293
RESEARCH METHODOLOGY GK Mbassa

Research methodology lectures new prof.. mbassa

Embed Size (px)

Citation preview

Page 1: Research methodology lectures new prof.. mbassa

RESEARCH METHODOLOGY

GK Mbassa

Page 2: Research methodology lectures new prof.. mbassa

INTRODUCTION TO RESEARCH

Objective Introduces researchers and

students to scientific research methods, enable them prepare research proposals in:

Page 3: Research methodology lectures new prof.. mbassa

Veterinary Sciences Animal health Animal production Biotechnology Medicine Biomedical and Laboratory

sciences Agriculture Wildlife Others

Page 4: Research methodology lectures new prof.. mbassa

CONTENTS

1. Introduction; Research planning and process

2. Types of research 3. Problem identification

process 4. Literature review on subject

Page 5: Research methodology lectures new prof.. mbassa

5. Factor-Outcome relationship 6. Measurements 7. Research designs 8. Data collection 9. Data processing, analysis,

and management

Page 6: Research methodology lectures new prof.. mbassa

10. Data presentation 11. Research project

description 12. Report writing 13. Research ethics

Page 7: Research methodology lectures new prof.. mbassa

Introduction

Research is a systematic for search or inquiry for information (new information)

Research purpose is to explore, describe, explain and control

Page 8: Research methodology lectures new prof.. mbassa

Stages in Research

Planning stage Data collection (gathering

the information) Data analysis (processing

data to yield knowledge)

Page 9: Research methodology lectures new prof.. mbassa

Interpreting the data (extracting the knowledge and information)

Results utilization phase

Page 10: Research methodology lectures new prof.. mbassa

(a) Building the concept (b) Problem search (c) Research justification (d) General objectives

Planning stage

Page 11: Research methodology lectures new prof.. mbassa

(e) Specific objectives (f) Assumptions (g) Limitations (h) Hypotheses themes, arguments

(i) Operational concepts

Page 12: Research methodology lectures new prof.. mbassa

(j) Planning of research and purpose;

(k) Literature review(l) Proposal write up

Page 13: Research methodology lectures new prof.. mbassa

Data collection (gathering the information)

(a) Population source of data

(b) Logistics of data collection

(c) Collection of samples from the population

Page 14: Research methodology lectures new prof.. mbassa

Data analysis (processing data to yield knowledge)

Facilities for data analysis Laboratory procedures Treated samples Control samples Recording of results Statistical procedures

Page 15: Research methodology lectures new prof.. mbassa

Interpreting the data (extracting the knowledge and information)

Data grouping and splicing Tables and figures Means and trends; Equivocal and unequivocal

conclusions

Page 16: Research methodology lectures new prof.. mbassa

Results utilization Identify beneficiaries of

results (solved problem, generated technology)

Professional Research Report Scientific briefs Communications of

knowledge and technology

Page 17: Research methodology lectures new prof.. mbassa

Seminars and workshops Policy changes Further research or

activities Publication of data Patent technology Apply/ Sell technology Sales of research products

and technology

Page 18: Research methodology lectures new prof.. mbassa
Page 19: Research methodology lectures new prof.. mbassa

Variables in Research

Variables are factors, parameters, attributes or qualities of the cases that are being measured or recorded, examples being sex, age, height, weight, colour, number etc and are either independent or dependent.

Page 20: Research methodology lectures new prof.. mbassa

Variables vary in their scores on the different attributes, observations, records or population numbers.

Page 21: Research methodology lectures new prof.. mbassa

Independent variables, also called predictor or explanatory variables are the factors that cause the variation in the dependent variable

Dependent variable is the outcome resulting from the independent variable

Page 22: Research methodology lectures new prof.. mbassa

TYPES OF RESEARCH

Several categories On basis of numerical

principles1. Qualitative2. Quantitative3. Both qualitative and

quantitative

Page 23: Research methodology lectures new prof.. mbassa

Qualitative; describes/analyzes culture and behaviour of humans, animals, plants, materials, cells, flowers, fruits, organic or inorganic matter, for example staining characteristics, organoleptic tests of foods such as taste, smell, colour, consistence, interactive, opinions, feelings. In social sciences qualitative research is called naturalistic inquiry or field studies.

Page 24: Research methodology lectures new prof.. mbassa

Qualitative research Ethnographies (observations of

groups); Phenomenologies, studying

subjects over a period of time; Case studies to investigate subject

over time

Page 25: Research methodology lectures new prof.. mbassa

Quantitative Research Uses numeric data to verify,

confirm, prove causation, correlation, corroboration or substantiation

Establishes cause-effect relationship, focusing on measurements, assigning of numerical events according to rules

Page 26: Research methodology lectures new prof.. mbassa

Application of quantitative research

Requirement for statistical tests Quantify extent of cause of effect Frequencies needed to explain

meanings, collects numerical data to explain phenomena;

Discovery of unexpected and in-depth investigation

Page 27: Research methodology lectures new prof.. mbassa

In finding effect, control of one or more factors required. Rigorous and rigid methodologies and all procedures are specified before data collection and consistently followed

Data analysis is statistical (deductive)

Scenario is artificial, as in a laboratory.

Page 28: Research methodology lectures new prof.. mbassa

Qualitative and quantitative research complement each other

Combined in biological systems to maximize strengths and minimize limitations of each.

Page 29: Research methodology lectures new prof.. mbassa

Other Research types criteria Basic sciences: —study of macro

and micro morphology & physiology

Applied sciences; —apply knowledge to develop drugs, vaccines, seeds, molecules, genes and more;

Observational (descriptive)- observe activities in natural or artificial systems and reports the findings by description

Page 30: Research methodology lectures new prof.. mbassa

Analytical: — test causal association between factor and outcome

Evaluative studies:—find out whether some factor introduced in a population has imparted any influence (retrospective studies)

Innovative: —new inventions; technology, materials and testing alternative technologies

Experimental (research designed to test hypothesis)

Page 31: Research methodology lectures new prof.. mbassa

Surveys; prevalence of diseases, distribution of a factor. A survey to screen and detect existence of a factor in a population, and determine (or not), the magnitude of the factor

Page 32: Research methodology lectures new prof.. mbassa

RESEARCH PROBLEM IDENTIFICATION

Purpose of research; solve problem Identification is essential Develop ideas from your subject Purpose new knowledge solve

problem Know client of the research results

Page 33: Research methodology lectures new prof.. mbassa

Characteristics of a good research question

Very clear; Specific; Elements, variables &

factors affecting variables definable and measurable;

Investigations on the question achievable within given time

Page 34: Research methodology lectures new prof.. mbassa

Clear relationship between research question and hypothesis;

Research question fitted into hypothesis and statistical tests;

Research question methods and assumptions well definable

Page 35: Research methodology lectures new prof.. mbassa

Source of data well definable; primary, secondary, routine or published data

Factors affecting research process defined; logistics and finances

Research on problem and question acceptable scientifically, socially, politically and budgetary

Research question has not been previously answered

Page 36: Research methodology lectures new prof.. mbassa

A researchable problem is found through:

A directly observed problem or puzzle in science, social, economic, cultural, development, political or other systems, animals, plants, materials, foods, diseases, machines, education, health, weather and many others;

Page 37: Research methodology lectures new prof.. mbassa

A thorough review of literature on the subject and on closely related subjects to find out what exists, what has already been done and what the exact problem is and how it can be solved in terms of experimental design, materials and methods

Page 38: Research methodology lectures new prof.. mbassa

Improvement of existing systems, machines, life of organisms, foods (palatability, nutritive value, contents, effects), forests, domestic animals, wildlife, cells, drugs, vaccines and other biological products

Page 39: Research methodology lectures new prof.. mbassa

To evaluate effectiveness of systems, materials, vaccines, drugs, shelf life of goods, animal/human nutrition, knowledge delivery in teaching, teaching aids, and others

Page 40: Research methodology lectures new prof.. mbassa

Once defined, decide type, kind of data to be collected and experimental designs.

Research is preceded by preparation of comprehensive research proposal to guide the research process.

Page 41: Research methodology lectures new prof.. mbassa

Research proposal is a step by step manual of all activities

It follows critical path of events achievable in sequences

Sound and effective research proposal is completed after thorough literature review of subject

Page 42: Research methodology lectures new prof.. mbassa

When elements, variables and factors are known, plan and type of research are decided based on nature of problem or question to be answered

Page 43: Research methodology lectures new prof.. mbassa

LITERATURE REVIEW ON SUBJECT OF

INVESTIGATION

Literature means published knowledge stored in any retrieval system, books, journals, periodicals, newsletters, microfilms, films, music, video and others.

