An open door and professional help are different offers
For a reader planning a library visit, “open” answers only the first question. A second question is whether someone with a library-professional role will be available to provide service. Finland’s official statistics distinguish these offers rather than placing every opening hour in one undifferentiated total. This study uses that distinction to describe the composition of reported opening volume in 2024. Its central result is not a verdict about whether libraries were sufficiently staffed. It is an account of how much recorded opening volume belonged to each source-defined mode.
On the matched set of 277 complete reporting rows, professional service accounted for 43.6445% of annual hours. Personnel-present opening without service accounted for 4.7343%, while access without library staff accounted for 51.6212%. These percentages share one denominator and therefore describe a coherent three-part composition. They do not show that professional service declined, that self-service replaced it, or that any particular reader could obtain the same mix at a nearby branch. A single year’s reporting cannot establish those changes or individual experiences.
The original contribution here is the joint reading of that composition, the distribution across reporting rows, and the consequences of weighting by hours. The official definitions and descriptive statistics are existing source methods and prior art, not inventions of this article. Bringing the measures together makes a practical distinction visible: a system can report substantial opening volume without that entire volume being professionally served. Conversely, professional service can remain present in every complete reporting row even when another mode accounts for the largest share of pooled hours.
What the three opening modes actually mean
The first mode requires library-professional personnel to be present and serving. Both conditions matter. It is narrower than having any employee somewhere in the building, and it is not a count of staff members or their working hours. The recorded unit remains opening time under a specified service condition. A reader interested in help from library professionals should therefore look for this category rather than treating all personnel-present hours as equivalent. The statistics distinguish the availability of professional service, not the particular questions answered during those hours.
The second mode records personnel present without service. Under the official 2024 instructions, those personnel need not be library professionals: they may include a guard, caretaker, or café worker. Its 117,998.25 hours on complete support cannot therefore be described as additional professional-service hours. Nor should the category be interpreted as a uniform staffing arrangement across reporting rows. What connects the observations is the source’s reporting definition, not a demonstrated common occupation, task, or level of assistance available to visitors.
The third mode means access without library staff present, associated with card- or code-type entry. “Self-service access without library staff” is more accurate than “absolutely unattended.” The definition does not establish that nobody else was in the building, that no other activity occurred there, or that every library resource was usable. For readers, the three modes are best understood as distinct service conditions attached to opening volume. They are not a ladder of building quality, safety, collection value, or the usefulness of an individual visit.
The unit is summed service-hours, not a national timetable
The denominator requires as much attention as the percentages. The official opening totals aggregate service-hour volume across service points and include mobile-library service time at stops. They are not unique clock-hours during which some library somewhere was open. If opening periods overlap across locations, their hours still contribute to the sum. Consequently, the complete-support total of 2,492,395.25 hours describes an accumulated annual quantity. It cannot be placed directly on a calendar as a continuous sequence or converted into the opening schedule of a typical branch.
The 2024 instructions also explain why mobile libraries belong in this interpretation. Their reporting includes owner-or-base vehicle recording, municipality-specific stops and activity, and stops treated as locations. The resulting opening measure includes service time at those stops rather than representing only fixed-building doors. Other service-place opening hours are excluded by the source. This scope matters when comparing the figures with an ordinary reader’s understanding of “library opening hours,” which may refer only to the branch they usually visit.
The selected observations are 278 named municipal reporting rows, not 278 branches, independent networks, or separate employers. The official system covers mainland Finland, excluding Åland, and uses municipal or joint-library reporting. The selected rows should not be relabelled as all Finnish municipalities or as independent local systems. Keeping those distinctions intact prevents two misleading shortcuts: dividing an aggregate into an imagined branch average, and treating a reporting-row distribution as a direct description of residents’ opportunities to visit. Neither inference is supported by the supplied fields.
What the matched composition establishes
The main calculation uses the same 277 complete rows for every mode. Their annual volumes are 1,087,793.50 professional-service hours, 117,998.25 personnel-present nonservice hours, and 1,286,603.50 hours of access without library staff. Together they total 2,492,395.25 hours. Dividing each mode’s sum by that common total gives 43.6445%, 4.7343%, and 51.6212%, respectively. The tables round these shares to two decimal places; the more precise percentages express the same calculation rather than a separate estimate.
The largest component is therefore access without library staff, and it exceeds half of the matched opening volume. The narrow interpretation is important: more than half of these accumulated hours were reported under that access condition. It does not follow that more than half of visits occurred then, that most readers preferred that mode, or that most branches offered it for most of their week. The study has no visit-by-mode counts, preference measurements, or branch-level schedules with which to answer those questions.
