Paper 2: Measuring Neglect: Instruments, Biases, and a Proposed Attention Index for Pacific War Operations


1. Why This Paper Comes First

The suite’s governing claim is that attention to Pacific War operations is distributed in a patterned way that does not track operational weight, casualties, or strategic consequence. That is an empirical claim. It can be wrong. Everything in Parts II and III depends on it, and if the distribution turns out to be roughly proportional to some defensible measure of importance, the mechanism papers are explaining a phenomenon that does not exist.

The temptation in work of this kind is to treat the empirical claim as obvious and move directly to interpretation. Anyone who has spent time with popular Pacific War material knows that Kwajalein is not discussed and Midway is. But “knows” here means “has an impression,” and impressions of attention distributions are unreliable in a specific and predictable way: an observer’s sense of what is neglected is shaped by his own reading, which is shaped by his own interests, which are precisely what led him to notice the neglect. The specialist who has read deeply on Eniwetok will experience the general silence about Eniwetok more acutely than the silence about Wakde, which he has not read about either.

So the measurement problem is not merely a matter of rigor for its own sake. It is a correction for a bias the author of a neglect argument is guaranteed to have.

This paper does three things. It specifies a set of instruments for measuring attention. It states what each instrument is actually measuring and where it fails. And it proposes a composite index and a unit-of-analysis scheme that the rest of the suite will use, so that later papers can refer to measured positions rather than reasserting impressions.


2. What Is Being Measured

Attention here means the volume of contemporary output — published, broadcast, exhibited, uploaded, taught, or preserved — that takes a given operation as its subject. It is deliberately not “importance,” “significance,” or “memory.” Attention is the observable trace; memory is the cultural formation that produces and is produced by it. Attention is countable. Memory is not, at least not directly.

This distinction has to be held firmly, because the interesting findings live in the gap between them. A place can be intensely remembered locally and generate almost no measurable attention in the instruments used here, which is exactly the situation Paper 12 examines for Guam and the Marshalls. A battle can generate high attention through a single adaptation without being deeply held in memory at all. Conflating the two would make both findings invisible.

Neglect is therefore a relational term, not an absolute one. No operation receives zero attention. The claim is about distribution: that attention is concentrated on a small number of operations far beyond what any defensible weighting of operational significance would predict, and that the concentration follows a pattern.


3. Unit of Analysis

The suite needs a stable list of operations to compare, and constructing it involves choices that affect results.

The proposed unit is the named operation or fleet action, with the following inclusion rules. An entry qualifies if it was assigned a code name at the theater level, or if it constituted a discrete fleet engagement, or if it involved an opposed landing of division strength or greater. Sub-operations within a larger operation are broken out where they were separately planned and separately assaulted — so Forager appears as Saipan, Guam, and Tinian rather than as a single entry, because those are the units at which attention actually attaches.

This produces roughly forty-five entries for the Pacific War as a whole and about sixteen for the Central Pacific Advance as bounded in Paper 1. The full list belongs in an appendix; the working core for the campaign is Makin, Tarawa, Kwajalein (Roi-Namur and Kwajalein Island), Majuro, Eniwetok (Engebi, Eniwetok Island, Parry), Hailstone, Saipan, Philippine Sea, Guam, Tinian, Peleliu, Angaur, and Ulithi.

Three problems with this scheme should be acknowledged rather than solved.

First, the bypass problem. Operations that consisted of not attacking — the neutralization of Rabaul, Truk, and the eastern Marshall atolls — have no natural unit. They are not battles and they generate no operation-level attention, which is Paper 9’s entire subject. The index cannot represent them, and the suite should say plainly that its central instrument is blind to the campaign’s signature strategic achievement. That blindness is itself a finding about how attention works.