Page 44: Research methodology lectures new prof.. mbassa

Literature review means reading, extensively with a purpose of updating knowledge on specific subject and keep list of titles of published material

Page 45: Research methodology lectures new prof.. mbassa

Bibliography; collection of titles of published papers and books on a subject indicating source of paper, no abstracts

Example all publications on mitochondria make bibliography of the mitochondria

Page 46: Research methodology lectures new prof.. mbassa

Literature review is analysis of current state of knowledge

Helps research proposal to state clearly what will be known after research, that is not known now

Summarizes current state of knowledge, giving up-to-date bibliography

Page 47: Research methodology lectures new prof.. mbassa

Purposes (aims) of reviewing literature

Determine relevant literature to study

Gain information on subject to current

level

Reveal investigations related to the

proposed research

Page 48: Research methodology lectures new prof.. mbassa

Show how other researchers did on similar problems,

Obtain a method or technique of

dealing with the problem

Determine valid approach to the

problem,Reveal other

sources of data,Obtain new ideas and approaches

Form research in historical &

associational perspective and in relation to earlier

attempts in solving similar problem

Page 49: Research methodology lectures new prof.. mbassa

Determine terminologies, parameters &

variables & their measurements

Be able to define general research questions within

current knowledge of literature

Be able to specify question in current

literature,Update knowledge to locate gap for

research

Page 50: Research methodology lectures new prof.. mbassa

Places of literature storage Library of Universities and

Institutions; Computers on world wide web

(internet); National libraries and bibliographic

centers; National Bureau of Statistics; National Archives; Ministries; United Nations Organization offices

Page 51: Research methodology lectures new prof.. mbassa

Two categories of sources of knowledge

Published Non published

Page 52: Research methodology lectures new prof.. mbassa

Published Books Journals Periodical publications Annual subject reviews Proceedings of conferences,

symposia and professional society meetings

Page 53: Research methodology lectures new prof.. mbassa

Non-published Ph.D. theses MSc. Dissertations Various reports Office documents Project reports Special collections and even

Minutes of meetings

Page 54: Research methodology lectures new prof.. mbassa

Review summarizes literature precisely and succinctlyGathers specific knowledge to which the research will add

Reflect on review of related literature, what others have written in relation to what is planned

Review literature from a comprehensive perspective, like an inverted pyramid, broad end first

Page 55: Research methodology lectures new prof.. mbassa

Constantly explain relatedness of proposed research to one in literatureDefine gap where new research will fill

Rewrite content of literature sources in own words and style, not copying texts of other authors

Read source, understand it, list points you remember, re-check points and join them into proper sentences

Page 56: Research methodology lectures new prof.. mbassa

Outcomes of literature review;- Framed hypothesis within current

literature Defined scope of research and

objectives Defined type or category of

research within current literature Defined variables in the research

within current literature

Page 57: Research methodology lectures new prof.. mbassa

Clear research plan Identified source of data (Research

sites) Identified action plan, logistics to

the research Ascertained objectives and

purpose of research Sequencing the activities Clear implementation strategy,

variables, parameters, factors

Page 58: Research methodology lectures new prof.. mbassa

Conclude literature review by giving specific objectives of what the research is going to achieve

Page 59: Research methodology lectures new prof.. mbassa

All the published materials must be indicated at two places in the researchers project plan / or text

In the text In the list of references

Page 60: Research methodology lectures new prof.. mbassa

Citing literature in the text

Literature is cited in text to indicate source of scientific findings being quoted from retrievable sources

1. Normally journals 2. Sometimes books 3. May quote a review paper

Page 61: Research methodology lectures new prof.. mbassa

References are required for

1. 0riginal findings 2. Knowledge established by previous researchers

3. Other information needing to indicate source

Page 62: Research methodology lectures new prof.. mbassa

a) Articles with single author

(1) Subject comes before quotation

(i) Name of author placed at end of subject of the sentence

(ii) Author's name followed by coma (,), the year of publication

(iii) Author's name & year of publication placed in brackets (parentheses)

Page 63: Research methodology lectures new prof.. mbassa

Examples Biotechnology is a primary tool of

control of foot and mouth diseases in cattle (Thompsen, 1995).

African goats produce more milk than Toggenburg breeds (Smith, 1989)

Lymphocytes cultured in vitro in RPMI 1640 medium proliferate within minutes (Belinger, 2000)

Total serum protein in African buffaloes lies within 100 ‑140 g/l limits (Mbassa, 1990)

Page 64: Research methodology lectures new prof.. mbassa

(2) Author before subject (i) Name of author placed at

beginning or middle of subject of sentence

(ii) Author's name not followed by comma but year of publication

(iii) Only year is in parentheses

Page 65: Research methodology lectures new prof.. mbassa

Examples Thompsen (2006) observed

that the native dense core protein of Babesia bovis has several epitopes, thus it is a candidate vaccine

Smith (2007) concludes that African goat milk contains an antiallergic factor

Page 66: Research methodology lectures new prof.. mbassa

According to Belinger (2006) lymphocytes cultured in vitro in RPMI1640 medium proliferate within minutes

Mbassa (2006) observed the total serum protein in African buffaloes to be 100‑140 g/l

Page 67: Research methodology lectures new prof.. mbassa

b) Articles with two authors

(i) Names of authors in order they appear in article placed at end of subject

(ii) Name of last author followed by coma (,) and year

(iii) Names of authors and year in parentheses

Page 68: Research methodology lectures new prof.. mbassa

ExamplesThe butter‑fat content of zebu

cattle milk is 5.2 % (Mtenga and Aboud, 2004)

Endemic stability in East Coast fever is well established in Lake Victoria Region (Mpangala and Mollel, 2005)

Page 69: Research methodology lectures new prof.. mbassa

Pars intermedia is not found in the pituitary gland of the greater flamingo, Phoenicopterus rubber rouseus (Mhowa and Dominico, 2007)

Page 70: Research methodology lectures new prof.. mbassa

Authors before subject of sentence

(i) Names of authors as in article placed before subject

(ii) Name of last author not followed by coma, but year of publication

(iii) Only year of publication in parentheses

Page 71: Research methodology lectures new prof.. mbassa

Examples Mtenga and Aboud (2004) found the

butter-fat content of zebu cattle milk to be 5.2 %

Mpangala and Mollel (2005) noted that endemic stability in East Coast fever is well established in Lake Victoria Region

Mhowa and Dominico (2007) did not find any pars intermedia in the pituitary gland of the greater flamingo, Phoenicopterus rubber rouseus

Page 72: Research methodology lectures new prof.. mbassa

(c) Articles with more than two authors

Mention first author, followed by words

"et al.,” in italics

Page 73: Research methodology lectures new prof.. mbassa

(i) Subject before authors Name of 1st author followed by

"et al.,“ placed after subject "et al ” in italics, then full stop

(.), coma (,) and year Name of author, "et al.,” and

year all placed in parentheses

Page 74: Research methodology lectures new prof.. mbassa

Examples Biotechnlogy is the primary tool in the

control of foot and mouth diseases in cattle (Thompsen et al., 2006)

African goats produce more milk than Toggenburg breeds (Smith et al., 2007)

Lymphocytes cultured in vitro in RPMI 1640 medium proliferate within minutes (Belinger et al., 2006)

Total serum proteins in African buffaloes lie within 100 ‑140 g/l limit (Mbassa et al., 2006)

Page 75: Research methodology lectures new prof.. mbassa

(ii) Subject after authors

Name of 1st author followed by "et al.” in italics placed after subject

"et al.“ not followed by coma (,) but year

Only year in parentheses

Page 76: Research methodology lectures new prof.. mbassa

Examples1. Thompsen et al. (2005) observed that

foot and mouth diseases is more severe in cattle than in goats

2. Smith et al. (2003) concludes that African goats produce more butterfat in milk than Toggenburg

3. According to Belinger et al. (2006) lymphocytes cultured in vitro, in RPMI 1640 medium proliferate within minutes

4. Mbassa et al. (2007) observed the total serum protein in African buffaloes to lie within 100 ‑140 g/l limits

Page 77: Research methodology lectures new prof.. mbassa

Compiling list of references

Gives detail of 1. Names of all authors

quoted 2. Only relevant references 3. References not cited, not to

appear 4. Names and, initials of all

authors said as "et al " in the text must appear in reference list

Page 78: Research methodology lectures new prof.. mbassa

Information for each author quoted

(i) Surname followed by coma (,)

(ii) Initials (eg. M. M.) (iii) Year article published (e.g.

2006) (iv) Full stop (.) (v) Title of article (e.g. " Peripolar

cells form the majority of granulated cells in kidneys of antelopes and goats”

full stop.

Page 79: Research methodology lectures new prof.. mbassa

(v) Full title (or std abbreviation of journal where article was published (e.g. Veterinary Parasitology" or Vet. Parasitol., Acta Anatomica or Acta Anat., Internat. J. Biotechnol.)

