A shared year does not create a shared reader
Three official survey instruments ask about book reading over the last twelve months: Mexico's MOLEC 2025, Switzerland's culture survey 2024 and the United States Survey of Public Participation in the Arts 2022. Their agreement on recall is the starting result of this audit, not evidence of identical measurement. A respondent can encounter different eligibility filters, purpose restrictions, confirmation questions and answer formats despite remembering the same year of activity. The comparison examines those observable differences without calculating national reader rates or claiming that one country reads more than another.
The unit of analysis is a particular instrument version and its related documentation. It is not a country, a language community or a representative selection of the world's reading surveys. The three versions were chosen purposively because original questionnaires and provenance were already available for inspection. Their different years are not treated as a time series. This bounded design can establish that particular item definitions differ, or that a proposed mapping is unsupported. It cannot establish how common those differences are across all surveys, how many people they affect, or which survey provides a better account of reading.
Four kinds of evidence must stay separate
The coding distinguishes explicit questionnaire wording, clarification in matching technical documentation, analytical inference and unresolved information. An explicit exclusion in a main question is stronger evidence about that question than the mere presence of a separate genre item. A dictionary clarification can add useful information, but it must be identified as a dictionary statement rather than silently inserted into the printed interview question. Likewise, a plausible consequence of a count structure is not proof of every admissible response in its programmed implementation.
These distinctions explain several otherwise tempting shortcuts. The Swiss questionnaire requests correction when an ebook subset exceeds a known book total; this establishes a correction prompt, not an independently verified hard software rejection. Its all-respondent count structure suggests a zero response, but the inspected passage does not print the full permitted numeric range. The US book-count question follows a yes/no screen, but the exact programmed routing was not established in the inspected passage. Missing documentation therefore remains missing documentation. It is not coded as a negative answer, a prohibited activity or an absent survey feature.
Population eligibility comes before the book question
Mexico's sample-design documentation defines the target population as people aged twelve or older who usually reside in occupied private dwellings within national territory. The questionnaire then introduces an explicit literacy filter for the selected person: the reading module continues for someone who can read and write, and otherwise closes. This is an important distinction between the wider survey population and the people reaching the reading questions. Describing the module simply as a survey of all adults would obscure both its younger eligible ages and its documented literacy condition.
The Swiss methodological profile identifies permanent residents aged fifteen or older in private households. Its book question is marked for all respondents within the survey, but that local routing label is not evidence that every possible population eligibility rule has disappeared. The US documentation identifies civilian, noninstitutional household residents, with a selected member aged eighteen or older and random assignment to a survey core. The book items belong to Core 1. Restricting an eventual data analysis to people eighteen or older would address one age difference, but would not align residency concepts, household coverage, literacy filtering, core assignment or the meaning of a book. No such harmonized population estimate is attempted here.
Three routes to reporting books
MOLEC separates self-perceived reading habit from material-specific reporting. Its first question asks about habitual reading, but that answer is not the gate defining whether the book section is administered. The book screen asks about the previous twelve months. An initial negative answer leads to a verification question that explicitly mentions an unfinished book and gives manuals or subject and professional books as examples. A positive initial screen or a positive verification answer can open the book section. The route therefore offers a documented second opportunity to recognize activity that did not produce an initial affirmative response.
The Swiss main book item instead directly requests a number of books read for private purposes, with ebooks included and comics excluded. Estimation is permitted. The US main item begins with a yes/no report of books read, excluding those required for work or school and including print and electronic books. A count question follows in the instrument. These are different observables: identifying with a habit, recognizing any eligible book and supplying a numerical count are not interchangeable tasks. Their distinction is visible in the source documents, but its effect on respondents cannot be quantified from wording alone. This audit does not infer that confirmation necessarily raises an estimate by any particular amount.
Purpose is not one universal leisure category
Mexico explicitly introduces the material questions as covering work, study or entertainment. Its later book-purpose item distinguishes pleasure from necessity, but that subsequent classification does not retrospectively restrict the opening book screen to leisure reading. Manuals, guides and recipe books appear in its book-type questions, alongside subject, professional, school and university books. The combined self-help, personal-improvement and religious category is retained as a combined category; the instrument does not license separating its components into independent counts using that answer alone.
