This is the old United Nations University website. Visit the new site at http://unu.edu

Guidelines for the development of research proposals following a structured, holistic Approach for a Research Proposal (SHARP)

Abstract
Introduction
Objectives of these guidelines
Objectives of a research proposal
Contents of a research proposal
Selected readings
Appendix. Metaplan technique

Rainer Gross, Darwin Karyadi, Soemilah Sastroamidjojo, and Werner Schultink

Rainer Gross and Werner Schultink are affiliated with the Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ) GmbH, in Eschborn, Germany. Darwin Karyadi and Soemilah Sastroamidjojo are affiliated with the Regional SEAMEO-TROPMED Center for Community Nutrition in the University of Indonesia in Jakarta, Indonesia.

Mention of the names of firms and commercial products does not imply endorsement by the United Nations University.

Abstract

SHARP (a Structured, Holistic Approach for a Research Proposal) is a structured method for developing a research proposal that can be used either by individuals or by teams of researchers. The eight steps in SHARP are (1) setting up a causal model, (2) establishing a fact-hypothesis matrix (FaHM), (3) developing a variable-indicator-method matrix (VIM), (4) selecting the study design, (5) defining the sampling procedure and calculating the sample size, (6) selecting the statistical methods, (7) considering the ethical aspects, and (8) setting up an operational plan. The objectives of the research proposal are to help the researcher to define the contents and to plan and execute a research project, and to inform potential collaborators and supporters about the topic. The proposal that is produced during the process can be submitted to agencies for possible funding.

Introduction

The following guidelines for writing a research proposal have been developed on the basis of experiences at academic institutions providing advanced training in nutrition. We have noted that students who plan and conduct their first research projects often have difficulty orienting themselves within the vast field of scientific information and tend to focus their proposals on a minimum of relevant issues, ignoring many essential items. Therefore, a structured method has been developed as an orientation tool to help guide young researchers into scientific research. Furthermore, the method has also been found useful for the development of research proposals by a team. This is especially important in interdisciplinary sciences such as nutrition, when individuals with different experiences and expertise have to be brought together. This instrument and procedure has been named SHARP: Structured, Holistic Approach for a Research Proposal.

SHARP should be implemented in eight consecutive steps:

1. Set up a causal model;
2. Establish a fact-hypothesis matrix;
3. Develop a variable-indicator-method matrix;
4. Select the study design;
5. Define the sampling procedure and calculate the sample size;
6. Select the statistical methods;
7. Consider the ethical aspects;
8. Set up an operational plan.

Objectives of these guidelines

The objectives of these guidelines for the development of a research proposal are to:

» improve, standardize, and maintain research quality and performance;
» stimulate the orientation and training of scientific personnel;
» facilitate comparisons among research projects;
» facilitate the development of proposals that can be submitted to agencies for research funding.

Objectives of a research proposal

A research proposal has two main objectives:

» to help the researcher to define the contents and to plan and execute his/her research project;

» to inform potential collaborators and supporters about the topic and the expected quality of the research.

» justify the chosen research project;

» describe the current state of knowledge on the research topic, considering all important relevant literature;

» formulate the hypothesis or research question;

» define the research strategy and methodology to be used to test the hypothesis or research question;

» discuss ethical considerations about the research methodology;

» define realistic, feasible, operational planning, based on the research methodology and general conditions;

» inform potential collaborating institutions and persons about the research project and enable them to identify the kind of support they can give;

» serve as an important tool for monitoring the research.

Contents of a research proposal

The presentation of a research proposal follows a logical sequence according to the following steps:

1. introduction to the research topic, formulation of the problem, and justification for its selection;

2. development of the causal model;

3. formulation of the hypothesis;

4. definition of all variables and their indicators;

5. selection of the study design;

6. description of the population to be studied;

7. description of the sampling procedure;

8. selection of measurement methods and statistical methods;

9. development of the operational plan, including time schedule, human resources, equipment, material, and budget.