(vi) Volume of journal, colon (e.g. 51:) (vii) First page of article, dash

(e.g.3894- (viii) Last page of article (e.g. 3902) (ix) full stop.

Page 80: Research methodology lectures new prof.. mbassa

Unless specially required, issue numbers in same volume of journal are not shown

Page 81: Research methodology lectures new prof.. mbassa

Articles published in journals

Single author articles example

DeVos, A. J. 1978. Immunogenicity and pathogenicity of Babesia bovis in Bos indicus cattle. Onderstepoort J. Vet. Res. 45:119-123.

Page 82: Research methodology lectures new prof.. mbassa

Two authors articles, Example

Fu‑Chu, He and Ghu‑tse Wu 1993. Molecular evolution of Cytokines and receptors. Exp. Hematol. 21:521‑524.

Page 83: Research methodology lectures new prof.. mbassa

Articles with more than two authors

All authors, with initials must be provided

Word “et al ” not allowed to appear in list of references

Page 84: Research methodology lectures new prof.. mbassa

ExamplesBrown, W. C., V. Shkap, Damine Zhu, T. C. McGuire, W. T, T. F. McElwain and G. H. Palmer 1998. CD4+ T lymphocyte and immuno-globulin 2 responses in calves immunized with Anaplasma marginale outer membranes and protected against homologous challenge. Infect. Immun. 66:5406-5413

Page 85: Research methodology lectures new prof.. mbassa

Luziga, C., Yamamoto Y., Horii Y., Mbassa G. and Mamba K. 2006. Phagocytotic removal of apoptotic endocrine cells by folliculostellate cells and its functional implications in clusterin accumulation in pituitary colloids in helmeted guinea fowl (Numida meleagris). Acta Histochemica 108:69-80.

Page 86: Research methodology lectures new prof.. mbassa

Quoting in a book with single author

(i) Name of author with initials, year of publication

(ii) Title of book (iii) Pages (e.g. 601‑623). (iv) Publisher (e.g. Lea and

Febiger) (v) Place of publication (e.g. Dar es

Salaam, Philadelphia, Toronto)

Page 87: Research methodology lectures new prof.. mbassa

Example Klaus, G. G. B. 1987.

Lymphocytes: A practical approach. 261. IRL Press, Oxford England

Page 88: Research methodology lectures new prof.. mbassa

Quoting from book of two authors

Example Losos, G. J. and Brown J. M. 1996.

Infectious tropical diseases of domestic animals. Langman Scientific and Technical. 742-795, Washington D.C.

Page 89: Research methodology lectures new prof.. mbassa

Quoting from a book with many contributors

Example Keller, G. , K. Roger Tsang and I.

Kakoma. 1988. Advances in the in vitro cultivation of Babesia species. In: Babesiosis of domestic animals and man. Miodrag Ristic Eds, CRC Press Inc. Boca Raton Florida, 71-79.

Page 90: Research methodology lectures new prof.. mbassa

Quoting chapter in a book

(i) Name(s) of author(s) with initials, year of publication

(ii) Chapter title(iii) Followed by In: title of book

followed by Name(s) of author(s) of book with initials, Pages, Publisher, Place of publication.

Page 91: Research methodology lectures new prof.. mbassa

Examples Habour, C. and A. Fletcher 1991.

Hybridomas: Production and selection. In: Mammalian Cell Biotechnology: a practical approach, by M. Buttler, Oxford University Press, Oxford, 109‑138

Page 92: Research methodology lectures new prof.. mbassa

(ii) Mathew, P., S. Michael and Juma K. 2002. Modeling of changes in belief. In: The psycho‑technological development for inductive changes in pastoralism. First Edition, by Williams P. K and Manus S. P. 2002. Oxford University Press, Nairobi Kenya. pp 234‑313

Page 93: Research methodology lectures new prof.. mbassa

(iii) Hames, B. D. 1996. One dimensional polyacrilamide gel electrophoresis. In: Gel electrophoresis of proteins, Edited by B. D. Holmes and D. Rickwood. IRS Press Oxford. 1‑147.

Page 94: Research methodology lectures new prof.. mbassa

Quoting from book with collection of articles (Multi-

authored book)

(i) Name of author(s) of article in the book

(ii) Title of article, year of publication

(iii) Followed by In: (iv) Name(s) of book authors (s)

with initials, then in parentheses (Editors)

Page 95: Research methodology lectures new prof.. mbassa

(v) Title of book (vi) Publisher's name(s) (vii) Place of publication (viii) Pages read (first ‑ last)

Page 96: Research methodology lectures new prof.. mbassa

OR (i) Name of author(s) of article in

book (ii) title of article (ii) Year published (iii) Followed by In: (iv) Title of book, followed by

Edited by then Name(s) of Editor(s) with initials, and pages read first ‑ last

Page 97: Research methodology lectures new prof.. mbassa

Brown, L. R. and C. Flavin 1999. A new economy for a new century. In: L. R. Brown, C. Flavin and H. French (Editors), State of the World; A World-watch Institute Report on Progress Towards a Sustainable Society, 3‑21. W.W. Norton and Company, New York

Page 98: Research methodology lectures new prof.. mbassa

Alternative quote from a book with many articles (multi-authored book)

(i) Name of author(s) of article in the book

(ii) year (iii) title of paper, followed by, In..

title of book, then Edited by, name(s) of author(s) with initials, publisher, place of publication and pages being referred to first – last

Page 99: Research methodology lectures new prof.. mbassa

Example Brown, L. R. and C. Flavin 1999. A

new economy for a new century. In: State of the World; A World-watch Institute Report on Progress Towards a Sustainable Society, Edited by L. R. Brown, C. Flavin and H. French, 3‑21. W.W. Norton and Company, New York

Page 100: Research methodology lectures new prof.. mbassa

Quoting from conference proceedings, symposia

(i) Name(s) of author(s) with initials

(ii) Year (iii) Title of article (iv) Followed by words;

Proceedings of (names of conference) held on (date & year) in (city and country)

Page 101: Research methodology lectures new prof.. mbassa

(v) Name of Editor(s) with initials

(vi) Publishers (vii) Place of publication (viii) Volume of the proceeding (ix) Pages being read.

Page 102: Research methodology lectures new prof.. mbassa

Examples Mbassa G. K., Young S., Kauzeni P. and

Nkangaga J. J. 2003. Quantitative impact analysis of wildlife conservation strategies: a study of impact of community-based conservation in Gonabis wildlife management area and rare antelopes in Game Reserves. Proceedings of the Fourth Scientific conference of the Tanzania Wildlife Research Institute Held at Impala Hotel, Arusha Tanzania Dec 3 to 6, 2003: 50-65;

Page 103: Research methodology lectures new prof.. mbassa

Quoting from non regular publications

Workshops, Meetings, Other sources

(i) Name(s) of author(s) if possible (ii) Year of publication (iii) Publisher or custodian of work

Page 104: Research methodology lectures new prof.. mbassa

Examples Mgongo F.O.K., Mellau, L.S.B., Mbassa,

G.K., Silayo R.S., Kimbita, E.N., Hayghaimo, A.A., Mlangwa J.E.D., Mbiha E.R., and Gwakisa P.S. 2007. Improving cattle selection and reproduction in the traditional pastoral sector with efficient control of tick borne diseases. Published by PANTIL-SUA Project Programme for Agricultural and Natural Resources Transformation for Improved Livelihoods. 27 pp

Page 105: Research methodology lectures new prof.. mbassa

RESEARCH/ EXPERIMENTAL DESIGNS

An experiment is planned inquiry to get new facts, confirm hypotheses,

A trial to test validity of prior set hypothesis

Page 106: Research methodology lectures new prof.. mbassa

Factors in selecting experimental designs

1. Solve problem or answer an un-resolved question

2. Contribute new information, improve understanding of matter

3. Develop/Test technology 4. Searching the unknown

Page 107: Research methodology lectures new prof.. mbassa

In basic research new scientific information is generated

E.g. genome nucleotide sequence of animal/plant, gene e.g. fibronectin in liver cell, gene for resistance against disease in cattle, beans etc

Applied research generates, tests and validates technology

Page 108: Research methodology lectures new prof.. mbassa

Experimental design selected to be optimal to answer research problem

New material

Research type

Availability of adequate

samples

Number of factors being

analysed

Statistical methods

Page 109: Research methodology lectures new prof.. mbassa

Research types

Innovative, survey, analytical, experimentalProspecti

ve, retrospec

tive

Living or non living matter, social, species

Quantitative, qualitative

In vivo or in vitro macro or microscopic

Page 110: Research methodology lectures new prof.. mbassa

Categories of experiments

Preliminary

•Large number of treatments to gather information, with or without replication

Critical•Compare responses of variable to different treatments, adequate units to detect differences