The US main book question excludes books required for work or school. Required is consequential: the wording does not exclude every book with a professional subject or every voluntarily chosen book relevant to employment. Switzerland asks about private purposes. It would be convenient to assign both instruments a common leisure-only label, but that translation would conceal an unresolved boundary between private reading and non-required professional reading. The clear Mexico/US difference on required books does not establish an equally clear three-way classification for every mixed-purpose activity. A book chosen for enjoyment that also helps someone's work illustrates why a subject label cannot substitute for the stated purpose restriction.
Unfinished books and the limits of generic verbs
Mexico's verification question explicitly allows a book that was not finished. That provision matters because it appears on a specific response path: the respondent has first denied reading books and is then prompted to reconsider. The subsequent annual count question does not independently restate a completion convention. The evidence therefore supports inclusion of unfinished books in the confirmation route, not a fully specified rule about how every partly read volume is converted into the numerical count. Neither an assumed full-book equivalent nor a page-based threshold is introduced in this analysis.
The inspected Swiss and US main questions do not establish a comparable explicit completion rule. Their ordinary verbs for reading are not sufficient grounds to code partial reading as either included or excluded. This uncertainty limits comparisons involving abandoned books, reference consultation and several books read in small portions. It also prevents the attractive but unsupported conclusion that Mexico's eligible set necessarily contains both other sets in every respect. Set inclusion requires known boundaries across all relevant dimensions, not one documented expansion on a single dimension. A source-backed partial statement is more informative than a complete-looking table built from invented exclusions.
Comics, religious books and the difference between separation and exclusion
Switzerland explicitly excludes comics from its main book total and measures them separately. The separate comic item includes graphic novels and manga. That is a documented relationship between the two questions, not merely an inference from their positions. Mexico also has a separate comic material screen, and the US instrument separately asks about graphic novels or books in comic-strip form. However, the presence of those separate items does not establish that respondents must exclude the same material from each generic book screen. The generic-book overlap remains unresolved for Mexico and the United States.
The US documentation illustrates why inspecting only the printed questionnaire can also be insufficient. The matching 2022 dictionary entry for the main book measure clarifies inclusion of religious texts, although that clarification is not printed in the displayed main question. A separate religious-reading item also appears in the instrument and retains the required-work-or-school exclusion. The audit identifies both pieces of evidence without combining them into a quotation the respondent never saw. Swiss religious-book scope remains unknown in the inspected book passage. These differences concern documentation and item boundaries, not the cultural value of comics, religion or any other genre.
Ebook inclusion does not supply an ebook denominator
All three selected instruments explicitly accommodate electronic books in their book questions. That agreement is useful, but the follow-up information differs. Mexico records digital and printed format indicators for reported books. Such indicators reveal whether a format was reported, not how the total number of books divides between formats. A person reading one digital book and another reading many digital books can supply the same affirmative format indicator. It would therefore be incorrect to convert those indicators into counts or into the proportion of all books read electronically.
Switzerland asks how many of the privately read books were ebooks. The subset clarification describes books read on tablets or eReaders, and its correction prompt compares the subset with the known total. These details must remain attached to any discussion of the resulting count. The US main question includes printed and electronic books but does not provide a matching ebook-count item within the audited main-question pair. Consequently, agreement on inclusion does not produce equivalent format denominators. The audit does not extend the Swiss device clarification automatically to every device, or infer a US ebook share from the inclusive wording alone.
Audiobook listening is not silently resolved
The US instrument has a separate audiobook-listening question covering recordings, broadcasts or web streaming. That is evidence that listening is measured, but it does not by itself resolve every possible overlap with the generic book-reading answer. Mexico's and Switzerland's selected main-book passages do not provide an explicit audiobook inclusion or exclusion rule. The appropriate entries are therefore unknown for the main measure, with the separate US listening item identified. Absence of the word audiobook from a selected passage is not evidence that listening was prohibited or impossible to report.
This distinction matters for anyone trying to create a single cross-survey audience by joining reading and listening categories. Adding two percentages would require knowing whether their respondent sets overlap, whether their universes and recall periods match, and whether their definitions cover the intended activity. A separate question supplies neither automatic disjointness nor permission to reinterpret the main book answer. No audiobook-adjusted rate is calculated here. The more modest result is a documented limit: the audited evidence cannot assign a universal book-reader status to a hypothetical person whose only activity was listening.
The same US version changes units between collection and release
The 2022 US questionnaire instructs entry of an exact book count from one through nine hundred, with nine hundred and one used for values above that limit. The matching released-variable dictionary, however, groups counts above fifteen into intervals. This is a difference between collection instructions and released analytical values within the same survey version, not a change inferred by comparing different years. The distinction would be missed by an audit that looked only at either the interview document or the public data dictionary.
