MSc Research Skills

Lecture:
Formulating research problems, objectives and
questions
Formulating research 1

Topics

1. The MSc research proposal

2. Research problems

3. Research objectives

4. Research questions
Formulating research 2

Topic:The MSc research proposal

1. Logical structure

2. “Research”, “Design”, “Social”, “Modelling” proposals

Formulating research 3

Common elements of a research proposal

A research proposal usually has a logical structure something like:

Problem ⇒ Objectives ⇒ Questions ⇒ Hypotheses ⇒ Methods

• The problem, objectives, questions and hypotheses are usually in one

chapter called Introduction.

• This is usually followed by literature review and methods

chapters. The thesis will then have several more chapters, covering:
* results,
* discussion,
* conclusions and
* recommendations.
Formulating research 4

Additional elements in the “research” thesis proposal

• Study area: location, map, description, suitability for answering the research
questions
Formulating research 5

Additional elements in the “design” thesis proposal

A “research”-level design must have:

• A high level of innovation; in particular it must create something really new,

or at least a new synthesis;

• It must result in a design that is demonstrably better in some sense than

the alternatives;

• The thesis must both define and demonstrate this superiority.

The hypothesis of the “research” thesis is then replaced with a statement of the
proposed innovation and evaluation criteria to assess this.

The “demand” for a design replaces the “research problem” of the research
thesis.
Formulating research 6

The “social” or “organizational” thesis proposal

Social analysis: the study of humans and human societies or their organizations.

• The “hypothesis” takes the same form as a research thesis, but the
research method is different; in particular the evidence can be subjective
and anecdotal, rather than the objective result of a measurement.

• Usually includes a section on Definitions or Concepts, where terms such as

“participatory”, “sustainable”, “equitable” etc. are well-defined, so that they can
be consistently identified in the research.
Formulating research 7

The “modelling” thesis proposal

The researcher builds a conceptual or (more commonly) computational model of
a process.

This is usually applied to some real or synthetic test data.

• Models are evaluated by their success in reproducing the behaviour of the

natural or social system.

• Key issues in a modelling thesis are therefore calibration and validation.

Formulating research 8

Research problem
A general statement of why the research should be done.

This is something that is:

• not well-understood and

• can be addressed by research.

Formulating research 9

Research vs. Implementation (design)

Note that a research problem is not an implementation

• Research: something is not known or understood;

• Implementation: a known method needs to be applied to a practical problem.

However, it may not be known if an implementation will be successful . . . this

can be a subject of research.

It may not even be known how to evaluate success . . . another research problem!
Formulating research 10

Problems in a social context

1. Why should anyone care about the outcome of this

research?
• What more will we know?
• What practical problems will then be solvable?

2. Who would use the results of this research? and for what?

3. Why should anyone sponsor this research?

Formulating research 11

Categories of research problems

• Social: something (that might be) wrong with human society;

• Environmental: something (that might be) wrong with the natural world;

• Management: a deficiency in managing a social or environmental problem;

• Technical: a deficiency in methods to solve problems;

• Information: a lack of information, facts that are not known;

• Knowledge: a lack of understanding: why things happen;

Social, environmental, management and technical problems often reveal an

additional information or knowledge problem.
Formulating research 12

Example 1 – Naivasha – General statement

>“Wind erosion is causing widespread destruction of crop land and pastures in the
rift valley of Kenya.”

>“We do not know the priority areas for intervention.”

>“It is impractical to monitor wind erosion over large areas by ground survey
or conventional aerial photography.”
Formulating research 13

Example 1 – Naivasha – social problems

• The destruction caused by wind erosion;

• The lack of effective intervention.

Formulating research 14

Example 1 – Naivasha – possible research problems

• It is not known how to monitor wind erosion over large areas in a cost-
effective manner;

• Priority areas for intervention have not been identified;

• There are no established methods for identifying priority areas;

• It is not known what land-use practices are most associated with wind
erosion;

• The physical and social causes of wind erosion in this area are not known;

• Interventions to minimize erosion are not known.

Formulating research 15

Example 2 – animation – general statement

>“The province of North Brabant (NL) is digitizing land-use plans and making them
available in this form to the public and professional planners.”

>“Not all planning objects are comparable, because some of them are uncertain or
fuzzy.”

>“Uncertainty and fuzziness are hard to perceive in traditionally-mapped data.”

>“Planners are not able to correctly judge how some planning objects that are
continuous in reality influence land-use options . . . because these
continuous features are represented by crisp boundaries on the map.”

>“Planning objects, of which the location, boundaries, orientation, size and/or

shape are not well-defined, can not be judged exactly.”