Innovative

•Generate new products, test them for their uses, compare with those in use

Page 111: Research methodology lectures new prof.. mbassa

Planning the experiment

State research problem State problem in hypothesis form

List objectives

Page 112: Research methodology lectures new prof.. mbassa

Select experimental design

Describe materials and methods in detail, including statistical analysis

Filter results from inputs

Page 113: Research methodology lectures new prof.. mbassa
Page 114: Research methodology lectures new prof.. mbassa

Filter results from inputs

Collect results

3. Perform experiment

2. State objectives

and methods

1. State problem & hypothesis

Page 115: Research methodology lectures new prof.. mbassa

Qualities of well- designed experiments

Equality in

sample sizes and

member properti

es to eliminat

e systematic errors

Uniform treatme

ntAppropri

ate statistical design

Clear variables Proper plan for interpretation of results

Simplicity in

designProvide

necessary data

Page 116: Research methodology lectures new prof.. mbassa

Experimental design to envisage

Proper results interpretation,Estimate errorError control

Biological materials vary greatly, due to inherent variability

(individual, intrinsic) and lack of uniformity in dispensing treatments

in the experiment

Page 117: Research methodology lectures new prof.. mbassa

Strategies to control errors

Replication in space (location) Increase sample sizes Replication in time Randomization Inclusion of controls Blocking certain natural variations Refinement of methods and

chemicals Minimize mechanical errors

Page 118: Research methodology lectures new prof.. mbassa

Control of errors

Replication in space (location), same time

Page 119: Research methodology lectures new prof.. mbassa

Replication in experimental units at the same time

Page 120: Research methodology lectures new prof.. mbassa

Replication in time, repeat experiment several times

Page 121: Research methodology lectures new prof.. mbassa

Randomization, select samples by random principles

Page 122: Research methodology lectures new prof.. mbassa

Inclusion of controls, omitting certain factors so that only a single factor is allowed to act on selected groups of treatments

Page 123: Research methodology lectures new prof.. mbassa

Blocking certain natural variations, all samples uniform in size, age, material, time of treatment and other factors

Page 124: Research methodology lectures new prof.. mbassa

Refine methods and chemicals, use very refined materials to eliminate blocking of active ingredients by impurities

HH

H

H

H O

O

OH

Page 125: Research methodology lectures new prof.. mbassa

Minimize mechanical errors, investigator and experiment dispenser to be perfect

Page 126: Research methodology lectures new prof.. mbassa

Common experimental designs

1. Single factor experimental design

Single factor varies, others constant Treatments; different levels of same

factor, e.g. testing suitable dosage for a growth factor, therapeutic drug, animal feed, mitogenic factor, mutagenic factor

Page 127: Research methodology lectures new prof.. mbassa
Page 128: Research methodology lectures new prof.. mbassa

Drug dosage studies in guinea pigs

Page 129: Research methodology lectures new prof.. mbassa

Results Results Results Results

Report

Page 130: Research methodology lectures new prof.. mbassa

Report

Publish

Patent

Sell technol

ogy

Disseminate

technology

Page 131: Research methodology lectures new prof.. mbassa

2. Experimental design for two or more factors

These are factorial experiments

Effects of many different factors are

investigated simultaneously, economic & time

serving

A factor is a treatment, consisting of all

possible combinations

Page 132: Research methodology lectures new prof.. mbassa

Magnitudes of effects are measured to determine contribution of each factor to effect

Factorial design reduces costs in animals, chemicals, feed, space,

time, labour, drug.

Example two or more milk replacers differing in composition investigated

in same group of animals at the same time in the same experiment

Page 133: Research methodology lectures new prof.. mbassa

(a). Completely randomized

Animal 1 Animal 4

Animal 7Animal 8

Animal 5 Animal 2

Animal 3Animal 6

Animal 9

Page 134: Research methodology lectures new prof.. mbassa

Uniform animals

(homogenous

Uniform treatments

Random selection at

all levels

Statistical methods T test, X square, GLM,

ANOVA

Page 135: Research methodology lectures new prof.. mbassa

(b) Completely randomized block designanimals in blocks can be replaced by farm plots, cells, etc

Animal 19

Animal 22

Animal 25

Animal 26

Animal 23

Animal 20

Animal 21

Animal 24

Animal 27

Animal 1

Animal 4

Animal 7

Animal 8

Animal 5

Animal 2

Animal 3

Animal 6

Animal 9

Animal 10

Animal 13

Animal 16

Animal 17

Animal 14

Animal 11

Animal 12

Animal 15

Animal 18

Animal 28

Animal 31

Animal 34

Animal 35

Animal 32

Animal 29

Animal 30

Animal 33

Animal 36

Page 136: Research methodology lectures new prof.. mbassa

Uniform study subjects

(homogenous)

Uniform treatments

Random selection at all levels

Measure effect of more than one

factor in an experiment

Economic and time serving

Variables include metabolic product,

growth level, disease occurrence,

immunity level

Page 137: Research methodology lectures new prof.. mbassa

Animals grouped as natural as possible & as homogenous as possible

to eliminate errors (homogenous units called blocks or herds, pens etc)

Then apply randomised

treatments for the blocks

Statistical methods are

ANOVA, GLM, T test, X square

Page 138: Research methodology lectures new prof.. mbassa

Analysis is performed by model; Yijk = +Ai + Bj + (AB)ij + eijk

Where Yijk = observation on kth animal receiving ith level of first factor A, = general mean common to all observations, Ai = an effect of ith level factor, first factor, Bj = an effect of jth level of second factor B, (AB)ij = an interaction effect due to combination between ith level of factor A and jth level of factor B and eijk = a random effect specific to each animal

Page 139: Research methodology lectures new prof.. mbassa

Report

Publish Patent

Sell technolog

y

Disseminate

technology

Page 140: Research methodology lectures new prof.. mbassa

3. Latin square experimental design when:-

Number of units very few

Limited facilities

Time consuming measurements

Same sample sizes of units at different

periods in sequence

Page 141: Research methodology lectures new prof.. mbassa

One/more animal per treatmentStatistics; sum of squares

Too few samples reduces power of experiment

Carry over effects on using same animals causes confounding

Page 142: Research methodology lectures new prof.. mbassa

4. Quasi-experiments

Design has some but not all of characteristics of true experiment

No random assignment of subjects to control and experimental conditions

Natural experiments, nature assigns subjects to conditions, e.g. trends in rainfall, wind, hurricane, crimes etc.

Page 143: Research methodology lectures new prof.. mbassa

Characteristics

Matching instead of randomization, similar locations, not control but as comparison, also called nonequivalent group design

Time series analysis (longitudinal study over time), impact analysis

Page 144: Research methodology lectures new prof.. mbassa

•not humans, contextual concepts; sociology, quality of life, anomalies, organization, disorganization, morale, climate, atmosphere

Unit of analysis

•Sufficient number of events to control threats to validity and reliability, independent variable is time

Validity

•Quasi-experiments are creative on causes of events, no control of independent variable•Baseline and natural interventions (legislation, relocation)

Control

Page 145: Research methodology lectures new prof.. mbassa

In quasi-experiments use trend, not cause, major ones are syndromes or cycles, minor ones are normal or abnormal events

Quasi-experiment research designs to involve many different, but interrelated variables, causal relationships can be modeled to identify spurious (false), intervening and suppressing variables

Page 146: Research methodology lectures new prof.. mbassa

Statistical methods in data analysis

Statistics is an applied mathematical science that provides an objective basis for analysing problems where the cause and effect of observations are not apparent

Page 147: Research methodology lectures new prof.. mbassa

There are many statistical methods in biological research

For living organisms it is biometry Social sciences (non-parametric

statistics) Living things include their cellular parts,

organelles, molecules in cells (amino acids, vitamins, proteins, minerals, carbohydrates, lipids, water, nucleic acids and their combinations).

Page 148: Research methodology lectures new prof.. mbassa

Statistical methods require 1. Quantitative measurements of

causal and effect factors, called variables (variates)

2. Populations 3. Samples 4. Measures of central tendency 5. Measures of dispersion and

hypotheses testing

Page 149: Research methodology lectures new prof.. mbassa

Variables include plant number, height, pods, weight, milk quantity, protein content, number of genes, number of genetic recombinations on chromosomes, many others

Qualitative and quantitative measurements constitute data

Qualitative variables are non-numeric for example colour, taste, smell, molecular reaction in a cell, stain uptake and others

Page 150: Research methodology lectures new prof.. mbassa

Data may be continuous, taking any value on a continuous scale (for example weight, height, milk yield)

Or discontinuous (discrete) appearing only as integral values eg, animal (cannot have half animal).

Page 151: Research methodology lectures new prof.. mbassa

Populations A population (a universe)

means all individuals of a particular experimental set; animals, plants, cells, molecules and others) and includes all possible values of measurements, ranges from small to infinite

Page 152: Research methodology lectures new prof.. mbassa

Samples A sample is a selection from

where observations, information or variables are recorded

The results of measurements in a sample can be extrapolated to the population

To be truly representative of the population, it must have been obtained by random selection or by purposive sampling

Page 153: Research methodology lectures new prof.. mbassa

In a random sample each member of the population has equal and independent chances of being included.