Second, the aggregation problem. Some operations are attended to as parts of larger wholes — Guam’s attention is partly absorbed into “the Marianas,” Angaur’s almost entirely into “Peleliu.” Disaggregating produces artificially low counts for the absorbed member; aggregating hides the internal asymmetry that is precisely the phenomenon of interest. The paper’s approach is to measure at the disaggregated level and report absorption as a separate observation.

Third, the naming problem. Several operations carry multiple names in common usage — Ulithi is more often referenced as a place than an operation, Hailstone appears as “the Truk raid,” Philippine Sea as “the Marianas Turkey Shoot.” Search-based instruments require a controlled vocabulary of aliases for each entry, and the alias list has to be built by hand and documented, because an undercounted alias produces a false neglect finding.


4. The Instruments

Seven instruments are proposed. Each measures something real and each fails in a specific way.

4.1 Monograph and Trade Publication Counts

What it measures. Sustained scholarly and commercial-press engagement. Counted as: books in WorldCat with the operation as subject or in title, filtered to English-language, with separate counts for academic press, trade press, and self-published or print-on-demand.

Why it matters. Books are the most durable form of attention and the one most likely to generate downstream output in other media. A battle with three trade books will get documentaries; a battle with none generally will not.

How it fails. It undercounts operations treated well inside campaign-level or theater-level books without receiving their own volume — which is most of the Central Pacific. Morison’s volumes cover everything; a subject-heading count registers that as coverage of “World War II naval operations,” not of Eniwetok. The correction is a supplementary measure: chapter-level and page-count treatment within the ten most-cited general works on the Pacific War, hand-coded. This is labor-intensive and worth it, because it is the single best discriminator between “not written about” and “written about only in aggregate.”

It also fails on recency: the self-published military history market has expanded enormously since roughly 2010, and raw counts will overweight operations that happen to have attracted a print-on-demand author. Segmenting by press type handles this.

4.2 Documentary and Film Depiction

What it measures. Access to the largest general audience, and the strongest single amplifier in the set.

Counted as. Feature films with the operation as principal setting; documentary episodes with the operation as principal subject, segmented by broadcast network production, streaming original, and independent; total runtime rather than title count, since a ninety-minute treatment and a four-minute segment are not equivalent.

Why it matters. This instrument has the most extreme distribution of any in the set and is where the amplification effect is most visible. The Pacific (2010) is very likely responsible for a large share of Peleliu’s present standing relative to Angaur, Biak, or Tinian, and the pre- and post-2010 comparison is a natural experiment the suite should exploit directly.

How it fails. It is dominated by adaptation availability, which is dominated by memoir availability, which means this instrument is partly measuring the same thing as instrument 4.5. The suite must not treat correlated instruments as independent confirmation. It also fails on the footage problem: operations filmed by combat cameramen whose material survived in usable form are depictable at low cost, and this advantages Tarawa specifically and enormously.

4.3 Video Platform Volume

What it measures. Contemporary popular engagement, and the most responsive instrument to recent shifts.

Counted as. Video count and aggregate view count by operation, segmented by channel tier (large general-history channels, mid-tier specialist naval and military channels, small and amateur), with a date filter permitting time-series analysis.

Why it matters. This is where the neglect is most visible and most current, and it is the instrument closest to the observation that prompted the suite. It is also the only instrument with enough volume for meaningful time-series work.

How it fails. Badly, in several ways, and the paper should be candid about it. View counts are driven by channel subscriber base rather than by subject, so a mediocre video on a popular channel outperforms a good one on a small channel by orders of magnitude. Platform recommendation systems create feedback: a subject that performs well gets recommended, gets more views, and attracts more creators, so measured attention partly measures prior measured attention. Titles are optimized for search rather than accuracy, generating alias noise. And the platform’s own metrics are not fully accessible, so counts are approximations from search results rather than true totals.

The correct use of this instrument is comparative and directional, not absolute. It is good for showing that Kwajalein sits two orders of magnitude below Midway. It is not good for claiming a precise ratio.

