The GeoSTORM IDEA Framework

One of the first difficulties encountered by graduate students is developing their research ideas.

Most students do not begin with a polished research question. They begin with an interest:

• I want to study drought.

• I want to use satellite data.

• I want to understand tropical-cyclone rainfall.

• I want to examine how land-use change affects local weather. 

That is a normal place to start.

GeoSTORM’s Research Themes present some of the broader scientific problems currently explored within the laboratory. These themes are not ready-made thesis topics. They provide possible directions from which students can identify a more specific problem that matches their interests, preparation, and the evidence available to them.

A broad topic, however, is not yet a research question. A dataset, model, or method is not a research question either.

A research question must identify what we are trying to understand, what evidence is needed, and what the study may contribute. Within GeoSTORM, we use the IDEA framework to help develop that line of questioning:

Investigate, Describe, Explain, Advance, or IDEA, is not a requirement to write four separate questions. It is a way of checking whether the questions form a coherent scientific investigation. 

Research as deliberate design

Research questions are rarely written correctly on the first attempt. They develop as the student reads the literature, understands the scientific problem, examines the available data, and discusses possible approaches with the adviser and committee.

In this sense, developing research questions has something in common with design thinking. The researcher first explores and defines the problem, considers possible approaches, tests whether these approaches work, and revises the design when necessary. Established design-thinking approaches similarly emphasise problem definition, iteration, and testing rather than following a rigid sequence once.

It also benefits from systems thinking. Weather and environmental problems usually involve several interacting components, processes, and scales. We therefore need to understand the larger system before deciding which part can reasonably be studied in one thesis. Systems thinking is particularly useful when relationships are interconnected and cannot be reduced immediately to a simple linear cause and effect.

This does not mean that every student must produce a formal systems model. It means that the student should be able to explain:

• what belongs within the problem being studied;

• which relationships matter;

• where the study boundaries are placed; and

• what cannot be concluded outside those boundaries.

The IDEA framework

IDEA is not a four-question template. It identifies four functions that should be considered when developing a coherent research project:

Investigate the problem. Describe the evidence. Explain the findings. Advance understanding.

These functions do not always appear as separate questions. Investigate usually shapes the main research question, while Describe and Explain are developed through supporting questions. Advance is normally expressed as the intended contribution of the study rather than as an additional research question.

Investigate

Investigate identifies the central phenomenon, relationship, or uncertainty that holds the study together. It usually leads to the main research question.

Instead of beginning with a method:

• How can satellite data be used to study drought?

we might ask:

• How effectively do selected remote-sensing indices characterise drought conditions in the study area?

The revised question begins with the scientific problem rather than the tool. The method remains important, but it is selected because it helps answer the question.

A useful prompt is:

• What specifically are we trying to understand, compare, evaluate, or resolve?

Describe

Describe identifies the patterns, differences, trends, distributions, or recurring conditions that must first be established.

Descriptive supporting questions may define the study’s scope by specifying:

• the phenomenon or population being examined;

• the study area and period;

• the conditions or groups being compared; and

• the characteristics or outcomes being measured.

Examples include:

• How do the magnitude and spatial extent of extreme precipitation vary with tropical-cyclone intensity and movement speed?

• How do these relationships differ between the Southwest Monsoon season from June to September and the Northeast Monsoon season from October to December?

Description is not necessarily simple. Developing a reliable climatology, classification, comparison, or spatial analysis may require substantial scientific judgement. What matters is that the descriptive work provides evidence needed to address the larger investigation.

A useful prompt is:

• What must first be observed, measured, or compared before the main question can be answered?

Explain

Explain moves the study beyond reporting patterns. It considers the relationships, processes, and interpretations that may account for what was observed.

This may involve:

• relationships among variables;

• physical or environmental processes;

• competing explanations;

• consistency with previous studies;

• methodological differences;

• uncertainty and limitations; and

• questions that cannot be resolved using the current evidence alone.

Examples include:

• What environmental conditions may account for agreement or disagreement among the drought indices?