The sample constitutes an experimental unit

Treatment in experiment is a procedure applied on a member of the sample, e.g. drug, milk, growth factor

Page 154: Research methodology lectures new prof.. mbassa

Measures of central tendency (location)

Average (arithmetic mean) Statistical mean (central location

or ordinary value) Median (central number when

observations are arranged in ascending or descending order)

Mode (most frequently occurring observation value)

Page 155: Research methodology lectures new prof.. mbassa

Measures of dispersion (Variability), first, second, third and fourth moments about the mean

First moment gives mean deviation about the mean

Second moment about the mean gives variance and standard deviation its square root

Page 156: Research methodology lectures new prof.. mbassa

Third & fourth moments about the mean measure degree of skewness and curtosis of frequency distribution about the mean

Range is the difference between the highest and lowest values

Page 157: Research methodology lectures new prof.. mbassa

Coefficient of variation (CV) or coefficient of variability (CV) measures precision of measurements

CV (%) = (std x 100) divided by the mean

Small CV indicates greater precision than in large CV, thus results are more reliable

Large CV indicate increased experimental errors

Page 158: Research methodology lectures new prof.. mbassa

Hypothesis testing is based on statistical decision

A hypothesis is logical assumption about characteristics of a population.

The researcher guesses about the results of experiment, the guess is the hypothesis, which is tested

Page 159: Research methodology lectures new prof.. mbassa

Examples a new gene may treat a disease a new drug is more effective than

standard drug a formulated material treats this

disease a new compound causes higher

plant growth or animal growth

Page 160: Research methodology lectures new prof.. mbassa

a new breed of plant or animal yields more product

a new breed of animal is resistant to disease

a gene knockout results in abolition of disease, fertility, transcription, mutation, and many other examples of hypothesis

Page 161: Research methodology lectures new prof.. mbassa

Statistical tests include T-test; 2 means tested, two way test Analysis of variance (ANOVA), where

more than two means are being tested, sample sizes in groups uniform

General linear models (GLM), more than two means tested, samples sizes in groups different

Page 162: Research methodology lectures new prof.. mbassa

Chi (X) square, tests if one method is better than the other one or two way

Many others (parametric, non-parametric)

Page 163: Research methodology lectures new prof.. mbassa

ANOVA partitions total variation into different sources, for example

among experimental units treated differently

among experimental units treated similarly

due to non-experimental variables

various interactions

Page 164: Research methodology lectures new prof.. mbassa

ANOVA works if the sample sizes in different treatments are equal.

If the sample sizes are different another similar but stringent test similar to T –test is used, called General linear models (GLM)

Page 165: Research methodology lectures new prof.. mbassa

Other statistical analytical methods

Regression Correlation Wilcoxon Shapiro-Wilk statistic Duncan’s multiple range

Page 166: Research methodology lectures new prof.. mbassa

RESEARCH PROPOSAL

A research proposal is an idea intended or advised on a researchable subject, put

forward or suggested systematically and complete

Research Pro posals are suggestions, intentions, plans, schemes or requests

Systematic planned design of protocol to conduct research in specific subject

Page 167: Research methodology lectures new prof.. mbassa

Objective •Provides objectivity & critical insight of planned research

Manual •Advanced manual to be followed in research process

Search •Defines information being sort

Page 168: Research methodology lectures new prof.. mbassa

What • Sets out problem to be researched• What information being sort

Why •Why this is important to be known

How •How it is to be extracted from population

Where and when •When and where will the research be done

Page 169: Research methodology lectures new prof.. mbassa

Addresses subject from

known towards

unknown, superficial to

detailed, general to

specificStates

objectives to achieve

Relates to collateral or related studies of

other researcher

s

Proposes data

necessary for solving problem

indicated, how to

collect data, treat,

process interpret

Determines needs to success

Page 170: Research methodology lectures new prof.. mbassa

Indicates how results

be presented,

used, published

and disseminated

;Gives total budget for completion of research

Research is to provide

new knowledge & technology

Good research requires

early thinking on subject &

collection of adequate

recent literature

Research proposal is divided into three or more

components

Page 171: Research methodology lectures new prof.. mbassa

Components of Research Proposal

TitleResearcher

sSummary

or Concept

Introduction

(background

information, problem

statement, justification

)Goal (aim)

Purpose and objectives,Hypothesis,

conceptual or theoretical frameworkLiterature review of subject

Page 172: Research methodology lectures new prof.. mbassa

Research methodolo

gy materials methods

data analysis

work plansExpected outputs

BudgetResults

dissemination

Literature cited

(References)

Logic framewor

k of goals,

purpose, outputs,

objectives &

activitiesAppendic

es

Page 173: Research methodology lectures new prof.. mbassa

Title

ShortClear

Reflects content of research

Describes content of research

Leads to understand concepts,

methods & output of proposed research

Page 174: Research methodology lectures new prof.. mbassa

Researchers

Researchers involved

named

Their qualifications and addresses mentioned

Their affiliations and telephone numbers

given

Page 175: Research methodology lectures new prof.. mbassa

Summary/concept

No more than 250 words on;

What is to be done

Why should it be done (problem)

How it will be done (what samples)

Where and when research is to be done

Who are the beneficiaries

Page 176: Research methodology lectures new prof.. mbassa

Selecting topic or subject of research

Identify interests or puzzles

Identify puzzling points; scientific, social,

economic, health, political, cultural

Identify keywords on topic

Express puzzle in specific keywordsDefine topic by

analyzing keywords

Formulate topic by searching for articles to

identify researchable problem

Page 177: Research methodology lectures new prof.. mbassa

Qualities of a good research topic

Researchable, instruments are easily formulated, population,

samples, objectives, measurable

Contributes new knowledge

Findings publishable

Provocative, open to varied views and interpretations;

Clear and focused, not vague or ambiguous

Page 178: Research methodology lectures new prof.. mbassa

Too wide subject, no limit of scope

Vague topic not possible for in-

depth studyToo complex study

subject

Poor timingLimited accessibility

to materials

Problems in topic selection

Page 179: Research methodology lectures new prof.. mbassa

Selection of Title

Title is heading, label

or tag, describes

study, mini abstract, portrays

summary of key idea(s)

Discusses topical issues in science, business, life and living or others,

analyzing factors enhancing or hindering success of generations

Formulated after

identification of research

topic (subject)

Page 180: Research methodology lectures new prof.. mbassa

Identify title keywordsReflect on key issues

Identify independent and dependent variables, link them in title

Evaluate title; clear, specific, independent & dependent variables identified

Control length to 12-15 words only

Steps in title selection

Page 181: Research methodology lectures new prof.. mbassa

Qualities of good and effective research proposal

title

Brief and specificSalient & have strong impact

Easier to see independent &

dependent variables

In line with objectives

FocusedSummary of what

study is about

Page 182: Research methodology lectures new prof.. mbassa

Portray aims of study

Reflect relationship between

independent & dependent variables

Show researchable subject with measurable

results

Unambiguous, not to cause

various interpretations

of the study

Page 183: Research methodology lectures new prof.. mbassa

Challenges in title selection

Not specific, varied

interpretations

Too long and too wordy

Difficult to understand

Lack of consistency, objectives not apparent

or different from problem statement &

or methodology

Page 184: Research methodology lectures new prof.. mbassa

IntroductionOpens study

Discusses background to research

States & defines problemAims & objectives stated & how

work will progress givenEstablishes existence of problem &

need/justifies investigation

Page 185: Research methodology lectures new prof.. mbassa

Sub-divisions of introduction

.