Title page

The purpose of the title page is to present a concise statement of the subject of the research and to identify the responsible researchers. The title page is the “main gate” of the research proposal, which invites the reader to enter the research proposal.

The title should:

» express the main message of the research topic;
» be relevant;
» be short;
» be clearly and precisely formulated;
» be exciting;
» be appealing.

» title of the research project;
» name of the principal researcher;
» date of submission of the research proposal (month and year);
» name and address of the institution of the principal researcher;
» telephone number, fax number, and e-mail address of the principal researcher.

» name(s) of the scientific collaborator(s) (e.g., supervisors and promoters);
» name(s) and address(es) of the institution(s) of the scientific collaborator(s).

Table of contents

The table of contents outlines the structure of the research proposal. An example is shown in table 1. The headings and subheadings are structured and numbered, and the appropriate page numbers appear at the right-hand margin. The headings of the table of contents are identical to those in the body of the report.

Abstract

The purpose of the abstract is to summarize in less than 200 words all important parts of the research proposal. The abstract should:

» describe the general objective of the study (justification);
» define the central hypothesis;
» describe the site and population to be studied;
» summarize the total time and budget necessary to carry out the research.

Purpose of the introduction

The introduction should justify the hypothesis of the proposed research. In particular, it should:

» summarize the relevance of the topic;
» give an overview of the status of international research in related areas;
» finally, lead to the objectives and hypotheses of the proposed research topic.

Title Page
Table of Contents
0 Abstract
1 Introduction
	1.1 Relevance of the topic
	1.2 Current state-of-the-art
	1.3 Causal model
	1.4 Fact-hypothesis matrix (FaHM)
	1.5 Objectives of the research project
	1.6 Hypothesis of the research project
2 Methodology
	2.1 Variable-indicator-method matrix (VIM)
	2.2 Study design
	2.3 Population under survey
	2.4 Sample size and sampling procedure
	2.5 Measurement methods and statistical analysis
	2.6 Ethical considerations
3 Operational Planning
	3.1 Time schedule
	3.2 Human resources
	3.3 Equipment and material
	3.4 Budget
4 Appendices
	4.1 Literature cited
	4.2 Draft of questionnaire
	4.3 Curriculum vitae of proposer

Relevance of the research topic

In the introduction, the research topic is placed in the general framework of current knowledge. The proposed research is justified, and after the general framework has been described, the topic is narrowed down. Research submitted for a doctoral thesis must be original; it must be substantially new.

Current state-of-the-art

The review of literature should focus on the knowledge required to test the hypotheses. Literature citations should be used to support factual statements. It is necessary to filter out ideologies, unsubstantiated dogma, open questions, and untested hypotheses. Rambling on, mentioning uncritically everybody who ever said anything related to the topic confuses more than contributes to the analysis of information available about the topic. The selection and discussion of the literature strongly indicates the intellectual capacity of the researcher and the time spent in preparing the research proposal. This requires careful review and analysis of the literature.

Modern computerized systems allow for the collection of an enormous number of literature citations. It maybe difficult for a less experienced researcher to judge the relevance of a citation. The establishment of a causal model (see below) facilitates the decision of relevance.

Causal model

During the development of a causal model (frame of thinking), all important variables are denned and their relationship to the central hypothesis is identified. The causal model is based on conclusions drawn from the literature. An example of a causal model is shown in figure 1.

The causal model helps to identify all relevant variables that contribute to a hypothesis and define the expected cause-effect relationships of the variables. In particular, a causal model helps to:

» select relevant literature for reading;

» identify all necessary variables that have to be controlled by the study, including confounding factors;

» identify hypothetical relationships between variables.

» A concise statement of the core situation is written down and enclosed in a box.

» Each direct (proximate) cause (variable) that could influence the core situation is identified and written down in a box below the box containing the core situation. The box(es) containing the direct cause(s) of the core situation are connected to the box containing the core situation with directional arrows.

» Each box should be consecutively numbered for later identification.