• How does changing terrain elevation alter the circulation and precipitation of a simulated landfalling tropical cyclone?

The wording must match the strength of the evidence. Observational studies may support language such as associated with, related to, or consistent with. Stronger causal claims require a design capable of isolating or testing the process concerned.

A useful prompt is:

• What relationships or processes may account for the observed patterns, and what can the available evidence reasonably support?

Advance

Advance refers to what becomes better understood, more clearly demonstrated, or newly possible after the main question and supporting questions have been answered.

The advance may take the form of:

• new empirical evidence;

• a better regional characterisation;

• an evaluation of available methods or datasets;

• clarification of an unresolved relationship;

• deeper process understanding; or

• a new or improved method, classification, or conceptual model.

Advance does not usually need to appear as a separate research question. It may instead be stated in the study rationale, objectives, significance, expected contribution, or synthesis.

A useful prompt is:

• If the study succeeds, what will we understand or be able to do that we could not do before?

At the MS level, this advance may emerge from answering a focused question set well. At the PhD level, the intended advance should help determine which major investigations are included and how their findings are brought together.

Applying IDEA at MS and PhD Levels

IDEA does not prescribe a fixed number of research questions. It helps organise the scientific problem, supporting investigations, and intended contribution.

For an MS thesis, this commonly takes the form of one main question, two or three supporting questions that provide description and explanation, and a focused contribution that emerges from the study as a whole.

For a PhD dissertation, one overarching scientific problem may be addressed through two or more linked main questions, each with its own supporting questions. The doctoral contribution emerges from integrating these investigations into a larger and more substantial advance.

This distinction is consistent with the standards of the UP Diliman College of Science for Graduate Programs. An MS thesis is expected to embody original and significant research and make a worthwhile contribution to scientific knowledge, while a PhD dissertation must embody original, independent, and significant scientific research.

Advance therefore matters at both levels. At the MS level, it may emerge from answering a focused question set well. At the PhD level, the intended advance should help determine which major investigations are included in the dissertation and how they are connected.

The official College of Science Graduate Student Guide provides links to the current graduate guidelines and Thesis and Dissertation Manual. IDEA is intended to support these standards, not replace them.

A final IDEA check

Before settling on a question set, ask:

Investigate: Is the central scientific problem clear?

Describe: Will the study establish the evidence needed to understand it?

Explain: Will the analysis examine relationships, processes, meaning, and uncertainty?

Advance: Is the intended contribution identifiable and appropriate to the degree level?

The questions will probably change as the research develops. That is part of the process. What matters is that they continue to form a coherent, defensible, and answerable investigation.

IDEA In Practice

The same research interest can be refined by asking what must be described, what may explain the findings, and what the study could ultimately advance. This section describes how a general question can be refined into a scientific question that is appropriate for graduate-level research work.

Investigate

• Why do some tropical cyclones produce more extreme precipitation over Luzon?

This identifies the scientific problem, but it is still broad.

Describe

• Among westward-moving tropical cyclones making their first landfall along Luzon between1978–2015, how did the magnitude and spatial extent of extreme precipitation vary with pre-landfall intensity and movement speed?

• How did these relationships differ between the Southwest Monsoon season from June to September and the Northeast Monsoon season from October to December?

This specifies the rainfall characteristics and storm conditions that will be compared.

Explain

• What environmental conditions may account for the differences in extreme precipitation?

This moves beyond describing patterns and considers the relationships and physical processes that may account for them.

This moves from documenting relationships toward interpreting their possible meteorological basis and recognising what observational analysis cannot resolve.

Advance

For the example above, the intended advance is to connect the historical relationships with controlled terrain-sensitivity experiments, clarifying how storm characteristics, monsoon conditions, and the Cordillera Mountain Range shape tropical-cyclone precipitation over northern Luzon.

This also demonstrates that Investigate may remain broad, while the Describe questions carry much of the operational scope. 

• 2026-07-19 First draft
• 2026-07-23 Clarified example in IDEA in practice