Background knowledge

Statement of problem

Justification

Aims & objectives, hypothesis,

research questions

Significance of study,

limitations, conceptual &

theoretic frameworks

Page 186: Research methodology lectures new prof.. mbassa

Background information

Knowledge on the problem in

context of current

literature & status

Gives setting of study

Determines user/client or beneficiaryExplains the

matter, what is known about the subject

Scientific, establish a cause and

effect relationshipPrerequisite knowledge

before problem is stated

Page 187: Research methodology lectures new prof.. mbassa

Background information reveals

What has brought about need for study;

Challenges faced due to the issue

Problem or opportunity exists & needs being

addressed;Opportunities for

improvement

Current view of research problem;

Familiarity with subject with clear linkage of flow of knowledge

Page 188: Research methodology lectures new prof.. mbassa

Qualities of good background information

Brief, specific, summary literature review;

Generates concerns on problem &

opportunity of solution

Gives a glimpse of the

research problem;

Gives idea on how the

proposal is structured

Uses simple, straightforwar

d fluent language;

Informative, persuasive,

states urgency of addressing problem so

resources be allocated

Page 189: Research methodology lectures new prof.. mbassa
Page 190: Research methodology lectures new prof.. mbassa

Challenges in writing background information (BI)

Confusion between BI & literature review

(LR). LR studies related areas, BI is short,

briefly on why to study & opportunities after

Confusion of BI with justification of study, but BI is to give brief overview of problem

Bad quotations

Lack of clarity, jargon, slang, trendy words,

abbreviations, colloquial, redundant phrases, confusing

language

Page 191: Research methodology lectures new prof.. mbassa

Qualities of good backgrounds

Brief, specific overview of problem

Simple straightforward language

Previous studies justifying study are

cited

Researchers show familiarity with current events and information

on problem

Page 192: Research methodology lectures new prof.. mbassa

Statement of problem

Identified problem is reason for research

Problem identification may involve

exploratory research (diagnostic survey) to

collect background data

One way; use logical

approach to establish causation

Problems have logical

sub‑components

(sub‑problems)

When solved

separately, sub‑problems resolve

main problem

Page 193: Research methodology lectures new prof.. mbassa

Sub‑problems are researchable units, making interpretation of

data more apparent, adding up to totality of problem

Research problem must be specific to make methods specific & appropriate, set precise limits

of problem area

Page 194: Research methodology lectures new prof.. mbassa

Qualities of sound & good statement of problem

Clear & conciseHas an impact on the whole topic

Indicates urgency of research & that

research is definitely needed

Problem researchable

through collection and analysis of

data

Page 195: Research methodology lectures new prof.. mbassa

Steps to write research problem

Reflection—start with idea, what

kind of question is to be answered

Present research ideas or puzzles,

then assess selected topic &

title

Reflect topic, independent &

dependent variables of investigation

Page 196: Research methodology lectures new prof.. mbassa

Identify uncertaintie

sState why is a problem,

how will communities be better

off after research

Formulation; when

problem is identified

formulate it clearly. Indicate

how it came out,

personal observation or previous

research

Justification - explain

repercussions to follow if problem is not addressed, use statement to show that research has to be performed

Page 197: Research methodology lectures new prof.. mbassa

Challenges in research problem formulation

Clarity• Research problem is lack of

clarity

Unity

• lacks unity and relationship to objectives, independent and dependent variables & literature review

Urgency

• Lacks urgency, no urgency for investigation, no evidence that if not addressed, repercussions are serious for country, people, lost opportunities

Emotion

• Statements lack objectivity, reinforce emotions over topic, problem not easily investigated by collection and analysis of data.

Page 198: Research methodology lectures new prof.. mbassa

Goals (Aims) & Objectives

Objectives are aims the research envisages to achieve, eventually the purpose of study

In any research there are indicators of

intention & direction of study

Goals or aims are intentions, goals or

what research strives to achieve, long term

objectives e.g. national development

Page 199: Research methodology lectures new prof.. mbassa

Aim is general statement

• Reflects intention or purpose of research, stated in terms not easily measurable

Aims assist in formulation of objectives

• Pinpoints purpose of study, reveals whether research is urgent or not

Page 200: Research methodology lectures new prof.. mbassa

Quality aims & goals of research proposal

Pragmatic;State purpose of

study, not referring to specific

achievements

State accomplishment of

group not individuals

Stated in general terms providing

direction for research

development

Broad enough to lead to specific

objectivesClearly stated & are

reflective

Page 201: Research methodology lectures new prof.. mbassa

Formulation of study aims, goals & purpose

Reflection – think, decide on what to

accomplish by end of the study, analyze title

Formulation – write purpose of study (what

to accomplish within time)

Analysis – analyze aims to confirm they

address research problem & questions

Page 202: Research methodology lectures new prof.. mbassa

Challenges in formulation of aims

Lack of clarity, purpose of study is

not articulated

Lack of cohesion - no clear link

between title, purpose, objectives

or problem

Over-ambitious aims – not

achievable by resources & time

available

Page 203: Research methodology lectures new prof.. mbassa

Objectives

Objectives; intentions or purposes stated in specific measurable terms

Results are evaluated via objectives Specific objectives constitute means

by which aim/ goal is achieved Specify what to be done Are operational, stating specific

tasks with measurable results to be carried out

Page 204: Research methodology lectures new prof.. mbassa

Objectives are vital because they guide

Methods, instruments, study area;

Data collected, analysis & report

Literature review;Precision on what to

accomplish;Study into defined

parts

Variables, Evaluation;Break aim into achievable &

measurable pieces, Consistent focus

activities in sequence

Page 205: Research methodology lectures new prof.. mbassa
Page 206: Research methodology lectures new prof.. mbassa

Assumptions, limitations

Factors facilitate completion of research

Factors preventing research to be done

Practical & theoretical limitations make results valid and applicable or

inapplicable

Page 207: Research methodology lectures new prof.. mbassa

Qualities of suitable objectives

SpecificMeasurable

Focused

Cover problem coherently &

logically

SystematicOperational

Realistic

Page 208: Research methodology lectures new prof.. mbassa

Methodology (materials & methods)

Describes systematically in detail materials, tools, experimental

designs, methods, logistics how to research to cover objectives

Methodology is to be clear

on experimental

procedure on…

Page 209: Research methodology lectures new prof.. mbassa

What data required;What standards data to meet;How to collect,

process, analyze data

Selected materials for testing;Study area;

Procedures for measuring variables;Units of

measurements

Explain how data will be analyzed;

Trimmed, spliced to show trends or associations;

Communication, dissemination &

utilization of results

Page 210: Research methodology lectures new prof.. mbassa

Work plans

Expected

outputs

Logical framework

matrix

Monitoring & evaluation

Page 211: Research methodology lectures new prof.. mbassa

Matrix of research work plan

Goal(Aim)

1

1

Objective/Output

1

2

Activities

123

123

Outcome to observe

123

123

Month to observed outcome

Jun 2009Dec 2009Mar

2010Jun 2010Sep

2010Dec 2010

Page 212: Research methodology lectures new prof.. mbassa

Logic framework matrix or logframe

Effective tool for

planning and

evaluation of research

States conditions necessary for research

to succeedSummarized in four

column table

Rows represent different levels of

project goal, objectives, outputs,

activities

Page 213: Research methodology lectures new prof.. mbassa

Columns indicate how the achieved objectives are measured, and assumptions for achieving results

In vertical logic are narrative summary at goal, purpose,

objectives or outputs, and inputs in terms of activities levels

Inputs include personnel,

physical and financial resources

Page 214: Research methodology lectures new prof.. mbassa

Outputs are measures of what comes after inputs

energy (inputs & activities cause outcomes [Inputs

are similar to independent variables])

outputs are outcomes caused by activities

[outputs are similar to dependent variables]

Page 215: Research methodology lectures new prof.. mbassa

The goal is

ultimate objective,

not the immediat

e objective

The immediat

e objective

is the purpose, which is the main output

Page 216: Research methodology lectures new prof.. mbassa

Goal and purpose

•Relationship between goal and purpose is less direct and causal

Factors •Many exogenous factors influence the goal

Contribution to goal

•Achieving the purpose is necessary but not sufficient to achieve goal

Page 217: Research methodology lectures new prof.. mbassa

In horizontal logic are

verifiable indicators & means of verification

evidence of achievement of

research project (in deed every project),

& how that evidence is found & measured

Page 218: Research methodology lectures new prof.. mbassa

Indicators & their means of

verification show criteria for attaining

objectives of nature, quantity, quality and time

Assumptions are factors which are not controlled by the research but

influence it (external factors)

Page 219: Research methodology lectures new prof.. mbassa

Narrative summary

Goal

Purpose

Objectively verifiable indicators

Measure of goal

End of project status

Means of verification

Sources of informatio

nMethods

used

Sources of information

Methods used

Assumptions (Risks)

Assumptions affecting Purpose,

Goal

Assumptions affecting the Output,

Purpose linkage

Logic framework

Page 220: Research methodology lectures new prof.. mbassa

Narrative summary

Inputs

Output 1

Activities

Objectively verifiable indicators

Nature & level of

resources, necessary

costs, planned

starting dateQuantitativ

e magnitude

s of outputs

and planned

datesQuantitative magnitudes of outputs

and planned dates

Means of verification

Sources of

information

Sources of informationMethods to

use

Sources of information, Methods to

use

Assumptions (Risks)

Initial assumptions

about the project

Assumptions affecting

Input/Output linkage

Assumptions affecting activities

Page 221: Research methodology lectures new prof.. mbassa

Narrative summary

Inputs

Output 2

Activities

Objectively verifiable indicators

Nature & level of

resources, necessary

costs, planned

starting dateQuantitativ

e magnitude

s of outputs

and planned

datesQuantitative magnitudes of outputs

and planned dates

Means of verification

Sources of

information

Sources of informationMethods to

use

Sources of information, Methods to

use

Assumptions (Risks)

