Written and maintained by CASRAI Editorial Board
Last updated
GrupLAC is the group-level record in Colombia’s ScienTI platform. Where CvLAC holds an individual researcher’s CV, GrupLAC holds the research group: its identity, its research lines, its declared age, its projects, its members, and the production those members claim. It is the record that the Ministerio de Ciencia, Tecnología e Innovación (MinCiencias) reads when it decides whether a group is recognised at all, and then whether it is classified A1, A, B or C.
The distinction most commonly missed by group leaders: a group does not have production of its own. Every product that counts toward the group’s category is a product one of its members registered in that member’s own CvLAC and then linked to the group. A group with excellent researchers and neglected CvLAC records measures as a weak group, because the model can only see what was registered inside the observation window.
This page states the requirements as they appear in MinCiencias’ own published measurement model — the document Modelo de medición de grupos de investigación, desarrollo tecnológico o de innovación y de reconocimiento de investigadores, internal code M601PR04G01, versión 02, issued as Anexo 1 to the 2024 national convocatoria (252 pages, published on minciencias.gov.co). Where a figure is not stated in that document, this page says so rather than guessing.
A note on the name: Colciencias is now MinCiencias
Colombia’s science agency was Colciencias (the Departamento Administrativo de Ciencia, Tecnología e Innovación) until it was elevated to a ministry — the Ministerio de Ciencia, Tecnología e Innovación, generally shortened to MinCiencias or Minciencias. The measurement model document itself shows the transition in its own history: it attributes the 2017 call (Convocatoria 781) to “Colciencias” and issues the current model under the “Ministerio de Ciencia, Tecnología e Innovación”.
The practical consequence for a group leader is search, not substance: a great deal of institutional guidance, university intranet pages and older training material still says “Colciencias”, and the group categories A1–C are still colloquially called “categorías de Colciencias”. Use “MinCiencias” for anything current — convocatoria terms, appeals, official correspondence — and treat a source that only says “Colciencias” as potentially describing a superseded model.
The three ScienTI-Colombia registries, and which one you are in
ScienTI-Colombia is not one database but three linked ones. Knowing which one a task belongs to prevents most of the wasted effort in a measurement cycle:
- CvLAC — the individual researcher’s CV. One per person. The person controls it. All production originates here. See CASRAI’s guide to CvLAC, Colombia’s national researcher CV platform.
- GrupLAC — the research group record. One per group. Only the group leader can edit it, including linking members. This page’s subject.
- InstituLAC — the institutional record. Universities, hospitals, companies and research centres use it to endorse (“avalar”) the groups they host, and to certify member conditions.
The three are relational: a group’s measured output is assembled from its members’ CvLAC records, and an institution’s InstituLAC profile aggregates the groups it endorses. This is the same layered structure a CRIS (Current Research Information System) implements — person, output, organisational unit — expressed as a national, state-run system rather than an institutional installation.
Registering a group in GrupLAC, step by step
The order matters. Several steps cannot be completed out of sequence, and the most common failure is a group that exists in GrupLAC but was never endorsed, and so is invisible to the convocatoria.
- Every intended member creates or updates a CvLAC. The group cannot claim a product that is not in a member’s CvLAC. Do this first, not after the group record is created.
- The prospective leader registers the group in GrupLAC on the ScienTI platform, declaring the group’s name, its research lines (“líneas de investigación”), and its date of creation. That declared date becomes the group’s edad declarada, and the model measures age as the age reached at the close of the convocatoria’s observation window. It is load-bearing — see the production requirement below.
- Register institutional membership (“pertenencia institucional”) and obtain the aval. The model is explicit that the group must previously have registered its institutional membership, and must be endorsed by at least one institution registered in InstituLAC. In practice the institution’s research office (vicerrectoría de investigación or equivalent) performs the endorsement from the InstituLAC side; the group leader cannot self-endorse.