Facts and hypotheses

The fact-hypothesis matrix is a systematic way to organize causal relationships proposed in the causal model (table 2). Each relationship is identified as either hypothetical or proven. In the first three columns of the matrix, each causal relationship of the model is summarized. Each direct cause is identified by a letter, and the respective dependent (effect) and independent (cause) variables are listed in the second and third Columns.

FIG. 1. Example of a causal model

TABLE 2. Fact-hypothesis matrix (FaHM)

Letter	Effect	Cause	References
A	Stunting	Energy intake	H
A	Wasting	Energy intake	H
A	Stunting	Vitamin A intake	H
B	Stunting	ARI	Neuvianz et al. (1990)
B	Stunting	Diarrhoea	Black et al. (1984)
B	Stunting	Helminths	Stephenson et al. (1980)
C
D
E
F
F
G

In the fourth column, the literature citation that confirms that the stated variable is a cause of the core situation is cited. If no literature source can be found, an H for a hypothetical cause-effect relationship is entered in the fourth column. A causal model with many hypothetical cause-effect relationships indicates a very uncertain frame of thinking, which makes the outcome of the research work very unpredictable and planning more complex.

Formulation of the objectives of the research project

The general objective of a research project is defined as the purpose of the project. It states what the project intends to accomplish or develop in relation to observed problems or situations. The following criteria should be considered when formulating the general objectives of the project. An objective must (1) be applicable to the situation, (2) be achievable and measurable, (3) not be ambiguous, and (4) be harmonious with societal and institutional goals and constraints.

Sometimes after the single, general objective has been stated, it is convenient to list a series of component objectives or specific objectives. These state specific research results and sequential goals to be achieved during the study. They are concrete, attainable results that can be measured and are readily identified when they have been reached. Specific objectives must be necessary and sufficient to reach the general objective of the research project.

Formulation of the hypothesis of the research project

The hypothesis of a research project is an educated guess, or a prediction of causal relationships that can be tested; an unanswered question that arises from a literature review of a topic leads to such predictions. Data are then gathered and analysed to test the hypothesis.

If a hypothesis refers to the relationship between two variables, the formulation should indicate clearly the direction of their relationship. The relationship stated in a hypothesis should be tested by measuring appropriate variables and judging whether the values of these variables could have occurred by chance. If the observations could have occurred by chance, then the researcher concludes that the null hypothesis is true and that the hypothesized relationship does not explain the data. To prove a relationship, the null hypothesis must be proven to be wrong.

TABLE 3. Variable-indicator-method matrix (VIM)

No.	Variable	Indicator	Method	Reference
0	Nutritional status	wt/age, wt/ht, ht/age	Anthropometric measurement	WHO (1983)
	Nutritional status	MUAC/age	Anthropometric measurement	UN (1986)
	Nutritional status	Hemoglobin status (g/L)	Cyanide	INACG (1985)
1	Infectious/invasive diseases	Period prevalence ARI	Questionnaire	WHO (1985)
	Infectious/invasive diseases	Period prevalence diarrhoeal disease	Questionnaire	Black et al (1984)
	Infectious/invasive diseases	Period prevalence ascariasis	Stool examination	Brown et al (1980)

Abbreviations: ARI, acute respiratory infection; MUAC, mid upper arm circumference.

» based on a known fact or theory,
» testable,
» specific,
» brief, but clear.

Definition of methodology

The purpose of the methods section is to identify and justify the research methods selected for the research. The choice of methods depends on the variables to be measured and the cause-effect relationships among them. One method for developing a community nutrition research project is the causal model approach, as defined above.

Variables and indicators

A variable is a characteristic of the study subjects (e.g., nutritional status), and an indicator is a measurement collected during research that is assumed to reflect the variable (e.g., blood haemoglobin level).

The variable-indicator-method matrix (VIM) is a systematic way to organize the relationship between variables of interest and potential indicators of these variables. The VIM should:

» relate every variable (cause) of the causal model to at least one indicator;
» describe the methodology by which each indicator will be surveyed;
» cite the literature source of the methodology selected.