>“Static graphic variables have been used to represent uncertainty and fuzziness,
but dynamic visualization methods have not yet been integrated with these.”
Formulating research 16

Example 2 – animation – social problem

Poor quality of planning decisions, because certain planning objects can not be
well-visualized by planners.
Formulating research 17

Example 2 – animation – possible research problems

This was given at the end of the general statement:

>“Static graphic variables have been used to represent uncertainty and fuzziness,
but dynamic visualization methods have not yet been integrated with these.”

So the problem is:

• It is not known how to integrate dynamic visualization methods into

representations of uncertainty and fuzziness, in spatial planning applications.
Formulating research 18

Difficulties defining research problems

Confusion between a social problem and a research problem, e.g.,

The social problem:

“In recent decades, urban sprawl of City X has dramatically increased,

especially new-style gated residential areas. The combination of private roads,
high buildings, poor sidewalk design, and few gates in these residential
areas results in difficult access to public transit.”

This leads to many other social problems: inequity, lost productivity, increased
automobile usage leading to congestion, pollution etc.; . . .

But, where is the research problem?

Formulating research 19

Possible research problems

(to be justified by literature review): all of these require research to solve.

1. It is not known what proportion of people use public transit, how they reach the
stops, how much time is required from various locations (3D) – an information
problem;

2. It is not known what are people’s motivations for using public transit vs. other
forms (automobile, walking, bicycle) – a knowledge problem;

3. The optimal placement of public transit stops and routes is not known – a
management problem;

4. Current methods for route optimization don’t take into account the time people
spend within tall buildings – a technical problem;

5.. . .
Formulating research 20

Research objectives
These are statements of what is expected as the output of the research. Each of the
objectives must be at least partially met at the end of the project.

There is usually a single general objective which is not operational.

This is broken down into a list of specific objectives which are then formulated as
research questions, which are then operationalized as research methods.
Formulating research 21

Examples of general objectives

Naivasha SFAP:

>“To determine the applicability of Small-format Aerial Photography (SFAP) to wind

erosion mapping and monitoring in the rift valley of Kenya, and the main factors
which affect its success.”

Naivasha causes:

>“To determine the causes of wind erosion in the rift valley of Kenya.”

Animation:

>“To develop methods to effectively visualize uncertainty and fuzziness in

animated representations by various combinations of graphic and
dynamic visualization variables.”

>“To select or develop a method by which the usability of uncertainty and

fuzziness display in spatial planning maps can be evaluated.”
Formulating research 22

Examples of specific objectives – Naivasha SFAP

>“To determine which wind erosion features, and of what dimensions, can be
visually interpreted on SFAP”

>“To determine the accuracy with which SFAP can be georeferenced with
single-receiver GPS and mosaicked into a seamless image”

>“To determine the costs of a SFAP mission in local conditions”

Formulating research 23

Examples of specific objectives – Naivasha causes

>“To determine factors related to wind erosion in the study area”

>“To determine which land-use practices are most associated with wind erosion”
>“To determine which soil properties are most associated with wind erosion”

>“To relate these factors with presumed processes”

>“To identify and quantify the proximate and ultimate causes of wind erosion in
the study area”
Formulating research 24

Research outputs
In some thesis projects there may also be a list of expected outputs, e.g.:

• maps

• databases

• computer programs

all of which are specified in detail.

These are logically part of the objectives (“an objective is to produce a map of . . .
”)
Formulating research 25

Research questions
These specify what the research will actually address.

• Each research question must be answered by the thesis, therefore it must be a

specific question to which an answer can be given.

• Questions follow objectives and may be simple re-statements in

operational form, i.e. where an experiment or sample can answer it.

• Questions are of two main types:

observational ‘What’, ‘where’ or ‘which’ questions;

analytical ‘Why’ or ‘how’ questions.
Formulating research 26

Question words – simple

• “Where?” (mapping)
* “Where (in the study area) is the most severe accelerated erosion?”

• “Is there” or “Does” (presence, existence)

* “Is there a water quality gradient with depth?”

* “Does water quality vary with depth?”

• “Which?” (identification)

* “Which land areas are currently used for smallholder cassava production?”
* “Which aspects of current land-use plans are most controversial?”

(continued . . . )
Formulating research 27

• “Can?” (technique), in the sense of “Is it possible?”

* “Can a light aircraft with GPS carry out a photo mission to specified accuracy
standards?” “Is it possible to see blow-outs on an air photo?”

• “What?” (results of a technique)

* “What is the accuracy of geo-referencing?”

• “What?” (is encountered in the field)

* “What are the most common species of trees planted in domestic gardens?”

• “How?” (observational)

* “How has water quality changed since the establishment of the

irrigation project?”
* “What, if any, are the change in water quality . . . ”.
Formulating research 28

Question words – complex

• “What is?” (effects)

* “What is the effect of increased grazing on vegetation density?”