Information is lost when several entered counts map to one released category. A released interval can support membership in that interval but cannot recover the original exact count without additional information. Assigning its midpoint would be an analytical assumption, not reproduction of the respondent's reported number. The top category also does not identify a finite mean by itself. No midpoint substitution, total-book estimate or intensity ranking is performed. Mexico's positive numerical count and Switzerland's requested numerical total are recorded as questionnaire units, while their public-data transformations are outside this bounded comparison. Unequal investigation of release formats must not become a claim that those other surveys have no release transformations.
Synthetic histories test logic, not prevalence
The illustrative histories accompanying the evidence table are invented logical probes, not survey records or descriptions of actual respondents. Assume first that a person satisfies each instrument's respective eligibility conditions and completed only one printed textbook required for work during the preceding twelve months. Mexico's broad purpose introduction and textbook category support its eligibility as book activity, whereas the US main question explicitly excludes the required book. Swiss private-purpose interpretation remains indeterminate in this probe. This establishes a supported Mexico/US definitional distinction without assigning a national incidence or a probability of answering yes.
A second history starts with an initial denial followed by recognition of an unfinished professional manual. It demonstrates Mexico's documented confirmation path, while leaving other instruments' completion and purpose boundaries unresolved. A third uses only a private comic: Switzerland explicitly excludes it from the main total, but Mexico/US generic-book overlap stays unknown. A fourth specifies two completed private ebooks read on tablets or eReaders within twelve months. Swiss total and subset can both reflect two; Mexico supplies a digital indicator rather than an ebook count; the US count is two only if the person is routed to the count question, whose exact programmed route remains unverified. An audio-only history remains unresolved for generic book status. None of these probes estimates response behavior.
What can and cannot be carried forward
The strongest common feature in the audited main items is twelve-month book recall. It is not evidence of identical calendar windows because interviews occur at different times, and it does not erase differences in eligible activity. The resulting comparison is best understood as a map of conditions that an analyst would need to resolve before pooling data. It identifies supported agreements, explicit divergences and unknown boundaries separately. It deliberately stops short of a universal conversion factor, a hierarchy of survey quality or a synthetic comparable reading percentage.
The compact downloadable coding table preserves original item locators and paraphrased evidence; the reproduction script regenerates the display comparison from explicit inputs without downloading respondent records. Original questionnaires, extracted pages and screenshots are not redistributed. Questionnaire-specific redistribution rights for the Swiss and US originals were not established, and no blanket licence is inferred from an agency's identity. The original analysis is distinguishable from those source documents. Future expansion would require a transparently revised sample, version-specific source checking and the same treatment of uncertainty. Adding more countries without those safeguards would enlarge a table without necessarily improving its comparability.
Same recall, different book definitions
Four coded fields across three purposively selected instrument versions; no reader rates.
Full chart and source
Scroll horizontally to inspect the chart. ↔
Chart and values
Scroll horizontally to see all columns.
| Field | MX_MOLEC_2025 | CH_KULTUR_2024 | US_SPPA_2022 |
|---|---|---|---|
| recall | Last 12 months for books Q2a/Q3a/Q4 | Last 12 months Q52000 | Last 12 months C1Q14A/B |
| purpose | Work/study/entertainment explicitly in section 1 introduction; Q8 pleasure/necessity is subsequent classification | Private purposes; do not silently translate to every non-required professional reading | Explicitly excludes books required for work or school; voluntary work-related reading not explicitly excluded |
| completion | Q3a explicitly permits unfinished books in negative-screen verification; Q4 does not restate completion | Unknown; completion not specified in Q52000 | Unknown; read-any not a documented partial-book rule |
| audio | Unknown in selected book items; no inference from absence | Unknown in selected book items | Separate C1Q17A listening item; not interchangeable with Q14A reading; explicit main-item audio inclusion/exclusion unknown |
Exact paraphrased codes from the supplied input. Unknown is not exclusion; this is not an instrument-quality score.