4.4 Museum and Exhibit Presence

What it measures. Institutional commitment, which is expensive and slow-moving and therefore a good indicator of settled rather than faddish attention.

Counted as. Presence and approximate floor area at a fixed set of institutions — the National WWII Museum, the National Museum of the Marine Corps, the National Museum of the United States Navy, the National Naval Aviation Museum, the National Museum of the United States Army, and the Pacific Aviation Museum — plus dedicated memorial sites and their visitation figures where published.

Why it matters. Museum space is the attention measure with the most direct funding consequence and the longest institutional memory.

How it fails. Sponsorship and service affiliation drive it heavily. A Marine Corps museum will treat Marine operations at length; the Army’s Central Pacific role will be underrepresented in the aggregate for structural reasons that have nothing to do with the phenomenon under study. This is a real problem and the correction is to report by-service normalized figures alongside raw ones. Geography also intrudes: mainland institutions are measured, Pacific-sited memorials are not comparably instrumented, and Paper 12 exists partly because of this gap.

4.5 Memoir and Testimony Supply

What it measures. The input to several other instruments, and — Paper 7 argues — a strong causal driver rather than merely a correlate.

Counted as. Published first-person accounts by operation; oral history holdings by operation in major collections, particularly the Library of Congress Veterans History Project and the National WWII Museum’s collection; unit histories.

Why it matters. This is the instrument the suite most needs to be independent, because Paper 7’s central claim is that testimony supply predicts attention. If it is measured well, the correlation with instruments 4.1 through 4.3 becomes evidence rather than tautology.

How it fails. Survival rates confound it. Operations with high casualty rates produce fewer surviving memoirists, and operations that annihilated the enemy garrison produce no enemy testimony at all. So memoir supply is partly a function of outcome, which is one of the variables the suite wants to hold separate. This has to be modeled explicitly rather than assumed away — the paper should report memoir count per thousand participants surviving, not raw count.

4.6 Preservation, Designation, and Wreck Discovery

What it measures. Physical and institutional anchoring of memory to place.

Counted as. National Park Service units and National Historic Landmark designations by operation; American Battle Monuments Commission sites; located and imaged shipwrecks associated with each fleet action, with discovery dates.

Why it matters. The wreck-discovery variable is the cleanest natural experiment available in the entire suite. Discoveries occur at datable moments, are exogenous to the historical significance of the battle, and are followed by measurable attention spikes. The Petrel expeditions of 2015–2019 supply a set of these events, and comparing pre- and post-discovery attention for the same operation controls for nearly everything the suite otherwise struggles to control for.

How it fails. Coverage is uneven — deep-water sinkings in the open ocean remain unlocated, which is not random with respect to battle type, since fleet actions fought far from land are systematically harder to survey than those fought in restricted waters. Land preservation is confounded by sovereignty: American designations attach easily to Guam and unevenly elsewhere in Micronesia.

4.7 Curricular and Reference Presence

What it measures. Baseline transmission to people not seeking the subject out.

Counted as. Presence in a sample of university survey course syllabi and secondary textbooks; article length, citation count, page-view figures, and edit volume for corresponding Wikipedia articles.

Why it matters. Reference presence is the floor of attention — what someone encounters without looking. The Wikipedia metrics in particular give a high-resolution, freely available, time-series-capable proxy for general interest, and page-view data can be cross-checked against the video instrument to see whether the two move together.

How it fails. Wikipedia article length reflects editor interest, which is not a random sample of general interest and skews toward subjects with active hobbyist communities — naval and armor subjects especially. Syllabus sampling is small-n and hard to make representative. Neither instrument distinguishes engagement from encounter.


5. Composition, Weighting, and What Not to Do

The instruments should be reported separately as the primary presentation. A composite index is useful for ranking and for the correlation work in later papers, but the disaggregated profile is more informative — an operation high on memoir supply and low on video volume is a different case from one low on both, and a composite hides that.