- Link the members. Only the leader can link members from GrupLAC — but the link only takes effect when the person who owns that CvLAC authorises it, and the leader then records the vinculación period. A member who never authorises is not a member for measurement purposes, however long they have worked with the group.
- Register at least one project in execution, and link the group’s products.
- Register for the convocatoria when it opens. This is a separate action from existing in GrupLAC, and skipping it is the single most consequential mistake on this page — see the next section.
The eight recognition requirements, verbatim from the model
MinCiencias defines a research, technological development or innovation group as “the set of people who interact to research and generate knowledge products in one or several themes, according to a short-, medium- or long-term work plan”. Recognition consists of verifying and validating that the group meets each of eight minimum requirements. All eight, not most:
| # | Requirement | What trips groups up |
|---|---|---|
| 1 | Be registered in the GrupLAC system of the ScienTI platform | Straightforward, and the only one most groups think about |
| 2 | Have a minimum of two (2) members | A single-researcher “group” is not a group under the model |
| 3 | Have one (1) or more years of existence (declared age) | Age is measured at the close of the observation window, not at registration |
| 4 | Be endorsed by at least one (1) institution registered in InstituLAC; institutional membership must have been registered beforehand | The most common silent failure — the aval is the institution’s action, on the institution’s timetable |
| 5 | Have at least one (1) research, technological development or innovation project in execution | “In execution”, not merely proposed or completed years ago |
| 6 | The group leader must hold an undergraduate (pregrado), master’s or doctoral degree; if only pregrado, it must have been obtained before the close of the observation window | Applies to the leader at the closing date of the convocatoria |
| 7 | New-knowledge or technological-development-and-innovation production in the observation window equivalent to a minimum of one (1) product per declared year of existence | This is why a generously back-dated creation date is dangerous — every declared year raises the bar |
| 8 | Social appropriation of knowledge, public science communication, or human-resource-training production in the window, again a minimum of one (1) product per declared year of existence | A publication-only group fails here; outreach and training are not optional |
Requirements 7 and 8 are the ones worth planning around. Because both scale with declared age, declaring an older founding date to look established directly increases the number of products the group must show. A group that declares eight years of existence must show at least eight qualifying new-knowledge or technological-development products and at least eight appropriation/communication/training products within the window.
Recognition, once granted, is valid until the final results of the next convocatoria are published — it is not a permanent status and not a fixed number of years. After recognition, the group and its production become visible on the ScienTI platform, and everything registered in the GrupLAC — products and members — is publicly consultable.
Recognition is not classification, and measurement is opt-in
These are two separate outcomes, and conflating them is the most expensive misunderstanding in the whole process:
- Recognised (reconocido) means the eight requirements above were verified. The group exists officially. It has no letter.
- Classified (categorizado) means the group was placed in A1, A, B or C.
The model states the gate plainly: only groups that meet the criteria to be a group and that registered to be measured (“que se hayan inscrito para ser medidos”) are categorised. A recognised group that did not sign up for measurement in that convocatoria simply has no category, regardless of how strong its production is. If an institution’s internal incentive scheme, a funding call, or an accreditation review asks for a group category, “recognised but not measured” reads to that reviewer as no category at all.
Note also that the historical model once had five categories — A1, A, B, C and D, under the 2008 model. D no longer exists. If a source lists a category D, it is describing a superseded model, and any thresholds it quotes should be discarded.