Data & downloads
Scroll horizontally to see all columns.
| Field | MX_MOLEC_2025 | CH_KULTUR_2024 | US_SPPA_2022 |
|---|---|---|---|
| recall | Last 12 months for books Q2a/Q3a/Q4 | Last 12 months Q52000 | Last 12 months C1Q14A/B |
| purpose | Work/study/entertainment explicitly in section 1 introduction; Q8 pleasure/necessity is subsequent classification | Private purposes; do not silently translate to every non-required professional reading | Explicitly excludes books required for work or school; voluntary work-related reading not explicitly excluded |
| completion | Q3a explicitly permits unfinished books in negative-screen verification; Q4 does not restate completion | Unknown; completion not specified in Q52000 | Unknown; read-any not a documented partial-book rule |
| audio | Unknown in selected book items; no inference from absence | Unknown in selected book items | Separate C1Q17A listening item; not interchangeable with Q14A reading; explicit main-item audio inclusion/exclusion unknown |
Exact paraphrased codes from the supplied input. Unknown is not exclusion; this is not an instrument-quality score.
Scroll horizontally to see all columns.
| Field | MX_MOLEC_2025 | CH_KULTUR_2024 | US_SPPA_2022 |
|---|---|---|---|
| language_mode | Spanish; interviewer questionnaire | German original; CAWI and CATI wording inspected | English; Census instrument; interviewer/person-or-name wording |
| universe | Age 12+ usual residents of occupied private dwellings in national territory; design section 1.2 | Age 15+ permanent resident population in private households; CHM p1 | Civilian noninstitutional household population; selected member 18+; randomly assigned Core 1; July 16-25, 2022 |
| literacy | Explicit Filter 3 requires can read and write; no terminates interview | Unknown: Q52000 ALL is within survey and does not establish absence of other eligibility restrictions | Unknown; do not infer a literacy restriction or its absence from book question |
| recall | Last 12 months for books Q2a/Q3a/Q4 | Last 12 months Q52000 | Last 12 months C1Q14A/B |
| completion | Q3a explicitly permits unfinished books in negative-screen verification; Q4 does not restate completion | Unknown; completion not specified in Q52000 | Unknown; read-any not a documented partial-book rule |
| purpose | Work/study/entertainment explicitly in section 1 introduction; Q8 pleasure/necessity is subsequent classification | Private purposes; do not silently translate to every non-required professional reading | Explicitly excludes books required for work or school; voluntary work-related reading not explicitly excluded |
| comics | Separate material screen Q2d/Q3d; main-book exclusion not explicit | Explicitly excluded from Q52000; separate Q52500 includes graphic novels/manga | Separate C1Q16E graphic/comic-strip books; main Q14A exclusion not explicit |
| manuals | Explicitly included in Q3a and Q9.1; subject/professional/textbooks Q9.2 | Unknown; may not infer exclusion from private-purpose requirement | Unknown; required work/school books excluded independent of genre |
| religious | Q9.3 combined self-help/personal improvement/religious category | Unknown | Same 2022 dictionary PEC1Q14A p7-6 clarification includes religious texts in main book measure; not printed in questionnaire Q14A; separate C1Q16D excludes work/school-required books |
| ebook | Included paper/digital Q3; formats separately Q5 | Explicitly included Q52000; numeric ebook subset Q52100; subset clarification describes reading on tablets/eReaders | Explicitly included Q14A |
| audio | Unknown in selected book items; no inference from absence | Unknown in selected book items | Separate C1Q17A listening item; not interchangeable with Q14A reading; explicit main-item audio inclusion/exclusion unknown |
| routing | Q1 habit is not a gate; Q2a yes OR negative Q2a followed by Q3a yes opens book section; Filter 3 first | Q52000 ALL; Q52100 only Q52000>=1; consistency check prompts correction when ebook subset exceeds known total; hard programmed rejection unverified; Q52500 ALL | Core 1 random assignment; yes/no Q14A precedes count; count conditional routing not printed in questionnaire passage and left unverified |
| answer_unit | Q4 positive numeric book count; Q6 ordinal frequency; Q7 numeric average minutes on a reading day | Numeric total count; zero inferred from ALL/count structure, explicit admissible range unverified; estimate permitted; -8 unknown and -9 no answer distinguished | Q14A yes/no; administered Q14B exact 1-900, 901 above 900; released PTC1Q14B binned above 15 |
| locators | MXQ printed/PDF pp10-11 Q1-10; MXD section 1.2 | CHQ PDF pp16-17 / printed 14-15 Q52000/Q52100/Q52101_pl01/Q52500; CHM p1 | USQ PDF 155 / instrument p8 Q14A/B; PDF 157-158 / pp10-11 Q16/Q17; dictionary section 7 pp6-7; overview attachment 3 |
| unknowns | Audio; whether comics also counted as books; exact unfinished-book convention for Q4 | Completion/audio/manuals/religious; equivalence of private purpose with non-required work/school | Completion; exact programmed Q14B routing; audio/comic scope of generic book screen; literacy |
Source-specific eligibility, taxonomy, routing and documentation. Dictionary clarification and inference are explicitly identified.
