The phrase “IRB is a floor, not a ceiling” really stuck with me. It challenges the idea that ethical approval equals ethical research and instead frames ethics as an ongoing responsibility. This feels especially relevant for digital research, where new risks can emerge after a study has already begun.

The idea of informational risk stood out to me because it reframes harm as something that can emerge later, even if a study seems harmless at first. The fact that “anonymized” data can be re-identified suggests that ethical responsibility doesn’t end at data collection but extends to long-term data storage and sharing decisions.

I found the principles-based approach compelling because it acknowledges that rules and laws often lag behind technological capabilities. Rather than offering rigid answers, these principles seem designed to help researchers reason through uncertainty, which feels especially important in digital research where consequences aren’t always predictable.

The emotional contagion experiment raises a challenging question about informed consent in platform-based research. While users technically agreed to Facebook’s terms of service, this feels very different from actively consenting to be part of an experiment that alters emotional content. It makes me wonder whether consent should be contextual rather than purely legal.

The use of real-world cases here makes it clear that ethical issues in digital research aren’t hypothetical but that they emerge from ordinary research decisions. What stood out to me is that all three examples involve studies that were innovative but controversial, which suggests that ethical risk often increases alongside methodological ambition.

This framing helps clarify what we’ll be doing in this course, not just learning techniques, but learning how to evaluate research choices. It makes me think about how we’ll need to justify data sources, sampling decisions, and ethics in our own projects.

The contrast between the optimism of data scientists and the skepticism of social scientists stood out to me. I like the idea of balancing both perspectives; being critical without dismissing new methods outright. That balance feels essential for responsible computational social science.

This raises ethical questions for me about informed consent. If researchers are using data generated while people are “living their lives,” how aware are participants that they’re being studied? I’m curious how the book later addresses privacy and consent in digital contexts.

This example really highlights the methodological shift from traditional lab experiments to digital-age research. It’s striking how scale and speed change what’s possible but it also makes me wonder how issues like sample bias or lack of experimental control compare to in-person lab studies.

I appreciate that ethics is treated as both its own chapter and something embedded throughout the book. That structure reinforces the idea that ethical considerations shouldn’t be isolated to a single stage of research, especially when methods like experiments or mass collaboration introduce new responsibilities for researchers.

I found the framing of the four research designs helpful because it clarifies how different methods enable fundamentally different kinds of knowledge. The idea that collaboration allows learning that isn’t possible through observation, surveys, or experiments alone makes me think about how method choice shapes not just results, but the kinds of questions we even consider asking.

The discussion of researcher power stood out to me, especially the idea that researchers can now affect people’s lives without their awareness or consent. This feels like a major shift from traditional social research and makes ethics feel less like an afterthought and more like a core design constraint in digital-age studies.

The Duchamp vs. Michelangelo analogy makes the readymade vs. custommade distinction really concrete. I like how this reframes data sources as design choices rather than shortcuts or “less rigorous” options. It also raises a question for me about trade-offs: when repurposing readymade data, how do researchers ensure alignment between the original purpose of the data and their research question?