Indicators of variables are specified in the third column. Each variable should be related to at least one indicator, which defines the variable precisely (e.g., variable: nutritional status; indicator: Z-score of weight/height index according to the NCHS reference population). The indicators should be selected according to the following criteria:

» Validity

- Does it measure (quantify or describe) what we assume it measures? For example, does the indicator we have chosen to show obesity indeed measure the fatness of a person?

» Feasibility/appropriateness

- Is the cost realistic?
- Is the equipment available?
- Is the methodology appropriate, and can data be obtained?

» Accuracy (getting the correct answer). This includes:

- sensitivity
- specificity

» Precision (reliability, reproducibility, repeatability). There are several kinds of precision, including:

- instrumental (precision of analytical instrument on same sample on different occasions);
- biological (precision of same subject on different occasions);
- intra-observer (precision of same tester on different occasions on same subject);
- inter-observer (precision of different testers on the same subject at same occasion).

Study design

After variables to be surveyed have been identified and indicators related to each of the variables have been defined, the study design has to be selected and presented in the text and as a diagram. Common types of study designs include the following (fig. 2):

» Observational study

- prospective study (prospective cohort study)
- retrospective study (case-control study)
- historical prospective study (retrospective cohort study)
- cross-sectional study

» Experimental study

- clinical trial
- community trial

FIG. 2, Common types of study design

The criteria of the selection of the most appropriate study design should be listed and advantages and disadvantages of the design should be discussed. Special attention should be given to possible confounding factors related to the selected study design.

It is highly recommended that whenever possible, double-masked tests (formerly called double-blinded) be used in intervention studies. “Double-masking” implies that neither the researchers nor the participants in the intervention are aware of the assignment of the treatment. This eliminates evaluator bias on the part of the researcher and behavioural bias on the part of the participant.

Population under survey

The proposal needs to define the criteria to be used in selecting the population to be surveyed and needs to distinguish the selection criteria from the descriptive data that will be collected to characterize the sample. For example, the selection criteria for a study might be women with two children under the age of five years - all other women and all men would be excluded. The study population would then be characterized by maternal age, number of dependent children, geographic area, etc.

Ideally, the sample selected from a population is representative of the entire population, and therefore the characteristics of the study sample describe those of the entire population.

Sample size and sampling procedure

After identification of the variables and their indicators and the selection criteria for the population to be surveyed, the sample size and sampling methods must be defined. Obviously the selection of the population, the sampling method, and the sample size must be coordinated. By convention, N represents the number of individuals in the population and n the number of individuals in the sample.

It is important that the sample size be large enough that statistically significant differences will be identified if they exist. The calculations concerning sample size depend on:

» the primary question that the researchers want to investigate;
» the way in which it is to be answered.

The following sampling methods are applicable to community nutrition research:

Probability sampling methods rely on formal random techniques to identify the units to be included.

» Simple random sampling. A fixed percentage of the population is selected using a formal random process, such as a random number generator or random number table.

» Systematic random sampling. The n sampling units are selected from the sampling frame at regular intervals (e.g., every fifth house). When systematic methods are used, the starting point in the first interval is selected on a formal random basis.

» Stratified random sampling. Before selection, the sampling frame is divided into strata based on factors likely to influence the variable being estimated (e.g., variable: nutritional status; factor: income). Then a simple random or systematic random sample is selected within each stratum.

» Cluster sampling. Primary sampling units are defined, which are logical groups or clusters (e.g., classrooms) of secondary sampling units (e.g., individual children). The clusters can be selected by systematic, simple, or stratified random methods, and all individuals within the primary sampling units (or clusters) are selected to participate in the research.

» Multistage sampling. This method is similar to cluster sampling, except that sampling takes place at all stages. As an example of two-stage sampling, one would begin as in cluster sampling by selecting a sample of the primary units (e.g., classrooms) listed in the sampling frame. Then within each primary unit, a sample of secondary units (e.g., individual children) is selected. This procedure differs from cluster sampling, in which all of the secondary units within each selected primary unit are taken.

