Investigation planning
'Planning' (class obo:OBI_0000339) is a sub-process of investigations. It comprises all preparations of an investigation that are undertaken before work on the material to be studied begins.
Figure: General structure of investigation planning (cf. key to symbols in network graphs).
Planning has one or more hypotheses (class obo:OBI_0001908 'Testable hypothesis') as input and a study design specification (obo:OBI_0500000 'Study design') as output. The study design, in turn, is applied to a specific investigation by means of a study plan (obo:OBI_0000260 'Plan'). This plan is the basis for the execution of the investigation (obo:OBI_0000471 'Study design execution').
Though not formally a part of planning (in terms of the ontology), the investigation plan is still part of preparations for a study. This structure shows that planning in the sense of class obo:OBI_0000339 is concerned with the abstract preparation of an investigation type (study design), while the plan represents the concrete preparation of a specific investigation.
This structure is outlined in the documentation of the OBI core. It concurs with, but is not directly reflected by axioms in the OBI itself. Here, class obo:OBI_0000260 ('Plan') is just a generic subclass of class obo:BFO_0000017 ('Realizable entity'), without any restrictions. The class obo:OBI_0000471 ('Study design execution') has a restriction on property obo:BFO_0000055 ('realizes') with instances as range that are in the domain of obo:RO_0000059 ('concretizes') and where class obo:OBI_0500000 ('Study design') is in the range. This restriction can be satisfied by instances of class obo:OBI_0000260 ('Plan'), but the class is not explicitly named.
Against the understanding conveyed by the documentation of the OBI core might stand the fact that class obo:OBI_0001636 ('Grant') has a restriction on property obo:BFO_0000050 ('part of') with class obo:OBI_0500000 ('Study design') as range. The definition of 'grant' in this context clarifies that it refers to a specific grant given by a specific institution for a specific purpose. This could be interpreted to make the related study design specific to a particular investigation. With this understanding, however, there would be no need for an additional plan (class obo:OBI_0000260). An alternative interpretation would see the grant specification referring to a specific grant that either funded the development of the research design itself or a series of investigations, all realising the design.
In the context of RDFBones, we go with the understanding of the documentation of the OBI core, where a study design is a general specification, while a plan is understood to be specific to a particular investigation.
Testable hypotheses that go into the planning process are formulated on the same level as its output, the study design. They are statements applying to all kinds of circumstances and do not refer to specific investigations analysing specific materials.
The OBI defines class obo:OBI_0001908 ('Testable hypothesis') as follows:
An information content entity that expresses an assertion that is intended to be tested.
This makes clear that all hypotheses that go into the planning process need to be tested by the investigation. If a research design is based on a number of hypotheses that are generally testable but hypothesis testing is not among the investigation's objectives, these hypotheses cannot be specified as input to the class obo:OBI_0000339 ('Planning').
The study design contains all information on what a certain type of investigation comprises, the materials and instruments used therein and the actions that are to be performed while conducting it. These specifications are encoded in instances of specific classes. Study designs are related to a type of investigation, not to individual investigations (see discussion above).
-> read more about study designs
As the study design relates to a type of investigation (see above), individual investigations require additional information on how the study design is to be applied to a specific body of material under given conditions. This information is represented by an individual of class obo:OBI_0000260 ('Plan'). This class implies no restrictions and researchers can relate documents and other types of documentation to it, according to their own preferences.
In biological anthropology, it is not always common practice to formulate a hypothesis that is to be verified or falsified by an investigation. In many cases, this might be seen as negligence. On the other hand, in the rather narrow understanding of 'investigation' in the context of RDFBones (cf. Investigations), hypotheses might not always be testable or hypothesis testing might not be the intent of an investigation. For example, age estimation from skeletal remains is based on a number of hypotheses that are generally testable. However, carrying out an osteological age estimation does not test these hypotheses, but is solely carried out to obtain age estimation ranges for the skeletal individuals under investigation. It would be possible to formulate a null hypothesis (e.g. 'Age at death cannot be estimated from skeletal remains') which then could be falsified by an investigation or not. It is questionable if this would be helpful and it is not what the OBI construct of hypothesis input is intended for.
So with any new investigation type, it has to be decided if there is a hypothesis that is to be tested by the investigations of this type. If hypothesis testing is a component of the investigation type, a testable hypothesis must be specified by any means. Only if the formulation of a hypothesis does not help in clarifying its objectives (cf. the example above), it can be omitted. Only hypotheses that can be tested by a specific investigation should be identified as input. On the other hand, it is possible that several hypotheses might be tested by one investigation type.
Figure: Relation of instances of class obo:OBI_0001908 ('Testable hypothesis') to subclasses of obo:BFO_0000001 ('Entity').
Class obo:OBI_0001908 ('Testable hypothesis') has a restriction on the object property obo:IAO_0000136 ('is about') with class obo:BFO_0000001 ('Entity') as range. This is intended to specify elements that are relevant to a certain hypothesis. As hypotheses are formulated in general terms (e.g. referring to a certain type of skeletal element rather than to a specific skeletal element), instances of class obo:OBI_0001908 ('Testable hypothesis') will be about subclasses of class obo:BFO_0000001 ('Entity'), not about their instances. Similarly, the instances describing individual hypotheses need to be specified as classified input of all planning processes being part of a certain type of investigation, i.e. this specification also has to relate to a class.
The main accommodation to a specific type of investigation is the specification of a study design. Class obo:OBI_0500000 ('Study design') has a number of subclasses, primarily to sort study designs by type. When formulating a new study design, heed should be taken to place the new instance in the correct subclass.
Figure: Definition of classes and instances in extensions and creation of instance data (cf. key to symbols in network graphs).
While data for an investigation are being entered, instances of subclasses of classes obo:OBI_0000066 ('Planning') and obo:OBI_0000471 ('Study design execution') can be created automatically by the software, based on queries (cf. investigations).
Each RDFBones extension implementing a new type of investigation has to define a new instance of obo:OBI_0500000 ('Study design') or one of its subclasses ('Study Design Investigation Type 1' in the graph above). This instance is automatically defined as output of investigation planning by a restriction on the subclass of class obo:OBI_0000066 ('Planning') that is related with the investigation type (ext:PlanningInvestigationType1 in the graph above). So the relation between any instance of this class ('Planning A' in the graph above) and the instance representing the study design can be automatically created by the software.
Likewise, a restriction on the class representing investigation planning (ext:PlanningInvestigationType1 in the graph above) defines which testable hypotheses (if any) go into the planning process. This relation can also be established automatically.
Finally, each investigation needs an instance of class obo:OBI_0000260 ('Plan') to apply the study design to a specific study. As this is the case for all investigations, regardless of their type, the required relations can also be automatically created.