Where a composite is used, the proposal is a rank-normalized mean across the seven instruments, with instruments 4.2 and 4.5 down-weighted jointly because of their known dependence. Weights should be reported and the analysis re-run under alternative weightings to show that the headline findings are not artifacts of a weighting choice. If the ranking of Kwajalein relative to Tarawa flips under a plausible reweighting, the suite needs to know that before Paper 4 builds an argument on it.

Three things the index should not be asked to do.

It should not be used to argue that any operation is over-attended. The suite’s claim is about disproportion in aggregate, not about particular battles receiving too much. Midway’s attention level is not a problem to be corrected.

It should not be treated as a quality measure. Volume of output says nothing about the merit of the output, and the Central Pacific is in fact well served by a small number of excellent works. That is the distinction Paper 3 formalizes and it should be flagged here.

It should not be extended to non-American memory without rebuilding the instruments. Every instrument above measures an English-language, largely American attention environment. Japanese memory of these operations, Australian memory of the Southwest Pacific, and Micronesian memory of the islands themselves are all differently structured, and applying these instruments to them would produce findings about instrument coverage rather than about attention.


6. The Baseline Question, Deferred

An index of attention is only half the argument. Disproportion requires a denominator — attention relative to what. Candidate denominators include casualties, forces committed, tonnage landed, duration, strategic yield as assessed by contemporary planners, and strategic yield as assessed retrospectively. Each produces a different picture, and the choice among them is contestable enough to require its own treatment.

That is Paper 3’s work. This paper’s output is the numerator alone: a defensible, documented, bias-labeled measure of how much attention each operation currently receives.


7. Expected Findings and Falsification Conditions

Stating predictions before measurement is the discipline that separates this from advocacy. Four predictions, with the conditions that would refute the suite’s thesis.

Prediction 1. Within the Central Pacific Advance, attention will be concentrated on Tarawa, Saipan, and Peleliu across all seven instruments, with the remaining thirteen entries clustering more than an order of magnitude lower on at least five instruments. Refuted if attention within the campaign is smoothly distributed, or if the concentration appears on only two or three instruments — which would suggest a media-specific artifact rather than a general pattern.

Prediction 2. The three high-attention operations will share the features Paper 8 identifies — casualty shock with surviving footage, moral controversy, adaptation-amplified memoir — while the low-attention operations will lack all three regardless of scale. Refuted if a high-attention operation lacks all three, or if a low-attention operation possesses two or more.

Prediction 3. Wreck-discovery events will be followed by measurable attention increases on instruments 4.3 and 4.7 within twenty-four months, and Central Pacific fleet actions with unlocated wrecks will show no comparable inflection. Refuted if discovery produces no detectable change, which would substantially weaken Paper 6’s place-anchoring mechanism.

Prediction 4. Memoir supply, normalized for survivors, will correlate strongly with instruments 4.1 through 4.3 and weakly with 4.4 and 4.6. Refuted if the correlation is weak across the board, which would require rewriting Paper 7.

The honest position is that Prediction 3 is the one most likely to fail, because discovery events cluster with other publicity and the twenty-four-month window may be too generous to isolate an effect. If it fails, Paper 6 loses its strongest quantitative support and has to rest on comparative and qualitative argument. The suite should be prepared for that rather than reaching for a weaker version of the same claim.


8. Feasibility

Not all seven instruments are equally practical for a single researcher without institutional support. Instruments 4.1 (with the hand-coded chapter supplement), 4.3, 4.6, and the Wikipedia component of 4.7 are achievable. Instrument 4.4 requires either site visits or cooperative institutions and may have to be reduced to published floor plans and exhibit inventories. Instrument 4.5 requires substantial archival query time. The syllabus component of 4.7 is the weakest link and could be dropped without much loss.

A reduced four-instrument version — publication, video, wreck and preservation, and reference presence — would support Predictions 1 and 3 and partially support 2. It would not support Prediction 4, and Paper 7 would then have to be written as an argument from cases rather than from correlation. That is an acceptable fallback and should be planned for rather than discovered halfway through.