» Judgement sampling. Representative units of the population are selected by the investigator.

» Convenience sampling. The sample is selected because it is easy to obtain. Using convenience or judgement sampling often produces biased results, regardless of whether the researcher believes he/she can select representative samples. Therefore, these samples should rarely be used for survey purposes.

» Purposeful sampling. The selection of units is based on known exposure or disease status (for example, children with severe diarrhoea admitted to the hospital). Purposeful sampling is often used to select units for analytic observational studies, but it is in- adequate for obtaining data to estimate population parameters.

The purpose of statistical analysis is to allow the researcher to draw conclusions from the data obtained. Therefore all aspects of the research proposal should be formulated so that valid statistical conclusions can be drawn.

Appropriate statistical methods should be selected based on the study design and the sampling techniques. Before analysing the data with a certain statistical test, it is essential to determine some characteristics of the data.

Research data are differentiated into at least three categories:

» Frequency or nominal data. Each value represents a characteristic or group membership (e.g., sex: male = 1, female = 2; place of origin: south = 1, central = 2, north = 3).

» Ranking or ordinal data. The values imply a relative rank of the characteristic, but not the magnitude of differences between ranks (e.g., formal education: none =1, can read and write =2, completed primary school =3, completed secondary school =4).

» Measurement value or fixed interval data. Values are from a scale with constant intervals and known size (e.g., size, weight, age, haemoglobin level). The choice of statistical test depends on several aspects of the hypothesis being tested:

The nature of the research question. The hypothesis can ask for a comparison between proportions of observations or between the mean values of two groups, or it can predict group membership based on the values recorded for several variables. There are many types of research questions, and table 4 shows which tests are appropriate for common types of questions.

Ethical considerations

Each study and survey conducted on human beings and animals needs approval from an official committee of a research institution. This committee evaluates the research proposal according to the following criteria:

» maximizing benefit
» avoiding harm and minimizing discomfort
» confidentiality
» conflict of interests.

TABLE 4. Selected statistical tests

Factors	Sampling	Data category
		Frequency	Ranking	Measurement values
				Non-normal distribution	Normal distribution
Tests for 2 factor steps	Independent sampling	c2 test	Siegel-Tutzey test	Kolmogoroff-Smirnoff test	F test
			U test		b test
	Joint sampling	Tests for indications	Wilcoxon test		t test
			Spearman rank-order correlation coefficient		Product-moment correlation
					Linear regression
Tests for >2 factor steps	Independent sampling	c2 test	H test		Variance analysis
					Student-Newman-Keuls test
	Joint sampling	Q test	Friedman test		Variance analysis
			Multiple comparison between Wilcoxon and Wilcox		Multi-various methods

The research proposal should provide information on how the individuals will be directed for treatment in case health problems are identified during the study. Subjects allocated to a control group should be able to benefit from the treatment after ending the study. For example, if in an iron-supplementation study anaemic individuals have been identified who are not allocated to the treatment group or who are in the treatment group but are still anaemic at the end of the study, adequate treatment has to be provided for them at the end of the study.

A copy of the declaration of informed consent that should be signed by the subjects or their caretaker must be available in the appendix of the research proposal.

Operational planning

The purpose of operational planning is to synthesize all research activities into a working plan. The research activities include additional activities that are not directly linked to the scientific work. For example, during the selection of the study population, individuals with diseases need to be treated, and at the end of the study, the community needs to be informed of the results of the study, which should lead to interventions, if relevant. The plan includes a time schedule, a summary of the human resources, equipment, and material needed to complete the plan, and a budget. The plan needs to be checked carefully to determine whether it is feasible and whether it will indeed allow the researcher to test the stated hypothesis.

If methods have not been validated, the plan needs to include a provision for a validation procedure. In studies in which more than one person takes measurements, carries out tests, or makes observations, provisions for training the research team are essential.

