Objectives and hypotheses Wim Buysse Ric Coe RUFORUM 28 November 2006 Research Methods Group.
-
date post
21-Dec-2015 -
Category
Documents
-
view
222 -
download
0
Transcript of Objectives and hypotheses Wim Buysse Ric Coe RUFORUM 28 November 2006 Research Methods Group.
Objectives and hypotheses
Wim BuysseRic Coe
RUFORUM 28 November 2006
Research Methods Group
Part 1. The importance of setting objectives
Research Methods Group
Where do you want to be in life 10 years from now ?
Research Methods Group
Where does my organisation/firm wants to be 10 years from now ?
Research Methods Group
Keeney, R.L.1992Value-Focused Thinking: A Path to Creative Decision making. Cambridge, MA: Harvard University Press.
Research Methods Group
Research Methods Group
Research Methods Group
A real life example
Research Methods Group
• You have a job in Nairobi• Your wife is offered a job in Kisumu• Alternatives: we live apart or ‘she’ can’t work• What are your values – you want to live in the
same town, and you both want to work• Focusing on values drives you to create
alternatives (as opposed to choosing among existing alternatives)
• Innovate – You convince your company you can start up a new office in Kisumu
Setting objectives
Research Methods Group
Setting objectives (in life/in management of an organisation) provides structure, purposiveness, focus, meaning and enables you to measure progress in the right direction or the lack of it.
Setting objectives in a research project
Research Methods Group
The hypothetico-deductive method
Research Methods Group
• induction = inferring a general law or principle from the observation of particular instances
• deduction = inferring from generals to particulars
Research = learning to solve problems
Research Methods Group
So far it has been a soup
Research Methods Group
"Science consists of both knowledge and the process by which this knowledge is created, research. Although research succeeds by building and testing better theories, the process involves constructs, concepts, and activities that are not themselves predictive. Non-predictive constructs serve science as logical devices, memory aids, inspirational prods, incentives to thought, political opinions, personal ideals, half-formed notions, odd beliefs, and unexpressed ideas. These elements are not 'bad' or unscientific. They form a prescientific soup from which each scientist draws inspiration and from which the disciplined human mind has constructed modern science."
Peters, Robert Henry (1991) A critique for ecology. Cambridge University Press, Cambridge, UK. 366 pp. ISBN 0 521 39588 7 - p. 21
synthetic or private phase
Synthetic or private phase
Research Methods Group
•Logic•Knowledge of discipline•Creativity•Imagination•Lateral thinking
The importance of setting objectives = structure
Research Methods Group
The objectives drive the rest of the study.
• your hypotheses• the kind of study that needs to be
done (experiment, survey, ....)• the treatments that have to be
compared• the population to be sampled• the data to be collected• the formal statistical analysis
needed• ....
analyticor public phase
Analytic or public phase
Research Methods Group
Ability to extrapolate findings beyond a given location and to make some kind of prediction rather than just a description.
Communicate findings => ‘public’ phase
When? For ALL types of research!
• Biophysical, social, economic, policy…• Lab, field, landscape,…• Participatory,…• Strategic, applied, adaptive• …
• Maybe NOT for activities with local ‘development’ outcomes only
Research Methods Group
Part 2. The confusion
Research Methods Group
The confusion
Research Methods Group
• research questions• overall objectives• specific objectives• one hypothesis• several hypotheses• null-hypothesis• hypothesis testing
• and their order
Basically
Research Methods Group
Objectives– Why you are doing the work
• Hypotheses– Elements of theory you will test– A testable hunch
• (Research questions)– Overall statement of what you are going to
research
Part 3. ‘Good’ objectives
Research Methods Group
Objectives
• The overall aim • Must be…
– Clear – Complete– Relevant– Reasonable– Capable of being met by the activities (experiments
and surveys)– NOT ‘To compare treatments’ or ‘To compare
populations’…
Research Methods Group
Example Objective
"To evaluate 4 improved mango varieties".– Which varieties?– Under what conditions (location,
management) are they to be evaluated?– What criteria will be used to evaluate them? – Linked to what problem solving strategy? – What is the science behind this?
Research Methods Group
"To determine whether the survival and growth rates, during the first two years, of improved mango varieties (Kent, Van Dyke, Tommy Atkins and Sensation) can be brought up to the level of the local variety by using a higher level of management ( addition of manure and water), when these trees are grown in crop fields of the coffee zone of Embu, Kenya"
Better?
Research Methods Group
A characterisation of the research process
observe
reflect
generalise
plan
Question theory
Formulate modifications
Design studies
Collect data
Interpret
Update theory
Research Methods Group
Hypotheses
• Modifications of, additions to, or new theory…
• …that generate alternative predictions…
• …that can be compared with data
• Usually stated as a ‘rule’ or a cause-effect relationship.