Evidence & method
How we calculated it
- Prospective structured coding protocol followed inventory keyword inspection; fixed purposive sample of three original versions, not a blinded preregistration.
- Five original questionnaire/design PDF hashes matched frozen provenance; main question pages visually inspected; matching US dictionary and overview separately identified.
- Code explicit questionnaire wording, dictionary clarification, analytical inference and unknown separately; silence never implies exclusion.
- Synthetic histories are illustrative logical probes, not real respondents, prevalence estimates or causal effects.
- Public script validates the supplied compact coding and derives the comparison with explicit input/output arguments; it does not independently recode source prose.
What the data cannot tell us
- No representative survey sample, pooled reader rate, country ranking or survey-quality ranking. All selected book recall periods are 12 months.
- Population, literacy, purpose and routing differences remain; restricting ages alone cannot establish equivalence.
- US exact programmed count routing, Swiss numeric range and hard rejection behavior, and some completion/audio/comic boundaries are unverified.
- Swiss ebook subset clarification names tablets/eReaders; broader device equivalence is not assumed.
- US religious-text inclusion comes from the matching dictionary, not printed Q14A. Required work/school reading is not all voluntary professional reading.
- Swiss/US questionnaire redistribution rights unverified; public bundle excludes whole questionnaires, source extracts/screenshots and microdata.
Original sources
- INEGI MOLEC 2025 questionnaire
INEGI MOLEC 2025 questionnaire: Spanish original 2025; printed/PDF 10–11, Filter 3 and Q1–10. Negative-screen confirmation explicitly mentions unfinished books; later count convention not restated.
Accessed October 30, 2026
- INEGI MOLEC 2025 sample design
INEGI MOLEC 2025 sample design: MOLEC 2025 sample design section 1.2: age 12+ usual residents of occupied private dwellings within national territory.
Accessed October 30, 2026
- BFS culture 2024 questionnaire
BFS culture 2024 questionnaire: German original 2024; PDF 16–17/printed 14–15. Q52000, Q52100, Q52101_pl01, Q52500; CAWI/CATI variants inspected. Correction prompt is not verified hard rejection; zero is structural inference.
Accessed October 30, 2026
- BFS culture methodological profile
BFS culture methodological profile: BFS culture methodological profile p1: age 15+ permanent residents in private households; mixed Internet/telephone mode in 2024.
Accessed October 30, 2026
- US Census July 2022 SPPA technical documentation
US Census July 2022 SPPA technical documentation: Matching 2022 questionnaire, overview and dictionary. Instrument printed 8/10/11, dictionary 7-6/7-7; religious-text clarification is dictionary evidence, not printed Q14A. Exact count routing remains unverified.
Accessed October 30, 2026
- INEGI terms of use
INEGI terms of use: Existing source pack terms inspected: attribution and nonmisleading reuse. No blanket questionnaire licence assigned to Swiss or US originals.
Accessed October 30, 2026
Cite & reuse this study
Alisson Bento. Same twelve months, different readers: auditing three book-reading instruments. Comparative analysis of INEGI MOLEC 2025, BFS culture 2024 and US Census/NEA SPPA 2022 instruments. Sources inspected October 30, 2026. https://alissonbento.com/research/book-reading-instrument-comparability-2022-2025/
Original analysis and concise English coding: Alisson Bento. Underlying instruments: INEGI MOLEC 2025; Swiss Federal Statistical Office culture 2024; US Census Bureau/NEA SPPA 2022. No official endorsement or official translation claim. INEGI terms https://www.inegi.org.mx/inegi/terminos.html support attributed, nonmisleading use of its information. Swiss/US questionnaire-specific redistribution rights were not established here; no blanket licence is inferred from government provenance or statistical-result reuse terms. Do not interpret this package as permission to redistribute questionnaire layouts, logos or third-party material. Only compact original paraphrases, source references and original analytical code are supplied. No full questionnaires, PDF extracts, screenshots or respondent records are included. No licence for the entire article is invented.
Corrections or editorial questions