Using common support is what allows the three percentages to be read together. A composition becomes ambiguous if one category includes observations unavailable to another, even when each separate sum is arithmetically correct. Here the matched denominator keeps the comparison explicit. For someone reading a summary dashboard, the useful first check is therefore not which percentage is largest, but whether all displayed components use the same reporting units and annual total. Only after that check does the relative size of professional service and the other modes have a stable meaning.
Why the pooled share is higher than the median
There are two legitimate ways to summarize professional-service share, and they answer different questions. The pooled share asks what fraction of all matched hours belonged to professional service. The median row share asks where the middle reporting row falls after the 277 within-row percentages are ordered. The pooled result is 43.6445%, while the median is 39.6100%. Their difference is 4.0345 percentage points, not a 4.0345% change over time. No earlier year enters this comparison.
In the pooled calculation, a reporting row contributes in proportion to its total opening volume. In the median, each reporting row supplies one position in the ordered distribution, regardless of how many hours it reports. The higher pooled result means that weighting the observed shares by their opening volumes produces a larger professional-service share than the middle row’s percentage. It does not establish a simple rule that every larger reporting unit has a higher share. That would require examining a relationship the two summary statistics alone do not describe.
The distinction changes what a reader should take away. For an account of the composition of accumulated hours, the pooled percentage is appropriate. For a description of the middle reporting row’s internal mix, the median is appropriate. Neither is an average resident’s likelihood of encountering professional help: that would require information about where people live, which locations they use, and when they visit. Presenting both summaries prevents a single aggregate from standing in for all local patterns, while avoiding the opposite mistake of treating the median row as representative of total service volume.
Quartiles describe variation without ranking libraries
The distribution of within-row shares adds information that the pooled composition cannot provide. For professional service, the first quartile is 29.67%, the median is 39.61%, and the third quartile is 58.69%. For personnel-present nonservice, the corresponding values are 0.00%, 4.50%, and 10.87%. For access without library staff, they are 35.26%, 52.98%, and 63.84%. These are summaries of reporting-row percentages on complete support, not uncertainty bands around the pooled complete-support calculation.
The three medians must not be added as though they formed a representative library’s timetable. Each is calculated separately after ordering rows by a different mode’s share; the middle position need not belong to the same reporting row. The same caution applies to the quartiles. They describe the marginal distribution of each component, whereas the pooled shares describe one common-denominator composition. A display that forces the medians into a 100% stack would silently turn three separate distributional summaries into a combination that the analysis has not observed.
For local interpretation, the quartiles can provide context for an actual reporting row’s mix, but not a quality grade. A comparatively high professional-service share may coexist with a small or large total volume; the percentage alone does not resolve that distinction. Likewise, a high share of access without library staff says nothing by itself about collection suitability or assistance during other periods. The useful application is to identify which aspect of provision needs closer examination—absolute hours, service conditions, or local scheduling—rather than to label rows as better or worse.
Four combinations reveal coexistence, not replacement
Every one of the 277 complete rows reports positive professional-service hours. Within that common condition, the observed annual combinations divide into four groups: 157 rows report positive hours in all three modes; 25 report professional service only; 80 combine professional service with access without library staff; and 15 combine professional service with personnel-present nonservice. These counts exhaust the complete support. They describe whether an annual mode total is positive, not how many locations within a row offer it.
This distinction between presence and volume explains why the combination table belongs beside, rather than instead of, the hours table. A reporting row with a small positive amount in a mode and one with a large amount receive the same presence flag. Thus, the 237 rows with positive no-library-staff access do not determine that mode’s 51.6212% pooled share by themselves. Similarly, the 172 rows with positive personnel-present nonservice do not imply that this mode occupies a large fraction of pooled hours; its matched share is 4.7343%.
For readers, the combinations show that an annual reporting unit can contain several service conditions. They do not indicate simultaneous opening arrangements, a sequence through the week, or a transition from one mode to another. In particular, the coexistence of professional service and self-service access cannot establish that the latter replaced the former. A replacement claim would need evidence about change and the relevant decision context. What is established here is more limited and more concrete: professional service is present throughout the complete-row set, alongside differing combinations of other reported opening modes.
Why Espoo’s blank cell stays missing
The original selection contains 278 named reporting rows, but Espoo’s 2024 row has a blank in the personnel-present nonservice field. That row is excluded from the main three-mode calculation, leaving 277 complete rows. Its other reported fields are not thereby declared invalid, and the exclusion is not a judgement about Espoo’s services. It is a denominator decision: a comparison requiring all three modes cannot treat an unreported component as though its value were known.