Research Methods Group
Example
Addition of manure and water will increase initial growth and survival of new mango varieties …
because…
they promote vigorous root growth.First part:
•Easy to test
•Not clear what it adds to understanding
Second part
•Harder to test
•Maybe adds much more to theory
Hypotheses should be…
• Useful– Fill a critical gap in the problem solving
process– Allow you to do something you cannot
currently do
Ocimum populations vary in the level of active ingredients
Variation in active ingredients in different Ocimum populations is large enough to exploit pharmacologically
OR
Highland populations of Ocimum have active ingredient profiles similar enough to coast populations that they may also be used as substitutes
Active ingredients in Ocimum are a plant response to stress so will be found in higher concentrations in more stressed environments
• Logical– Based on current understanding and theory
(social, economic as well as biophysical)– Avoid alchemy
Deep rooted legumes can increase soil N in sandy soils if the dry season water table is shallow.
Fallowing with Myfavourite treespecies will increase soil N on sandy soils.
Research Methods Group
• Testable– Logically– Logistically
Soils can regenerate after serious erosion if left undisturbed for long enough
Changes in soil fertility following forest clearance are caused by changes in soil bacteria populations.
Research Methods Group
• How many…? Where…? What is the difference…?– often local description only
• How…? Why…?– potential to
• generalise to other situations• contribute to the body of scientific knowledge• generate abstraction
Research Methods Group
Most useful hypotheses
• Include ‘How?’ or ‘Why?’
Adoptability of improved fallow varies by gender
Women find improved fallow less adoptable than men because of labour shortages, low priority for maize production and lack of information
Research Methods Group
Expressing the hypothesis
• Words– may get long winded
• Simple models
Soil CPlant C
litter
roots
decomposition
erosion
Research Methods Group
• Empty tables or graphs
Forest Annual monocrops
Soil quality
Time since conversion
Research Methods Group
Time since conversion
Forest
Annual monocrops
Research Methods Group
Part 4. The pitfalls
Research Methods Group
Common problems
• Status of hypothesis already known– researcher not sufficiently familiar with current theory– hypothesis is trivial (‘Species differ…’)
• Study is just a realisation of a phenomenon in a new environment – without hypotheses of why it may be different in that new
environment
• Hypothesis does not take problem forward– Apply the ‘So what?’ test.
• Hypothesis stated as statistical null hypothesis– does not direct design or test– can always be demonstrated with a poor study– often known to be false
Research Methods Group
Common problems
• ‘I am doing an exploratory study. I don’t what to pre-empt results or restrict what I might find by writing hypotheses’– …but then you don’t even know what to
measure!– Hypotheses are there but vague and unstated– Writing them down, however complex, can
only help.
Research Methods Group
In practice• Expect to cycle through problem-objectives-hypotheses several times until it all makes sense• Do ‘thought experiments’. Write down (in tables or graphs) the
sort of results which would confirm or refute your hypotheses• Share your problem-objectives-hypotheses statements with as
many people who may give constructive help as possible.• Don’t be afraid to state hypotheses that challenge common
views – but only if you present a strong case.• Recognise that the finished product of problem-objectives-
hypotheses-design-results-conclusion you read in a paper rarely started out that way as the research was conceived and designed
• Read books on scientific method
Research Methods Group
Part 5. The toolbox
Research Methods Group
Talk
• Feedback
• Constructive criticism
Take initiative.
Research Methods Group
Question
Decision Protocol 2.0 USDA Forest Service
Research Methods Group
Problem and goal tree
DFID. 2002. Tools for development handbook. A handbook for those engaged in development activity.
Research Methods Group
Advanced ideas
Research Methods Group
Underutilised idea:
• Pre-testing hypotheses with quantitative models
Research Methods Group
Outcome Impact on design
1. No relevant model Add model building to objectives
2. Hypothesis not confirmed by model
(a) model assumptions are unrealistic
Add model refinement to objectives
(b) model assumptions are realistic
Examine your assumptions, modify them and produce revised hypothesis3. Hypothesis
confirmed by the model.
Hypothesis is a consequence of current knowledge. There may be more useful objectives for field trials.
Read
Chapter 5 and 6
Research Methods Group
Read
Chapter 3.2
Research Methods Group
Read
Ford, E. David (2000) Scientific Method for Ecological Research.
Cambridge University Press. 564 pp.
ISBN 0 521 66973 1
Research Methods Group
Peters, Robert Henry (1991) A critique for ecology. Cambridge University Press, Cambridge, UK. 366 pp. ISBN 0 521 39588 7