Time schedule

The length of time needed to complete each part of the research project needs to be clearly identified. It is very important that the estimates be realistic, because the allocation of resources (human resources, equipment, etc.) may need to be coordinated with other projects. All aspects of the project must be included: planning, validating survey instruments, training in data collection and analysis, providing information to the community and to institutions, follow-up treatment, etc. A balance needs to be reached: usually, the longer a project takes, the more expensive it becomes; however, a hurried project may not yield reliable results.

Figure 3 provides an example of a time schedule for a Master of Science research project. It shows how to represent the time required for each activity and indicates which portions of the project can be conducted simultaneously. The schedule starts after the acceptance of the research proposal. The literature research for the development of the proposal must begin before the schedule starts. However, literature study is not finished with the acceptance of the research proposal.

Much time can be saved if the researcher has met with and explained the activities to all parties involved in the research, such as local representatives of governmental offices, non-governmental organizations, and the individuals and the community to be studied.

FIG. 3. Example of an operational plan for a Master of Science research project

Research tools such as equipment, material, and questionnaires have to be checked and validated. It may therefore be necessary to carry out a reconnaissance or pilot study before beginning the main study.

Sufficient time and resources must also be available for the treatment of subjects who are diagnosed with health problems. Treatment should also be available for those individuals not chosen to participate in the study. Additionally, treatment must be organized for subjects who participate in the study and still have health problems at the end of the trial. The same issues must be considered in cross-sectional studies.

Research that takes more than one year, such as doctoral research, needs special attention from supervisors. At certain phases of a research project, the principal researcher must inform all those involved about the progress of the research activities. Important reporting times are:

» after a reconnaissance or pilot study, before the main study is to start, so that the collaborators will have a chance to influence the research design and implementation;

» every six months during the research work.

Human resources

All human resources needed to complete the project must be identified. Individuals are not usually named in the research proposal. Human resources include:

» technicians
» consultants
» drivers
» translators
» data entry personnel.

» How many of each type of person will be needed;
» Whether they will be needed full-time or part-time;
» How much each person will be paid;
» How long they will be needed;
» When they will be needed.

All equipment and material needed to complete the project must be identified. Expensive or specialized items must be precisely specified. Other items can be listed in general categories (for example, office supplies). Rented, leased, or shared equipment should be considered expensive items and should be precisely specified. Some or all of the following may be needed:

» paper for reproducing questionnaires
» equipment to be used in the field
» equipment to be used in the laboratory
» transport
» computers, fax machine, and copy machine
» postage and communication costs

» What is needed;
» How much it will cost, including tax and shipping, etc.;
» When it will be needed.

The budget should include all projected expenditures. Generally, a budget is stated on a yearly basis. Within each year, the budget is divided into sections such as:

» personnel (salaries, wages, fringe benefits)
» equipment
» materials and supplies
» printing and publication
» travel
» rental or lease of facilities
» other (utilities, phone, insurance, advertising)
» overhead» contingency

Statement of assumptions

The statement of assumptions identifies external influences that are risks to successful implementation of the research project. This demonstrates how the research project is or can become dependent on its environment and facilitates the evaluation and reduction of any risks that threaten the success of the project. The proposal should discuss any implicit assumptions that may not be readily accepted by the potential collaborators or research supervisors of student projects.

Examples of assumptions that might be important in nutrition research projects during implementation include the following:

» Enough qualified interviewers can be hired locally;
» The population will accept research methods (e.g., blood sampling).

» The political and economic situation will remain stable;
» The health system will continue its current immunization schedule;
» The schools will continue to serve lunch to the children.

Monitoring

Monitoring is an important tool for guaranteeing the success of a research project. It needs to be done continuously, with particular attention at the beginning and end of each stage of the research. Monitoring helps inform all research workers about the situation and possible problems of the research project. In doctoral research, monitoring is carried out by supervisors and proposers. The research proposal is the basis upon which the programme is monitored.

