questions”
limitation · 6 Limitations and Future Work · confidence 0.94
“Next, we walk through three case studies of problematic information where we have applied this framework: (1) Anti-immigrant visual propaganda (4.1): Contemporaneous, weekly analysis of visual anti-immigrant propaganda on X and TikTok in the US across 2024, a Presidential Election year. (2) Jesus AI Slop Imagery on Facebook (4.”
validation scope · 4 Case studies · confidence 0.96
Limits
Method limits
The framework is explicitly interpretivist and qualitative, prioritizing transferability over generalizability and therefore not fitting all research questions. It also depends on substantial human interpretation, training, and iterative mixed-method coordination.
Deployment limits
The demonstrated cases are concentrated in Global North contexts and were conducted by English-speaking researchers in American and European universities. The paper also notes that more work is needed for video and that other human-analysis modes remain unexplored.
Boundary conditions
The framework is most appropriate when researchers need grounded visual analysis rather than purely statistical inference, can support collaborative human analysis, and can maintain care practices while integrating computation in service of human interpretation.
Position in field
This paper positions itself as a methodological intervention in HCI and social computing, addressing the relative lack of visual methodological innovations compared with text-centered approaches and proposing a reusable framework intended to shape future visual research norms and training.