What each category actually requires
These are the criteria as stated in the current model document. Every numbered condition must be met — the categories are conjunctive requirement sets, not a score band. Quartiles are computed within the OECD broad knowledge areas, so a group competes against groups in its own area, not against all of Colombian science.
| Condition | A1 | A | B | C |
|---|---|---|---|---|
| Group Indicator | In quartile 1 (top 25%) | At or above quartile 2 (top 50%) | At or above quartile 3 (top 75%) | Greater than zero |
| High-quality production | TOP-type products indicator in quartile 1 | TOP-type or Type A products indicator > 0 | TOP-type or Type A products indicator > 0 | TOP-type or Type A products indicator > 0 |
| Social appropriation / public communication | Either indicator > 0 | Either indicator > 0 | Either indicator > 0 | Either indicator > 0 |
| Human-resource training (CTeI) | Type A indicator > 0* | Type A indicator > 0* | Type A indicator > 0, or Type B at/above quartile 2 | Type A or Type B indicator > 0 |
| Required senior member | One emérito, sénior or asociado researcher, contractually employed by an SNCTI institution and resident in Colombia | Same as A1 | One emérito, sénior, asociado or junior researcher, or a linked member with a doctorate, contractually employed by an SNCTI institution and resident in Colombia | Not required |
| Cohesion indicator | > 0 | > 0 | > 0 | Not required |
| Minimum years of existence | 5 years | 5 years | 3 years | 2 years |
* For groups outside the university sector, the model allows the Type A human-resource-training requirement to be substituted with TOP-type products, an obtained “patente A”, a plant variety, an animal breed, or five (5) TOP-type articles within the convocatoria’s five-year observation window.
Three things fall out of this table that are not obvious from the category letters:
- A1 and A differ mainly in two conditions, not in kind. The requirements are otherwise identical; A1 additionally demands quartile-1 placement on both the Group Indicator and the TOP-products indicator, where A needs quartile 2 or better and only a non-zero high-quality indicator. A group sitting just outside A1 is usually short on concentration of top-tier output, not on breadth.
- C is the only category with no cohesion requirement and no senior-member requirement. A young group of two members with modest but real output across all four product families can be C. It cannot be B until it has a cohesion indicator above zero — which requires at least one co-authored product between its own members.
- Age is a hard floor at every level. No amount of production makes a four-year-old group A1.
How the group’s score is assembled from its members’ CvLAC records
This is the mechanism the individual-vs-group distinction actually rests on. Four points from the model change how a group should behave day to day:
Production is normalised, not counted
The production indicator uses a natural logarithm of the number of products in a subtype divided by the “period of observation” — and the model defines that period as the smaller of the product type’s observation window and the group’s own period of existence. A young group is therefore not penalised for having existed for less time than the window, and a large group cannot win by sheer volume in one subtype: the log dampens it. The model also applies an explicit “eliminación de efectos de escala” step for the same reason.
Production registered before the group was created does not count
The model states directly that production registered in the applications, produced before the group’s creation, is not taken into account. Recruiting a productive researcher does not import that researcher’s back catalogue into the group.
Cohesion is co-authorship inside the group, and it has a data-entry trap
The cohesion indicator is computed from co-authorships registered in CvLAC, on new-knowledge and technological-development/innovation products, between members of the same group. The cooperation indicator does the same thing between members of different groups.
The trap is mechanical and it silently costs groups their category. For a co-authored product to register as co-authorship, the model instructs that the author be entered with the coauthor’s cédula number — so the system links directly to the corresponding CvLAC — or with the name exactly as it is registered in that coauthor’s CvLAC. A name typed freehand, or in a different form from the CvLAC record, produces a product with no detected internal co-authorship. Since B, A and A1 all require a cohesion indicator above zero, a group can do genuinely collaborative work and still fail the cohesion condition purely on data entry.
Observation windows differ by product type
The general observation window in the model is five (5) years, with differentiated windows by product type:
| Product type | Observation window |
|---|---|
| Research articles type A1, A2, B, C | 7 years |
| Research articles type D | 5 years |
| Books resulting from research | 7 years |
| Patentable products | 10 years |
| Plant variety / new animal breed certificates | 10 years |
| Research-creation in Arts, Architecture and Design | 10 years |
| All other products | 5 years |
For the 2024 convocatoria the model set the observation window as 1 January 2019 to 31 December 2023. Only products whose observation window is still current are included when the group’s production and collaboration indicators are computed — so an article that was decisive in one cycle can drop out of the next one without anything about the group changing.