Initial assumptions

about the project

Assumptions affecting

Input/Output linkage

Assumptions affecting activities

Page 222: Research methodology lectures new prof.. mbassa

Monitoring & evaluation plan

Goal(Aim)

1

1

Objective/Output

1

2

Activities

123

123

Output quantity

123

123

Month to observe output

Jun 2009Dec 2009Mar

2010Jun 2010Sep

2010Dec 2010

Page 223: Research methodology lectures new prof.. mbassa

List all literature cited in the text

Page 224: Research methodology lectures new prof.. mbassa

DATA COLLECTION

Samples may be whole mammals, birds, reptiles, amphibians, fish, higher or lower plants, heminthic worms; nematodes, cestodes and trematodes,

Page 225: Research methodology lectures new prof.. mbassa

fungi, bacteria and virus or their parts, cells, tissues and organs, secretions such as milk, sweat, tears, hormones, antibodies, vitamins, proteins, urine, bile acid, bile, cell, embryo, saliva, semen or other

Page 226: Research methodology lectures new prof.. mbassa

They may be macromolecules such as nucleic acids, proteins, fats, carbohydrates or product, metallic or non metallic, feeds, pathologic tissues, blood, plasma, serum, dyes or various compounds, simple and complex

Page 227: Research methodology lectures new prof.. mbassa

DATA ANALYSIS AND PRESENTATION

Biological samples collected from plants or animals are processed in the laboratory and analyzed to yield results by standard biological methods

Page 228: Research methodology lectures new prof.. mbassa

Samples may be whole mammals, birds, reptiles, amphibians, fish, higher or lower plants, heminthic worms; nematodes, cestodes and trematodes, fungi, bacteria and virus or their parts, cells, tissues and organs, secretions such as milk, sweat, tears, hormones, antibodies, vitamins, proteins, urine, bile acid, bile, cell, embryo, saliva, semen or other

Page 229: Research methodology lectures new prof.. mbassa

They may be macromolecules such as nucleic acids, proteins, fats, carbohydrates or product, metallic or non metallic, feeds, pathologic tissues, blood, plasma, serum, dyes or various compounds, simple and complex

Page 230: Research methodology lectures new prof.. mbassa

Each biological product is analyzed by a standard biological procedure in various instruments such as microscopes, celloscopes, spectrophotometers, scans, dyes, chromatograph, and many others

Page 231: Research methodology lectures new prof.. mbassa

The results of field data, observations, morphology, measurements, questionnaires, innovation data; new material, drug, vaccine, machine, bacteria, feed to be found and tested are described

Page 232: Research methodology lectures new prof.. mbassa

Impact data for a process going on, qualitative and quantitative data information obtained is also presented in detail.

Data is first entered in electronic form and storage (Data Base or Spreadsheet), then cleaned, statistical methods for analysis chosen and executed

Page 233: Research methodology lectures new prof.. mbassa

Presentation of Results

Descriptive texts Tables Figures (graphs, pictures,

diagrams, charts, histograms, box plots or other drawings)

Page 234: Research methodology lectures new prof.. mbassa

Describes in detail salient features of results obtained referring to tables and figures which explicitly or brightly show the characteristic or trend being reported

Tables classify data to facilitate comparisons and reveal relationships or trends of the data being presented

Page 235: Research methodology lectures new prof.. mbassa

Tables are self-explanatory, giving more illustrative information in support of the description in the text.

There are many varieties of organizations of research results in tables, depending on the type of experiment, its design and treatment structure

Page 236: Research methodology lectures new prof.. mbassa

Tables are numbered in the proper sequence relating to the flow of ideas and/or the sequence of events in the text but typed on separate pages

In the text, the first letter of the word table(s) is capitalized e.g. Table 1, Tables 4 and 5, even when it appears in the middle of the sentence

Page 237: Research methodology lectures new prof.. mbassa

Tables may be arranged in different ways, but the format that demonstrates results most effectively is selected

Many types of data have decimal points. It is important to reduce decimal places to the precision required in the parameter being determined

Page 238: Research methodology lectures new prof.. mbassa

Figures in form of pictures, diagrams, charts are commonly used in biological presentations, the most commonly used are line graphs, bar or pie charts in two or dimensions with X, Y or Z axes

Line graphs; suitable for presenting a relationship between continuous (quantitative) variables e.g milk or rice grain yield response to varying rates of fertilizer levels

Page 239: Research methodology lectures new prof.. mbassa

Bar charts are used for discrete (discontinuous) data e.g. frequency distribution

Pie charts are used to present the relative magnitude of the components of a whole unit

Page 240: Research methodology lectures new prof.. mbassa

In degree of accuracy in presenting data, graphs can be ranked in descending order of line graphs, bar and lastly pie charts

Page 241: Research methodology lectures new prof.. mbassa

In bar charts, Y axis begins at the zero so that both the relative and absolute bar heights reflect accurately magnitude of treatment means and their differences, truncating it exaggerate the differences

Page 242: Research methodology lectures new prof.. mbassa

A line graph cannot be valid unless it is based on a minimum number of three treatments i.e three data points)

Tables are more accurate in data presentation

Page 243: Research methodology lectures new prof.. mbassa

DATA INTERPRETATION

1. Analysis of causation of effect by a factor

2. Interpretation in multi-causal factors and multi-effects

3. Statistical association 4. Application of statistics in

interpreting data

Page 244: Research methodology lectures new prof.. mbassa

1. Analysis of causation of effect by a factorA Variable is a property, a factor, or a characteristic of a system, an individual, a group, or a system that changes or causes a change in quality or quantity of another property or characteristic of any system

Page 245: Research methodology lectures new prof.. mbassa

variable Outcome Outcome

Page 246: Research methodology lectures new prof.. mbassa

Putative causes of an outcome (or effect) such as transcription level, translation level, death of a cell, disease in an animal, milk produced, protein produced, etc

Page 247: Research methodology lectures new prof.. mbassa

Putative causes are exposure or risk factors (as independent predictor or variable, or explanatory variables) producing the outcome of interest (the effect)

Page 248: Research methodology lectures new prof.. mbassa

(e.g. disease, gene transcript, response, productivity, cell secretion, absorption, reproduction, .growth or others). An effect is a measurable response and is a dependent variable

An Outcome or Effect is a measurable outcome of action of a cause

Page 249: Research methodology lectures new prof.. mbassa

A determinant is any factor that when altered produces a change in the frequency or magnitude or characteristics of a dependent variable

Page 250: Research methodology lectures new prof.. mbassa

Factors are such as age, breed or sex in animals and many others in non animal biological or non biological systems

Page 251: Research methodology lectures new prof.. mbassa

Association of causal and outcome variables may be true associations or spurious associations (chance, bias, confounding)

Page 252: Research methodology lectures new prof.. mbassa

Many determinants are external to the system (animal, cell, culture, tissue others)

Internal factors relate to the intrinsic elements of the system (such as the pathogenesis of a disease in an animal).

Sufficient cause is amount of factor that reaches a threshold to cause an outcome

Page 253: Research methodology lectures new prof.. mbassa

Purpose of data analysis in many experiments and research is to determine whether suspected factor (agent, independent variable) is the cause of a specific outcome or response (dependent variable).

Page 254: Research methodology lectures new prof.. mbassa

Early (Henle-Koch) Guidelines for determining causal-effect relationship Independent variable casual

(agent) must be present in every case of that outcome (dependent variable);

Independent variable must not be present in other systems (where the outcome of interest is not there);

Page 255: Research methodology lectures new prof.. mbassa

Independent variable must be isolated from dependent variable (tissues, cultures);

Independent variable must be capable of inducing the response or the outcome of interest under controlled experimental conditions

Page 256: Research methodology lectures new prof.. mbassa

Henle-Koch guidelines are not sufficient because outcome of interest may not be caused by a single agent, i.e. dependent variable may be affected by many independent variables (multi-causal effects)

Page 257: Research methodology lectures new prof.. mbassa

Causal agent (independent variable) may also cause many effects (many outcomes) or many dependent variables. There are;

(1) Multiple aetiological factors/ agents of effects (e.g. disease, genes, secretions, hormones) (multi-causality in addition to mono-causality for one causal factor)

Page 258: Research methodology lectures new prof.. mbassa

(2) Multiple effects of single causes

(3) Causes may also be affected by other factors (quantitative causal factors or determinants)

Page 259: Research methodology lectures new prof.. mbassa

2. Interpretation in multi-causal factors and multi-effects

Guidelines for multi-causal agents and multi-effect outcomes & many outcomes are analyzed by;

Page 260: Research methodology lectures new prof.. mbassa

Methods of agreement Methods of difference Methods of concomitant

variation Methods of analogy Methods of residue

Page 261: Research methodology lectures new prof.. mbassa

Method of agreement

An outcome (an effect) occurs under a variety of circumstances but there is a common factor. This factor is the cause of the outcome

Page 262: Research methodology lectures new prof.. mbassa

Method of difference If circumstances where an effect

(dependent variable) occurs are similar to those circumstances where causal factor ( independent variable) does not occur, except where there is one factor difference, this factor or its absence is the cause of outcome (effect of interest).