• “What is?” (relation)

* “What is the relation between increased grazing and vegetation density?”; this
must be answered with a statistical model.

• “Why?” (causes)

* “Why does increased grazing affect vegetation density?”; this must be

answered with some proposed mechanism.

• “How?” (function)

* “How does increasing pesticide use in surrounding farmland

affect reproductive success of migratory bird species in the lake?”
Formulating research 29

Example of research questions (Naivasha SFAP)

>“What are the photo-interpretation elements for different wind erosion features?”
(e.g. in this case the blowouts may be darker because of the different ash in
subsoil; elongated form in wind direction etc.)

>“Can blow-outs and dunes caused by wind erosion be seen on SFAP, and if so, of
what dimensions?”

>“What is the smallest wind erosion feature than can be recognised, measuring
both vertically and horizontally?”

>“Can sufficient group control points be established to convert the set of SFAP
photos to orthophoto mosaic?”

>“What is the accuracy of such a conversion, using a single GPS receiver

for ground control?”

(continued . . . )
Formulating research 30

>“What is the cost of a SFAP mission and how does this compare with conventional
survey?”

>“What is the time required to organise a SFAP mission and produce an wind
erosion assessment, and how does this compare with conventional survey?”
Formulating research 31

Example of research questions (Naivasha causes)

>“What are the land-use practices in the study area?”

>“Which of these are most associated with wind erosion features?”

>“What is the quantitative relation between the intensity of specific land uses and
wind erosion?”

>“What is the physical process which relates the intensity of a specific land use to
wind erosion?”

>“What are the synergistic or antagonistic effects of specific land uses and other
causative factors?”

>“What is the principal cause of wind erosion in the study area?”

Formulating research 32

Example of research questions (animation)

>“Which planning objects are uncertain and fuzzy in spatial planning maps?”

>“What characteristics of these objects play a role in the plan preparation phase of
spatial planning?”

>“How can these objects be represented in an interactive animated way

by combination of graphic and dynamic visualization variables?”

>“How can the annoyance of some users by some animated effects, e.g. moving or
blinking objects, be eliminated, while still communicating the uncertainty?”

>“Which combinations of variables can best aid spatial planners in making better
decisions?”
Formulating research 33

Research Questions related to Research Objectives

One way to organize research questions is to list them as a sub-list under each
research objective.

This shows which questions, if answered, will meet each objective.

Note that objectives are declarative sentences, whereas questions are

interrogative sentences.
Formulating research 34

Example of objectives and related questions (Naivasha SFAP)

1. “To determine which wind erosion features, and of what dimensions, can be
visually interpreted on SFAP”
(a) “What are the photo-interpretation elements for different wind
erosion features?”
(b) “Can blow-outs and dunes caused by wind erosion be seen on SFAP, and if
so, of what dimensions?”
(c)“What is the smallest wind erosion feature than can be recognised, measuring
both vertically and horizontally?”

2. “To determine the accuracy with which SFAP can be georeferenced with
single-receiver GPS and mosaicked into a seamless image”
(a) “Can sufficient group control points be established to convert the set of
SFAP photos to orthophoto mosaic?”
(b) “What is the accuracy of such a conversion, using a single GPS receiver
for ground control?”
(continued . . . )
Formulating research 35

3. “To determine the costs of a SFAP mission in local conditions”

(a) “What is the cost of a SFAP mission and how does this compare
with conventional survey?”
(b) “What is the time required to organise a SFAP mission and produce an wind
erosion assessment, and how does this compare with conventional
survey?”
Formulating research 36

Hypotheses
Various definitions of this term, here we use:

Hypothesis: “[An] idea or suggestion that is based on known facts and is used as a
basis for reasoning or further investigation” (Oxford Advanced Learner’s
Dictionary, 1995)

These are the researcher’s ideas on what the research will show, before it is carried
out. They are statements that can be:

• proved,

• dis-proved, or (most likely)

• modified by the research.

They are based on previous work, usually discovered in the literature review. They
should match the research questions one-to-one.
Formulating research 37

Another definition of hypothesis in this sense is anticipated results; another term

for this is proposition.
Formulating research 38

Hypotheses should be as specific as possible

Example: Given the research question “What is the effect of grazing intensity on
vegetation density?” we can formulate the corresponding hypotheses:

• Wrong: “Grazing affects vegetation density”

• Right: “Above a threshold (to be determined), vegetation density is reduced

linearly (coefficient to be determined) with grazing intensity, measured as
animal-months.

The “to be determined” could be filled in with reference to results reported in

the
literature review, or from first principles.

The first hypothesis is too general, “affects” could be anything.