Reporting is an essential part of monitoring. Appropriate reporting by the person responsible for the research is needed for monitoring the activities carried out. The report should consist of the following topics:

» Implementation

- overview of activities carried out during the report period
- institutional affiliations, contacts, and collaborations
- organizational set-up (e.g., personnel hired, logistics)
- collaboration with individuals and population to be studied
- evaluation of implementation (if relevant, reasons for changes)

» Changes of research design (if relevant)

» Preliminary results

- main results of data analysis
- reasons for not achieving the objectives (if relevant)

» Assumptions not directly related to programme

» Plans for the next reporting period

» Finances

- budget spent
- budget planning for the next report period

Appendices

The appendices contain background documentation that provides additional detail on any aspect of the proposal. The detail should complement the proposal text.

Literature cited

The list of cited literature is an important indication of the quality of the research proposal because it shows the thoroughness of the literature search and the understanding of the technical background needed for the project. The cited literature verifies:

» the relevance of the proposed research topic;
» the originality of the hypothesis;
» the validity, accuracy, and precision of the proposed research methodology.

Specific formats for citation in text and in the literature cited section of a research proposal will be dealt with below.

Draft of questionnaire

Questionnaires are an important tool of epidemiological and operational research. Given the nature of original research, a unique survey form must be developed for each research project. The nature of the study and the setting in which the data will be collected will influence the design and structure of the data-recording form or questionnaire; however, the following general principles should be considered:

» The title of the study should appear at the top of the survey form and should be clear and sufficiently detailed to inform collaborators of the general purpose of the survey.

» Questions must be clearly worded, straightforward, and necessary. Initially, it is useful to list all of the variables about which information is required; then structure the questions so that the answer(s) to each question provide the appropriate data.

» Questions should be grouped according to subject matter or another logical basis, such as the temporal relationship of events, to facilitate communication with the respondent.

» It is desirable to record the answers as measurements or continuous variables (e.g., the actual age). Data can be grouped later, if necessary.

» The layout of the questionnaire should assist the analysis and/or computer entry of data. Copying data by hand should be avoided, because each time a number is written down the probability of introducing an error increases.

» Asking questions correctly is as much an art as it is a science. Nonetheless, certain principles should be followed:

- Avoid asking leading questions that suggest a right answer to the respondent.

- Make sure that there is an obvious answer to each question.

- The terminology used in the question should be tailored to the way the respondents use the words and names.

» surveys of individual data
» surveys of household data
» surveys of structural data (data on the village, city, suburb, district, etc.)

The survey forms for each project should be standardized to simplify recording and later reading of the data. For this purpose the following rules should be observed:

1. Each survey form should have a header that provides the following information:

» title of the study
» name of the responsible institution
» type of form (e.g., household form, individual form)
» household number
» individual number (if relevant)

2. Each household must be assigned a unique household number. This number should be entered in the header of the survey form BEFORE commencing the survey. In this way if the pages of the survey become separated, the identity of the household will be clear, and also no two households will receive the same identification number (household number).

3. In case more than one individual of a household is assessed (for example, more than one child below five years of age), besides the household number each individual must be assigned a unique individual number. Consequently, each page of the individual survey form should have both the household number and the individual number, which should be filled in BEFORE commencing the survey.

4. Each variable must have a sequence number.

5. The text of each variable is fully written out, so that the surveyor has no doubts in his/her mind during the interview. This means that the questions on the survey form should be short but specific.

6. The answer categories for closed questions are coded and the codes identified on the survey form (e.g., male = 1; female = 2). For size measurements (such as height, age, and weight), the appropriate unit (such as cm, months, or kg) is shown on the survey form.

7. A box is available on the right-hand side of the form for each variable to provide an identified space for coded answers.

8. All possible responses to a closed-category question are assigned to a number (including categories for “other” and “unknown,” if appropriate). There should be no opportunity to record unique answers, because these add considerably to the workload of data entry and analysis, and where several data collectors are engaged, it is impossible to maintain adequate standardization. The reconnaissance or pilot study is an important stage to convert open questions into closed ones.

9. To reduce errors, answers such as “no answer,” “don’t know,” or “others” should always have the same code (e.g., 99 or 9,88 or 8, and 77 or 7).