Notes

  1. The formulation that impressions of neglect are shaped by the observer’s own reading owes something to the general problem of availability in judgment under uncertainty; the classic statement is Tversky and Kahneman (1973). No claim of methodological sophistication is intended here beyond the observation that a specialist’s sense of what is missing is not a random sample.
  2. The Central Pacific boundary used throughout is Paper 1’s — Galvanic through Stalemate II — with Iwo Jima and Okinawa excluded. Readers who prefer the inclusive boundary should note that adding those two operations would raise the campaign’s aggregate attention substantially while leaving the internal distribution, which is this paper’s actual subject, largely unchanged.
  3. On the alias problem: “Operation Hailstone” and “the Truk raid” refer to the same February 1944 carrier operation, and a search-based instrument that captures only the code name will undercount it severely, since popular usage overwhelmingly prefers the descriptive form. Similar care is needed for Philippine Sea, where “Great Marianas Turkey Shoot” is the more common popular label and attaches to only part of the engagement.
  4. The pre- and post-2010 comparison around The Pacific is proposed as a natural experiment, but the design is imperfect: the miniseries covers Guadalcanal, Cape Gloucester, Peleliu, and Okinawa, so it does not offer a clean treated-versus-untreated split within the Central Pacific. Peleliu is effectively the only Central Pacific operation treated, which makes it a single-case observation rather than a comparison. It remains worth doing for that single case.
  5. Sledge’s memoir (1981) is the clearest instance in the whole suite of a single testimony determining an operation’s standing. The suite should resist any implication that this is a defect. The book is excellent and its influence is deserved. The point is only that its existence, rather than Peleliu’s operational weight, accounts for the attention differential between Peleliu and Angaur — two operations conducted concurrently by adjacent formations against the same garrison command.
  6. Y’Blood (1981) and Tillman (2005) are cited here as instances of the pattern noted in section 5: the Central Pacific is not badly served by its literature, only thinly served by volume. A count-based instrument registers thin coverage and cannot register good coverage, which is exactly why Paper 3’s distinction between quantity and quality of attention is necessary rather than pedantic.
  7. On the wreck instrument: the Petrel discoveries are used as the primary event set because they are well documented and closely spaced. Researchers should note that the vessel’s operations were suspended following its owner’s death in 2018, and that subsequent survey work has proceeded under different sponsorship, which introduces a discontinuity in discovery rate that must not be mistaken for a change in attention dynamics.
  8. Morison’s volumes seven and eight are the backbone of the chapter-level supplement to instrument 4.1 and should be coded first, since their allocation of pages across operations is itself an early attention distribution and may be doing more causal work than the suite currently credits. A useful side finding would be whether present-day attention correlates with Morison’s page allocation more strongly than with any operational variable.
  9. The suite’s use of “attention” as a countable outcome distinct from “memory” as a cultural formation follows the general shape of the distinction in memory studies without adopting any particular theoretical apparatus. Winter (2006) and Rosenzweig and Thelen (1998) are the useful reference points for the gap between institutional and popular historical consciousness that Paper 12 develops.
  10. Instrument 4.4’s service-affiliation bias deserves more than the paragraph given it. The Army divisions that took Kwajalein Island, Eniwetok Island, Angaur, and much of Saipan are represented in the museum landscape by institutions with far less Pacific-specific floor space than the Marine Corps museum devotes to the same campaign. Any raw floor-area measure will therefore encode a service-memory asymmetry that Paper 8 treats as a finding in its own right, and the two papers should cross-reference rather than duplicate.

References

Alexander, J. H. (1995). Utmost savagery: The three days of Tarawa. Naval Institute Press.

American Battle Monuments Commission. (n.d.). Cemeteries and memorials. https://www.abmc.gov

Blair, C. (1975). Silent victory: The U.S. submarine war against Japan. Lippincott.