The available municipal sums match the workbook footer: 2,562,888.25 total hours, 1,119,167.50 professional-service hours, 117,998.25 personnel-present nonservice hours, and 1,325,722.50 hours of access without library staff. The corresponding original Finnish column heading is “Aukiolotunnit: omatoimikirjaston aukioloajat”; “unattended” in the audit table is an analytic column key, not the source’s English wording. The middle field has one missing reporting row, while the other listed fields have none. These available-field totals therefore have different completeness characteristics.
Accounting closure does not supply the missing observation. Even if reported components appear to reconcile with a total, that arithmetic is not permission to replace a blank with zero. The footer comparison answers whether available selected values reproduce the published aggregate, not whether every component was reported for every row. Readers should consequently keep 2,562,888.25 available-field total hours separate from the 2,492,395.25-hour matched denominator. Both totals have a legitimate descriptive role, but substituting the former into the main composition would change the support without resolving the missing middle-mode value.
Sensitivity measures influence, not uncertainty
The leave-one-row-out calculation asks how the pooled professional-service share changes when each complete reporting row is omitted in turn. Across those omissions, the resulting share ranges from 41.2572% to 43.9842%. The published table rounds the endpoints to 41.26% and 43.98%. This is a finite sensitivity exercise applied to the observed set of complete rows. It shows how dependent the pooled summary is on individual reporting-row contributions without asserting that any omitted row is erroneous or dispensable.
The result is not a confidence interval. These complete descriptive data are not being treated as a random sample from which sampling uncertainty is estimated. The range also does not quantify missing-data uncertainty, reporting error, or possible alternative definitions of opening hours. Its question is narrower: with definitions and observed values fixed, how much can one-row deletion move this particular pooled statistic? That makes it useful for interpreting aggregation, but unsuitable as an error bar suggesting an unknown population percentage lies within those endpoints.
The endpoints are not alternatives from which to select a preferred headline. The main reported share remains 43.6445% on all 277 complete rows. Removing a row changes the observed support, so the resulting percentage answers a different version of the aggregation question. For a reader evaluating a summary, the sensitivity range is best used alongside the pooled-versus-median comparison: one shows the effect of single-row omission, while the other shows the effect of changing the summary’s weighting logic. Neither establishes a ranking of local service provision.
How to use the results when planning or comparing access
For an individual visit, the annual statistics are a guide to the distinction one should check, not a substitute for a local timetable. The relevant question is whether the intended location offers professional service during the planned visit or access under another mode. Annual totals cannot identify that time window. They also cannot establish whether a particular mobile-library stop, fixed branch, or resource will be available. A reporting row’s positive annual professional-service hours do not guarantee professional help at every place or opening period within it.
For comparing published local summaries, a disciplined reading starts with four connected elements: the reporting unit, the total annual volume, the service-mode definition, and the support used for each percentage. A proportion without its total cannot tell the reader how much professional-service time was recorded. A total without its mode mix cannot show how much was professionally served. A mode label without the official definition may confuse non-library personnel with library professionals. Different support across components can then undermine an otherwise plausible comparison.
These checks also set a boundary around accessibility claims. Recorded opening volume does not measure travel time, entry eligibility for a particular person, accessibility conformance, unmet demand, or the ability to complete a specific task. It would be equally unwarranted to dismiss self-service access as useless or to count it as professionally served access. The study supplies a vocabulary and a measured composition for distinguishing those offers. A reader can use that distinction to ask a more precise local question while leaving outcomes and personal suitability to evidence that actually addresses them.
A dated descriptive account, not a story of historical change
The observation period is January–December 2024. The official annual download index lists an update dated 9 April 2025, and the supplied sources were accessed on 30 October 2026. Those dates describe different things: the year of library activity, a listed source update, and retrieval. None turns this analysis into a measure of 2026 opening conditions. This is a retrospective analysis of published 2024 returns; no earlier activity year is used to establish a trend.
The evidentiary foundation is the official statistical system, its 2024 instructions, and the published workbook. The original analysis consists of the matched three-mode composition, within-row distribution summaries, positive-mode combinations, reconciliation of available sums with the footer, and leave-one-row-out sensitivity. This is a descriptive synthesis of supplied fields, not a new definition of library service or a first-ever method. Its scope is deliberately narrower than explanations of staffing decisions, workforce qualifications, evening or weekend provision, or the effects of self-service on readers.