Formulating research 39

Example of hypotheses (Naivasha SFAP)

>“Blow-outs and dunes caused by wind erosion can consistently be seen on SFAP
at a nominal photo scale of 1:5 000.”

>“Both blow-outs and dunes with a vertical relief difference of as little as 1 m, and
an minimum horizontal dimension of 5 m can be seen.”

>“It is always possible to find sufficient points for direct linear transformation
within a single SFAP.”

>“SFAP can be converted to an orthophoto mosaic with a horizontal accuracy of

5 m using GPS ground control.”

>“The cost of a SFAP mission is an order of magnitude less than a conventional air
photo mission.”

>“The time required to organise a SFAP mission and produce an wind erosion
assessment is less than two weeks.”
Formulating research 40

Example of hypotheses (Naivasha causes)

>“The principal land uses are small-scale subsistence farming, paddock grazing
of cattle, and extensive grazing.”

>“Wind erosion is found only in paddock grazing.”

>“No erosion is observed until grazing intensity reaches a threshold, after which
the extent increases exponentially with grazing intensity until the whole area is
destroyed.”

>“Overgrazing leads to removal of the surface cover (grasses), exposing the soil to
the full kinetic energy of the wind.”

>“Fine-grained volcanic ash soils are more susceptible to wind erosion,

when exposed by overgrazing, than coarse-textured ash and lacustrine
soils.”
Formulating research 41

Example of hypotheses (animation)

>“Replacing blinking graphical objects with subtle low-frequency changes in

colour enhances comprehension and reduces user fatigue”.

Here the proposed design decision is stated; the truth of the statement must be
tested during the research.
Formulating research 42

Statistical hypotheses
Another use of “hypothesis” is in frequentist statistical inference.

Here the so-called null hypothesis (abbreviated H0) is a numerical statement about
some population that is to be tested on the basis of some sample; the so-called
alternate hypothesis (abbreviated H1 or H a ) is its complement.

Example: “There is no difference in mean height between third-grade boys and

girls in school district X”; or “The difference in mean height between third-grade
boys and girls in school district X is 5 cm.”

Null hypothesis H0 is really . . .

Action taken True False
Reject Type I error committed success
Don’t success Type II error committed
reject
This definition is too narrow for use in a research proposal, especially because
frequentist testing is not the only approach to statistical inference.
Formulating research 43

Assumptions
These are preconditions for research:

• taken as true;

• not questioned or verified during this research;

• difficult to specify;

• made explicit and justified if questionable.

If an assumption is false, the research is (at least partly) invalid or infeasible.

Two kinds: conceptual and logistical

Formulating research 44

Conceptual assumptions

• Laws of nature (gravity, light, . . . ): not stated

• Laws within a discipline (chemistry, soil science, . . . sociology?)

• Facts taken as true, not interfering with the factors being studied. Examples:
* “Soils are fairly homogeneous in a study area, so any differences
in biodiversity are due to other factors (the ones we will study ).”
* “Social structure in the study area is based on strong kinship ties (so we will
survey kinship, not economic relations).”

• This latter is called ceteris paribus, Latin for “with other things the same”
(see next slide)
Formulating research 45

“Ceteris paribus”

• A research question leads to one or more hypotheses;

* i.e., proposed or possible answers to the question

• So, the methods chosen to test the hypothesis only consider those
factors mentioned in the hypothesis;

• If there are other factors (untested) that in fact cause experimental or

observational differences, conclusions based on the “answer” to the hypothesis
(true, false, modified) are wrong or incomplete.

• This is especially important in causal hypotheses: “The reason that X occurs

is
Y ”; when only Y1 , Y2 . . . are considered, but the real reason is Z.

UT/ITC Enschede
Formulating research 46

Logistical assumptions
These are also called “risks”: if these conditions are not met, the research will not
be possible. Examples:

• “The study area is accessible”

• “Permission to access the study area will be granted by local authorities”

• “A translator will be assigned to the research team”

• “Samples will be processed by a laboratory correctly and within a given time”

• “A model will be updated by its author prior to the time we need it”

• “ITC will acquire a license for a specialized computer program”

Be careful with these!!

Formulating research 47

Verifying assumptions
By definition these can not be verified, but their

• plausibility (concepts) and

• feasibility (logistics)

must be argued and defended in the research

proposal. Special care should be taken with ceteris

paribus:

• have you identified other possible factors than the ones named in
the hypotheses?

• can you argue that they are not important (i.e., constant) in your
case?
Formulating research 48

Summary and next step

At this point the research has been structured as:

1.Social, contextual problems

2.Research problems

3.Research objectives

4.Research questions, several per objective

5.Research hypotheses for each question

6.Research assumptions, not to be tested

The next step is to select research

methods to answer the questions.