Note that the article types A1/A2/B/C/D here are journal classifications, assigned through Publindex, Colombia’s national journal classification system, and are a different scale from the group categories A1/A/B/C. The two share letters and mean different things. Where a group publishes is therefore not a neutral choice: it determines both the weight and the shelf life of that product inside the measurement.
Member types: what the system decides about your people
When a CvLAC is linked as a group member, ScienTI automatically classifies that person into one of four types — investigadores, investigadores en formación, estudiantes de pregrado, and integrante vinculado — which are further subdivided into eight subtypes. The leader does not choose the type; the system assigns it from the characteristics evidenced in the CvLAC.
Assignment follows a stated priority order, taking the first that matches: Investigador Emérito, then Sénior, then Asociado, then Junior, then linked member with a doctorate, then doctoral student, then linked member with a master’s or clinical specialty, and onward. The types are also refreshed automatically across the whole system at each convocatoria.
Two operational consequences. First, a member’s own CvLAC completeness changes the group’s options: the A1/A senior-member condition is satisfied by having an emérito, sénior or asociado researcher, and that recognition is itself derived from that person’s CvLAC record. Second, the member must authorise the link — so a group’s roster is only as accurate as its members’ willingness to accept the vinculación in their own CvLAC.
What could not be verified, and where to check it
Following CASRAI’s practice of stating verification limits rather than filling gaps:
- The convocatoria currently open. The most recent measurement call this page could confirm is Convocatoria 957, described by MinCiencias as a national call for actualización y transición for the recognition and measurement of groups and recognition of researchers, associated with the 2024 model and its 1 January 2019 – 31 December 2023 window. MinCiencias’ own archived results page lists calls only through Convocatoria 737 (2015) and is not a current index. Confirm the open call, its dates and its exact terms of reference on minciencias.gov.co before acting — convocatoria numbers, windows and closing dates change every cycle.
- Quartile cut-off values. The model defines quartiles statistically and computes them within OECD knowledge areas from the distribution of that cycle’s participating groups. There is therefore no fixed numeric threshold for A1 or A to publish, and any source quoting one as a permanent figure is describing a single past cycle’s outcome, not a rule.
- Relative product weights. The model does assign relative scores per product category, used to weight each product’s contribution to the production indicator; they are tabulated across the document’s annexes. This page deliberately does not reproduce individual weights, because they are revised between model versions and a stale weight is worse than none. Read them from the version of the model attached to the call you are actually entering.
- The word “transición” in the 957 title suggests the model is in a transitional state. The 2024 model document itself does not describe a successor model, so what follows it is unconfirmed here.
Common GrupLAC problems and how to fix them
The group is registered but does not appear in the convocatoria
Almost always the missing aval institucional (requirement 4) or the missing registration for measurement. Both are actions by someone other than you — the institution’s InstituLAC administrator, and the leader at the point the call opens. Check both before assuming a platform fault.
A member’s products are not counting toward the group
Check three things in order: the member authorised the vinculación; the leader recorded the vinculación period; and the product’s date falls inside the applicable observation window and after the group’s declared creation date.
The cohesion indicator is zero despite internal co-authorship
Re-enter the coauthors on those products using the cédula number, or the exact name as registered in the coauthor’s CvLAC. This is the fix for the most consequential silent failure in the model.
The group’s category dropped without any change in activity
Products age out. Articles carry a 7-year window and most other products 5, so a cycle’s shift moves the window forward and drops older output. A stable category needs a stable rate of production, not a stock of it.
Duplicate or split researcher records
Because everything is keyed to CvLAC, a researcher with two CvLAC records splits their production and can break co-authorship detection. Resolve at the CvLAC level first; GrupLAC will not fix it.
