Complexity Explorer Santa Few Institute

Foundations & Applications of Humanities Analytics (fall 2021)

Lead instructor:

This course is no longer in session.

15.1 Guest Lecture: Nan Z. Da » Test Your Knowledge: Explanations

Q1. What is the conclusion of the study on local newspapers and municipal bonds described in the lecture?

A. U.S. cities whose local newspapers generated more revenue had more highly-valued municipal bonds.
B. U.S. cities that experience more local newspaper closures paid lower borrowing costs when issuing municipal bonds.

C. U.S. cities that experience more local newspaper closures paid higher borrowing costs when issuing municipal bonds.
D. None of the above.

Correct answer: (C)   Da cites this study in order to demonstrate that quantitative analysis can elucidate the value of cultural output, provided that said analysis is done well and makes explicit the sense in which the value of a cultural output is being measured, without assuming that this measurement is totalizing. The study also illustrates how the intentions of certain actors (i.e., the intentions of public officials when they are or are not held accountable by a local newspaper) can be analyzed quantitatively. 

The paper Da cites can be viewed here.


Q2. What is Da's primary critique of the use of metadata in cultural analytics projects that aim to measure bais on the part of authors?

A. Most metadata on the identities of authors is highly discretized, with only a few possible authorial races, genders, etc., represented, which limits the power of statistical analyses.
B. In general, there is too much metadata on most authors for this metadata to be managed by existing statistical software packages.
C. Most digital humanities data sets provide no metadata about the authors of different parts of a corpora.
D. None of the above.

Correct answer: (A)   As is clear from the lecture, Da is concerned that most author metadata is essentially binary (e.g., male or female, white or non-white) which limits the statistical power of any given analysis. This also reflects what to her is a significant abstraction from the real-world dynamics of creativity in which subtle differences in authorial identity can have significant impacts on output. 


Q3. What is Da's primary critique of viewing digital humanities data sets as "utilitarian structures", particularly with respect to how this view of data influences measurement?

A. By viewing digital humanities data sets as utilitarian structures, we lose the ability to make any nuanced measurement of any aspect of a corpus.
B. By viewing digital humanities data sets as utilitarian structures, we come to see everything within a corpus as easily quantifiable.

C. By viewing digital humanities data sets as utilitarian structures, we bias ourselves towards taking "easy", outcome-focused measurements that neglect any understanding of authorial intent.
D. None of the above.

Correct answer: (C)   Da's primary issue with a "utilitarian structures" understanding of cultural analytics is that it lends itself to outcome-based measurements in which the intent of an author does not matter. Options A and B are too strong:  while an outcome-based approach does seem to bias us towards quantification and (arguably) a lack of nuance, it would be unfair to say that, even if we take an outcome-based approach to all humanistic data, everything within a corpus becomes easily quantifiable, or that nothing nuanced can be said.