Why General Search Misses Asian Primary Literature

The blind spot

General academic search was built for one world: English-language journals indexed in commercial databases. Google Scholar, Web of Science and Scopus do not index CNKI, Wanfang or VIP, and they do not index Garuda, Neliti or SINTA either. They do not index the institutional repositories of Vietnamese and Thai universities, nor the Indonesian OJS journals, where a large share of the local empirical work is published.

That is the uncomfortable part: the literature you most need to cite for a study about a domestic market is precisely the literature that general search cannot see. A supervisor in the target country expects those references. One search on Scholar will simply not produce them.

These are not sources you can reach by being clever with a query string. Indexing is the publisher's decision and the database's business; a journal published in Hanoi or Bangkok is often invisible to a search that runs entirely in English, and a paywall or a national licence model can put it out of reach even when you know the exact title. The literature is there. The index is not.

Three tiers, not one search box

WriterOne treats sourcing as a routing problem and splits it into three layers:

• Target-country native-language sources: the elite universities, core journals and authoritative institutions of the country you are studying, searched in the local language.
• The English literature track: leading US and European universities plus SSCI core journals, which supply the theory and the cross-country comparisons.
• Domestic journal databases: the national indexes that general search does not cover.

The three layers are searched separately, and every hit keeps a record of the layer it came from. A claim about Indonesian consumer behaviour and a claim about e-commerce theory do not deserve the same kind of source, and they should not be written as though they did.

A literature review is not a reading list. Its critical review only means something when it rests on the sources the market actually produces.

Reach: what each country really opens up

Below is the accessible volume for the 14 core markets (the whitelist actually covers 142 target countries), in the fixed order universities / journals / institutions:

• China: 227 / 16 / 20 — CNKI, Wanfang, VIP.
• Indonesia: 138 / 14 / 20 — Garuda, Neliti, SINTA.
• Vietnam: 128 / 14 / 13 — the VJOL national database.
• Thailand: 127 / 14 / 13 — the ThaiJO national database.
• Malaysia: 142 / 12 / 14 — statistics office, central bank.
• Philippines: 141 / 11 / 13 — central bank, statistics office.
• Singapore: 148 / 12 / 13 — statistics office, monetary authority.
• Japan: 146 / 11 / 13 — Bank of Japan, Ministry of Internal Affairs statistics.
• South Korea: 142 / 11 / 11 — global Tier-1 institutions.
• United States: 130 / 15 / 12 — Federal Reserve, federal statistics.
• United Kingdom: 135 / 11 / 12 — the national statistics office.
• Germany: 142 / 11 / 12 — the national statistics office.
• France: 141 / 14 / 12 — the central bank, the national library.
• Australia: 148 / 12 / 13 — the central bank, statistics office.
• The second track is shared: leading US and European universities plus SSCI core journals.

Read those fourteen numbers side by side and the differences show the shape of the problem exactly: China's reach is the largest not because it has more English literature, but because it also has the largest domestic journal ecosystem; Singapore and Australia lead on university count because their whitelists each admit 30 verified universities; Vietnam and Thailand land close together because the two countries have similarly sized universities and similarly sized domestic journal databases. These numbers are not a measure of "enough" either — they measure not how much literature you need, but how many doors you will meet once you open the first one.

Principles: gate, archive, funnel

First, the whitelist gate. Every source carries its tier and the reason it was admitted. No entry stops at "high-quality source"; the reason is written down, because a reason you cannot check is not a reason.

Second, the archive. Every search string is stored with the stage it belongs to, the source type, the language, the raw hit count and the number kept. Every excluded record is stored with its exclusion reason: a landing page instead of an article, a full text below the length threshold, a duplicate title, outside the whitelist. When your supervisor asks why a paper is missing from your review, the answer is one line in a file, not a recollection.

Third, the funnel. The PRISMA screening funnel records each step and the exclusion reason at that step, one by one. How many remain after duplicates are removed, how many drop out at the title screen, how many full texts cannot be obtained, how many fail the methodology test — every number traces back to the records behind it.

Conclusion: the first question is "where do I look"

Most people begin a literature review with a search string. That is the second question. The first question is where the literature of your market actually lives, and the answer is a whitelist you maintain yourself, not a set of defaults you inherited.

Search strings do not close a blind spot. A source map does. And the map is not something you inherit either; it is a list you keep current, country by country, as the institutions and indexes of each market change.