Method of difference is basis for keeping all factors constant except for one factor in experimental research design

Page 263: Research methodology lectures new prof.. mbassa

Methods of concomitant variation If independent factor (Causal factor)

and effect (outcome, response, dependent variable) have a dose dependant relationship factor may be cause of response

Independent factor (variable) or agent whose strength or frequency varies directly with occurrence of outcome convinces a causal- effect relationship

Page 264: Research methodology lectures new prof.. mbassa

Method of analogy

If the distribution of an outcome is sufficiently similar to another factor, it may be that there is a causal-effect relationship

Page 265: Research methodology lectures new prof.. mbassa

Method of residue

If the factor only explains X% of the outcome other factors must be identified to explain the remainder or (100-X%)

Page 266: Research methodology lectures new prof.. mbassa

to increase the experimental precision

Control variables that are not of interest

Increase the sample size Repeat experiment in another

location Sampling without bias

Page 267: Research methodology lectures new prof.. mbassa

Guidelines to concept on inferences in causation (developing causal inferences)

Incidence of outcome must be higher in the material animal, cell or other system which has been exposed to the putative causal agent than in non exposed systems

Page 268: Research methodology lectures new prof.. mbassa

Exposure should be more common in cases where outcome occurs/has occurred than in those where outcome (effect) has not occurred

Exposure to putative cause must precede outcome

There should be a spectrum of measurable responses on dependent variable

Page 269: Research methodology lectures new prof.. mbassa

Elimination of putative cause results in elimination of the outcome

Preventing or modifying the independent variable decreases or eliminates the outcome or effect

The outcome must be reproducible experimentally

Page 270: Research methodology lectures new prof.. mbassa

The sequence for the researcher in assessing causation is;

to demonstrate that association exists

To assess likelihood that causal association exists

To elaborate nature of causal association

Page 271: Research methodology lectures new prof.. mbassa

3. Statistical association

For factor to be causally associated with outcome, rate of outcome in exposed members must be different from not exposed members.

To evaluate probability that sampling error may have accounted for observations a statistical test is required

Page 272: Research methodology lectures new prof.. mbassa

Member status

Outcome present

Outcome absent

Total

Exposed a b a+b

Not exposed

c d c+d

Total a+c a+b+c+d

A 2x2 table displaying the relationship between two dichotomous variables, one the factor, the other for the outcome

Page 273: Research methodology lectures new prof.. mbassa

Proportional or rate of interest Exposed to factor in population p(F+) Outcome (effect) is positive in

population (E+) Affected and exposed to factor p(F+

and D+) Affected in exposed members

p(D+/F+) Affected in non exposed members

p(D+/F-) Exposed to factor in effected

members p(F+/D+) Exposed to factor in non affected

members p(F+/D-)

Page 274: Research methodology lectures new prof.. mbassa

Working out rates from statistics (representing parameters)

F+ = (a + b)/n D+ = (a + c)/n F+ and D+ = (a/n) D+/F+ = a/(a + b) D+/F- = c/(c+d) F+/D+ = a/(a+c) F+/D- = b/(b+d) n=a+b+c+d

Page 275: Research methodology lectures new prof.. mbassa

A statistical test would be the chi-square X2

X2 = [Іaxd)-(bxc)l- O.5n] x n (a+b) x (c+d) x

(a+c) x (b+d)

In a 2x2 table all statistics have one degree of freedom, the critical value for significance at 5% level is 3.84

Page 276: Research methodology lectures new prof.. mbassa

4. Application of statistics in interpreting data

Statistical difference is a function of;

The magnitude of difference The variability of difference The sample size

Page 277: Research methodology lectures new prof.. mbassa

Strength of association between causal factor and effect (outcome, response) is called Relative Risk, Risk ratio, incidence ratio, prevalence ratio.

Relative Risk is calculated as ratio between rate of response in members exposed to causal factor and rate of response in members not exposed to causal factor.

If there is no association between cause and outcome RR is 1.

Page 278: Research methodology lectures new prof.. mbassa

The greater the departure of RR from 1 larger or smaller, the stronger the association

Attributable fraction (AF) is the proportion of response (outcome) in the factor exposed group that is due to the factor

Page 279: Research methodology lectures new prof.. mbassa

Measures of association for independent proportions in 2x2 tables.

RR = [a/(a+b)]/[c/(c+d)]

Odds ratio = ad/bc

Page 280: Research methodology lectures new prof.. mbassa

Effect of factor in exposed members here is some effect in factor negative group/members, not all outcome in factor exposed members is due the factor

In calculating attributable rate assumption is that other factors which lead to some outcome in factor negative group operate with same frequency and intensity in factor positive group

This absolute difference is called Attributable Rate (AR)

Page 281: Research methodology lectures new prof.. mbassa

AR is calculated by subtracting the rate of response (outcome) in the un-exposed group from the rate in the exposed group

AR is the rate of outcome response in the group due to the exposure

The larger the AR the greater the effect of the factor in the exposed group

AR = [a/(a+b)]-[c/(c+d)] AF = AR/[a/(a+b)] or = (RR-I)/RR

Page 282: Research methodology lectures new prof.. mbassa

Causal Inferences in Observational Studies

In interpreting ARs and AFs assumption is that there is a cause and effect relationship

However statistical associations do not in themselves represent a causal association

There are important precautions needed to eliminate this problem

Page 283: Research methodology lectures new prof.. mbassa

Solutions for statistical deficiencies

Select sampling method better for measuring associations, such as cohorts

Refine independent and dependent variables by using cause specific rather than crude factors, this strengthens associations

Seek other variables that produce or explain associations or lack of associations (eliminate confounding)

Page 284: Research methodology lectures new prof.. mbassa

Confounding variables A confounding variable is one

associated with the independent variable and the dependent variable

Confounding variables may be determinants of an outcome

For example a disease e.g. mastitis may be associated with age, castration is associated with age, therefore the effect of age is to be taken into account

Page 285: Research methodology lectures new prof.. mbassa

Control of confounding variables

Exclusion (restricted sampling), select units with only one level of confounding or without confounding

Matching, Equalize the frequency of confounding variable in two groups being compared, e.g. cohorts

Analysis, stratify data and display in a series of 2x2 tables, one table for each level of confounding

Page 286: Research methodology lectures new prof.. mbassa

A method that summarizes associations in multiple tables is called Mantel-Haenszel technique, here the Odds ratio (OR) is the measure of association.

OR = (Σad/n)/( Σbc/n) These methods are not mutually

exclusive, all may be applied at the same time

Page 287: Research methodology lectures new prof.. mbassa

Criteria of judging causal inference

1. Time sequence, the factor to cause an outcome must precede the outcome, well designed cohort studies give best experiments

2. Strength of association, is measured RR or Odds ratio. The greater the . departure from 1, the strong the association

Page 288: Research methodology lectures new prof.. mbassa

1. Dose-Response relationship, there exists an association if increasing amount of causal factor produces higher and higher level of outcome

2. Coherence, an association is more likely to be causal if it is sensible

3. Consistence, an association is likely to exist if it is supported by similar findings under different conditions

Page 289: Research methodology lectures new prof.. mbassa

4. Specificity of association, a single causal factor (if crude) may produce a number of effects (outcomes). Refining the causal variable produces a better outcome thus strengthens the association

Page 290: Research methodology lectures new prof.. mbassa

Elaborating Causal Mechanisms

If an association of two variables is causal the nature of the association may be determined by

Indirect and direct causes the causal association is direct when there is no intervening variable between the factor and the outcome.

Page 291: Research methodology lectures new prof.. mbassa

Both the independent and the dependent variables are measured at same level of organization (individual, cell, tree, fruit, group of trees, farm). All other causes are indirect.

Necessary and sufficient causes; another dimension of classifying determinants. A necessary cause is one without which the outcome cannot occur. Sufficient cause is one that always produces the outcome

Page 292: Research methodology lectures new prof.. mbassa

Path model causation; Path model provide other ways for conceptualizing, analyzing and demonstrating the causal effects of multiple factors. Variables are ordered and causal effects flow along arrows and paths

Statistical methods are applied to estimate the relative magnitude path coefficients of each path

Page 293: Research methodology lectures new prof.. mbassa

Displaying effects of multiple factors; Produce ven diagram, important when the risk values increase steadily in number of putative causal factors. Calculate RR or OR of outcome for each combination of independent variables relative to the lower risks (group member outcome).