The clearest conclusion is therefore a distinction rather than a slogan. On complete matched support, access without library staff accounts for the largest share of reported annual opening volume, while every included row retains positive professional-service hours. Hour weighting produces a higher professional-service share than the median row percentage, and a missing component prevents the full selected set from serving as a common three-mode denominator. Read together, those findings describe how access and professional service coexist in the records. They do not tell us whether any local balance is sufficient, why it arose, or how well it serves particular readers.
Opening hours are not all professionally served
Mainland Finland · 2024 · 277 complete reporting rows
Full chart and source
Scroll horizontally to inspect the chart. ↔
Chart and values
Scroll horizontally to see all columns.
| Mode | Annual hours | Pooled share (%) | Median row share (%) | First quartile (%) | Third quartile (%) | Rows with positive hours |
|---|---|---|---|---|---|---|
| Library-professional service | 1,087,793.50 | 43.64 | 39.61 | 29.67 | 58.69 | 277 |
| Personnel present, not serving | 117,998.25 | 4.73 | 4.50 | 0.00 | 10.87 | 172 |
| Access without library staff | 1,286,603.50 | 51.62 | 52.98 | 35.26 | 63.84 | 237 |
277 named reporting rows; Espoo excluded for a blank middle-mode cell. Pooled shares use summed annual service-hours, not unique clock-hours. Medians do not sum to 100%.
Which opening modes are reported together?
Mainland Finland · 2024 · 277 complete reporting rows
Full chart and source
Scroll horizontally to inspect the chart. ↔
Chart and values
Scroll horizontally to see all columns.
| Professional | Personnel present without service | No-library-staff access | Reporting rows |
|---|---|---|---|
| 1 | 1 | 1 | 157 |
| 1 | 0 | 0 | 25 |
| 1 | 0 | 1 | 80 |
| 1 | 1 | 0 | 15 |
Positive annual hours in each mode; not simultaneous schedules. All 277 complete rows have positive professional-service hours.
Data & downloads
Scroll horizontally to see all columns.
| Mode | Annual hours | Pooled share (%) | Median row share (%) | First quartile (%) | Third quartile (%) | Rows with positive hours |
|---|---|---|---|---|---|---|
| Library-professional service | 1,087,793.50 | 43.64 | 39.61 | 29.67 | 58.69 | 277 |
| Personnel present, not serving | 117,998.25 | 4.73 | 4.50 | 0.00 | 10.87 | 172 |
| Access without library staff | 1,286,603.50 | 51.62 | 52.98 | 35.26 | 63.84 | 237 |
277 named reporting rows; Espoo excluded for a blank middle-mode cell. Pooled shares use summed annual service-hours, not unique clock-hours. Medians do not sum to 100%.
Scroll horizontally to see all columns.
| Professional | Personnel present without service | No-library-staff access | Reporting rows |
|---|---|---|---|
| 1 | 1 | 1 | 157 |
| 1 | 0 | 0 | 25 |
| 1 | 0 | 1 | 80 |
| 1 | 1 | 0 | 15 |
Positive annual hours in each mode; not simultaneous schedules. All 277 complete rows have positive professional-service hours.
Scroll horizontally to see all columns.
| Measure | Value |
|---|---|
| Source reporting rows | 278 |
| Complete exact-closing rows | 277 |
| Excluded row | Espoo, 2024; blank staff-present-nonservice mode |
| Complete-support annual hours | 2,492,395.25 |
| Pooled minus median professional share (pp) | 4.03 |
| Leave-one-row-out professional share: minimum (%) | 41.26 |
| Leave-one-row-out professional share: maximum (%) | 43.98 |
Missing values are not zero. Leave-one-row-out range is a finite sensitivity diagnostic, not a confidence interval.
Scroll horizontally to see all columns.
| Field | Municipal available sum | Workbook footer | Missing reporting rows |
|---|---|---|---|
| total | 2,562,888.25 | 2,562,888.25 | 0 |
| professional_service | 1,119,167.50 | 1,119,167.50 | 0 |
| staff_present_not_serving | 117,998.25 | 117,998.25 | 1 |
| unattended | 1,325,722.50 | 1,325,722.50 | 0 |
Available-field sums equal this footer but are not the 277-row complete support. Espoo middle mode remains blank. Field labels are analytic column keys, not verbatim source headings; unattended denotes access without library staff.