Why this matters for research administrators
For a Colombian institution, GrupLAC categories are not an internal metric. They feed institutional accreditation processes, they are cited in MinCiencias funding calls — which typically require applicants to hold an active CvLAC and named group members to be registered in GrupLAC — and they are commonly used in internal incentive and promotion schemes. A research office that treats CvLAC upkeep as the individual researcher’s private housekeeping is, in effect, leaving the institution’s group categories to chance.
The administrative implication is specific: the highest-leverage intervention is not writing more papers, it is a pre-convocatoria data audit — verify avales are current in InstituLAC, verify every intended member has authorised their vinculación, verify internal co-authorships are entered by cédula, and verify that declared group ages are defensible against the per-year production requirements. Those four checks address the failure modes above, and none of them require any new research output.
Colombia is not unusual in running a national researcher registry of this kind; it is unusual in running a three-tier one that measures groups rather than only individuals. Argentina’s SIGEVA and CVar handle CV registration and funding administration together; India’s VIDWAN expert database operates at the individual and institutional level. None of them make a group’s classification a direct function of what its members individually registered. That is the specific thing ScienTI does, and the specific reason GrupLAC rewards administrative discipline as much as research strength.
For the wider standards context — persistent identifiers, CRIS interoperability and how national systems relate to ORCID iDs — see CASRAI’s CRIS and identifiers cluster. For the integrity framework these registrations sit inside, see Colombia’s research integrity oversight.
Frequently asked questions
What is GrupLAC?
GrupLAC is the research-group registry of Colombia’s ScienTI platform, operated by MinCiencias. It holds a group’s identity, research lines, declared age, projects, members and products, and it is the record against which MinCiencias recognises and classifies research groups.
What is the difference between CvLAC and GrupLAC?
CvLAC is the individual researcher’s CV, controlled by that researcher. GrupLAC is the group record, editable only by the group leader. The group’s production is assembled from products its members registered in their own CvLACs and linked to the group — so the group’s category is a function of its members’ individual record-keeping.
What are the GrupLAC categories A1, A, B and C?
They are the classification levels MinCiencias assigns to recognised groups that registered to be measured. A1 is the highest. Each is a set of conditions that must all be met, covering the group’s overall indicator quartile, high-quality production, social appropriation, human-resource training, a required senior member (for A1, A and B), cohesion, and a minimum age of 5, 5, 3 and 2 years respectively. Category D existed under the 2008 model and no longer does.
How many members does a research group need in Colombia?
A minimum of two. That is requirement 2 of the eight minimum recognition requirements in MinCiencias’ measurement model.
Is a recognised group automatically classified A1–C?
No. Recognition and classification are separate. Only groups that meet the requirements and registered to be measured in that convocatoria are categorised. A recognised group that did not register for measurement has no category.
Is GrupLAC the same as Colciencias?
Colciencias was the agency, not the platform, and it is now MinCiencias — the Ministerio de Ciencia, Tecnología e Innovación. GrupLAC is one of the three registries (with CvLAC and InstituLAC) in the ScienTI platform that MinCiencias operates. Group categories are still colloquially called “categorías de Colciencias”, but current terms of reference use MinCiencias.
Do international collaborators need to be in GrupLAC?
Only if they are formally linked as members of the Colombian group. Linking requires the person to hold a CvLAC and to authorise the vinculación. Note that the senior-member condition for categories A1, A and B specifically requires a researcher contractually employed by an SNCTI institution and resident in Colombia, so an international collaborator cannot satisfy it.
Why did our group’s category fall?
The most common causes are products ageing out of the observation window, a lapsed institutional aval, members who never authorised their vinculación, and a cohesion indicator of zero caused by co-authors entered by freehand name instead of cédula. All four are data conditions, not research conditions.
Where do I check the current convocatoria?
On minciencias.gov.co. The measurement model is published under Sistemas de información → Modelo de medición de grupos, and each call publishes its own terms of reference with the observation window and dates. Verify against the open call rather than a summary, including this one.