Bureau of Yards and Docks. (1947). Building the Navy’s bases in World War II (Vols. 1–2). U.S. Government Printing Office.

Camp, D. (2009). Leatherneck legends: Conversations with the Marine Corps’ old breed. Zenith Press.

Carter, W. R. (1953). Beans, bullets, and black oil: The story of fleet logistics afloat in the Pacific during World War II. U.S. Government Printing Office.

Costello, J. (1981). The Pacific war, 1941–1945. Rawson, Wade.

Crowl, P. A. (1960). Campaign in the Marianas. Office of the Chief of Military History, Department of the Army.

Crowl, P. A., & Love, E. G. (1955). Seizure of the Gilberts and Marshalls. Office of the Chief of Military History, Department of the Army.

Dyer, G. C. (1972). The amphibians came to conquer: The story of Admiral Richmond Kelly Turner (Vols. 1–2). U.S. Government Printing Office.

Gailey, H. A. (1983). Peleliu, 1944. Nautical & Aviation Publishing.

Hallas, J. H. (1994). The devil’s anvil: The assault on Peleliu. Praeger.

Hoffman, C. W. (1950). The seizure of Tinian. Historical Division, U.S. Marine Corps.

Hornfischer, J. D. (2016). The fleet at flood tide: America at total war in the Pacific, 1944–1945. Bantam Books.

Isely, J. A., & Crowl, P. A. (1951). The U.S. Marines and amphibious war: Its theory and its practice in the Pacific. Princeton University Press.

Library of Congress. (n.d.). Veterans History Project collection. American Folklife Center. https://www.loc.gov/vets/

Miller, E. S. (1991). War Plan Orange: The U.S. strategy to defeat Japan, 1897–1945. Naval Institute Press.

Morison, S. E. (1951). Aleutians, Gilberts and Marshalls, June 1942–April 1944 (History of United States naval operations in World War II, Vol. 7). Little, Brown.

Morison, S. E. (1953). New Guinea and the Marianas, March 1944–August 1944 (History of United States naval operations in World War II, Vol. 8). Little, Brown.

National WWII Museum. (n.d.). Digital collections. https://www.nationalww2museum.org

O’Brien, P. P. (2015). How the war was won: Air-sea power and Allied victory in World War II. Cambridge University Press.

Poyer, L., Falgout, S., & Carucci, L. M. (2001). The typhoon of war: Micronesian experiences of the Pacific War. University of Hawai’i Press.

Prefer, N. N. (2012). The Battle for Tinian: Vital stepping stone in America’s war against Japan. Casemate.

Rosenzweig, R., & Thelen, D. (1998). The presence of the past: Popular uses of history in American life. Columbia University Press.

Rottman, G. L. (2004). The Marshall Islands 1944: Operation Flintlock, the capture of Kwajalein and Eniwetok. Osprey.

Sherrod, R. (1944). Tarawa: The story of a battle. Duell, Sloan and Pearce.

Sledge, E. B. (1981). With the old breed: At Peleliu and Okinawa. Presidio Press.

Smith, H. M., & Finch, P. (1949). Coral and brass. Charles Scribner’s Sons.

Spector, R. H. (1985). Eagle against the sun: The American war with Japan. Free Press.

Tillman, B. (2005). Clash of the carriers: The true story of the Marianas Turkey Shoot of World War II. NAL Caliber.

Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5(2), 207–232. https://doi.org/10.1016/0010-0285(73)90033-9

Winter, J. (2006). Remembering war: The Great War between memory and history in the twentieth century. Yale University Press.

Y’Blood, W. T. (1981). Red sun setting: The Battle of the Philippine Sea. Naval Institute Press.

Unknown's avatar

About nathanalbright

I'm a person with diverse interests who loves to read. If you want to know something about me, just ask.
This entry was posted in American History, History, Military History and tagged , , , , , . Bookmark the permalink.

Leave a Reply