Evidence & method
How we calculated it
- Use the official 2024 workbook and the official 2024 statistical instructions, especially pages 17–18, for reporting scope and opening-mode definitions. The observation period is January–December 2024; source retrieval on 30 October 2026 is not the observation period.
- Retain the 278 named municipal reporting rows and original missing values without municipal-code joins, repairs, or imputation. The main common-support analysis uses 277 complete, exactly closing rows, excluding Espoo because its personnel-present nonservice cell is blank.
- For each opening mode, divide its summed annual hours across the 277 complete rows by their common total of 2,492,395.25 hours. Calculations use Decimal arithmetic from published values, independently checked with exact rational arithmetic; table percentages are rounded to two decimal places.
- Calculate each complete row’s mode shares using its own total, then summarize those distributions with the supplied medians and quartiles. Compare the pooled professional-service share with the median row share in percentage points, not as a temporal change.
- Classify the 277 complete rows by positive annual hours in each mode. Separately compare available-field sums across the 278 selected rows with the workbook footer, preserving the missing middle-mode value.
- Recalculate the pooled professional-service share after omitting each complete row in turn. Report the minimum and maximum as descriptive sensitivity endpoints, not confidence limits; no inferential error bars are used.
What the data cannot tell us
- The 278 selected observations are named municipal reporting rows, not branches, independent networks, separate employers, or a claim to complete coverage of all Finnish municipalities. Mainland scope excludes Åland.
- The main three-mode composition covers 277 complete rows. Espoo’s blank personnel-present nonservice cell is not zero, and footer reconciliation does not justify imputing it.
- Opening hours sum service-hour volume across service points and mobile-library stops. They are not unique clock-hours, professional working hours, branch timetables, or weekly access available to every resident; other service-place opening hours are excluded by the source.
- Personnel-present nonservice may involve guards, caretakers, or café workers rather than library professionals. Access without library staff does not establish that nobody else is on site.
- No inference is made about historical change, staffing reductions, causal replacement, workforce qualifications, evening or weekend schedules, or reasons for a reporting row’s mode mix.
- The data do not measure visits by opening mode, resident reach, unmet demand, user outcomes, accessibility conformance, or service quality. Percentages and quartiles are not quality rankings.
- Separate row-share medians and quartiles do not form a common composition. The leave-one-row-out range measures finite deletion sensitivity, not sampling uncertainty or reporting-error bounds.
- The original XLS workbook has a container defect associated with a normal corruption warning; no repaired file is substituted. The analysis retains the selected original values and missingness rather than treating that warning as evidence of particular numerical errors.
Original sources
- Finnish Public Libraries Statistics: official scope and licence
Annual January–December statistics cover mainland Finland excluding Åland, using municipal/joint-library reporting rather than individual branches; statistical data are explicitly CC BY 4.0.
Accessed October 30, 2026
- Finnish Public Libraries Statistics: annual download index
Links the 2024 workbook and lists an update dated 9 April 2025; retrieval in October 2026 is distinct from the 2024 reporting period.
Accessed October 30, 2026
- Official Finnish public-library 2024 workbook
Original 1,084,416-byte file; SHA256 f1eb98dc7b73bc567dd50b137dd821edea8bba029e0caaaf272a68cbd69d96b6. Supplies the 278 named reporting rows and footer; missing cells remain missing and the container defect is not repaired.
Accessed October 30, 2026
- Official public-library statistical instructions for 2024
Pages 17–18 define mobile-library reporting and the three opening modes; mode 2 may involve non-library personnel, and mode 3 means no library staff.
Accessed October 30, 2026
- Creative Commons Attribution 4.0 International
The provider’s statistical-data licence requires attribution, a licence link, and identification of changes; it implies neither endorsement nor blanket rights to article text or artwork.
Accessed October 30, 2026
Cite & reuse this study
Alisson Bento. (2026). Open Doors and Professional Service: Finland’s 2024 Library-Hours Mix. https://alissonbento.com/research/finland-library-opening-service-modes-2024/
The official provider assigns CC BY 4.0 to the statistical data: Suomen yleisten kirjastojen tilastot 2024, OKM. Licence: https://creativecommons.org/licenses/by/4.0/. Alisson Bento’s original calculations transform selected published fields into matched-support shares, row-distribution summaries, positive-mode combinations, footer comparisons, and sensitivity results. Reuse of the source data requires attribution, a licence link, and identification of transformations; no provider endorsement is implied. This statement does not assign a blanket licence to the article or extend the statistical-data licence to unrelated report artwork, books, or other third-party material. The instructions PDF’s artwork is not redistributed.
Corrections or editorial questions