10. If no exact answer can be expected in certain cultural contexts, the question should be omitted. It is better to obtain less, but accurate, information than large quantities of erroneous or ambiguous information.

11. The respondent should not read the questionnaire in advance. This will maintain the spontaneity of the answers.

12. Each variable should have its own unique code name that should consist of not more than eight digits. All code names should be listed in the Description of the Variables (a list of all variable codes and their description). For example:

Variable code: ELECTRIC house has electricity. However, since the variable codes are only important for the data analysis and have no further implication for the enumerator, the code need not be stated on the form.

13. Some data needed for analysis are based on further manipulation of the information obtained by the interviewer. For example, to ensure that the question “How many people are there in the household?” is accurately answered, it is best to ask for the name of each member of the household. The interviewer writes down the names and later sums up the total. The variables that are not asked directly are marked on the questionnaire by square brackets [].

INSTITUTION, NUTRITION SURVEY					Household number
Household
35)	Question: If you were to have another child, would you prefer a boy or a girl?
	1) Boy			8) Don’t know
	2) Girl			9) No answer
	3) Boy or girl, it does not matter
36)	Measurement: Weight of the mother (00.1 kg)
37)	Measurement: Height of the mother (cm)
38)	Observation: Could you please show me which kind of salt you use currently in cooking and as table salt?
	1) Packaged iodized salt		3) Coarse, rock, or brick salt
	2) Packaged salt, without label about iodination		8) Don’t know, not sure
	9) No observation made
39)	Measurement: Presence of iodine in tested sample
	1) No colour change		8) Don’t know, not sure
	2) Blue colour change occurs		9) No observation made
40)	Question: During the last 7 days, how often have you frequented a meeting with more than 10 persons?

Samples of survey forms are presented in the Appendix. These are divided into the household level on one hand and the individual level (child) on the other. The household-level sample forms are further divided into samples adapted to urban and rural households, as there are some important differences in socio-economic and ecological descriptions of the two environments.

The sample forms are provided to show appropriate layout and structure. The content and language of the forms used must always be adjusted to the circumstances of the survey area and the respondents.

Curriculum vitae of the proposer

The research proposal finishes with the curriculum vitae of the proposer. The following information should be included in the curriculum vitae:

» full name and academic degrees
» place and date of birth
» office and rank
» higher education degrees
» work experience
» scientific publications
» awards

The form of the research proposal is particularly important, because:

» Often it is possible to draw conclusions about the quality of the content of the proposal from the quality of the form of the research proposal. For instance, an unstructured presentation may be a sign of conceptual problems in the logical framework of a research project.

» A well-structured and attractively presented document facilitates reading and understanding the proposal, even if the subject is complex and complicated, as is the case for many research proposals.

Miscellaneous

Abbreviations are usually used for generic names, but the full generic name should be given when the insect (or other animal, bacterium, or plant) is first mentioned anywhere in the proposal. Thus the mosquito would be referred to first as Anopheles hyrcanus var. sinensis, and later as A. hyrcanus var., or A. h. sinensis, or simply A. sinensis.

Several mosquitoes belong to genera having the same initial letter, and confusion is therefore possible if initials are used for them all. For instance, if Anopheles and Aedes are named in the same paper (as often happens), it is wise to refer to both genera in full throughout, even at the expense of much repetition. Similarly, if both Entamoeba coli and Escherichia coli were named in the same paper, it would obviously be wrong to use the contraction E. coli. In general, all Latin words should be written in italic letters.

Names of diseases derived from proper names should not be written with initial capital letters once they have become generally accepted: e.g., brucellosis, bilharziasis, leishmaniasis.

Measurements should always be given in metric units. Abbreviations of metric units should be written according to international standards.

Tables

Tables must be self-explanatory, although interpretations need to be explained in the text. Each table should begin with a heading consisting of the table number followed by the title. The table number must be referred to in the text. Tables should be numbered sequentially through the research proposal.

The title should summarize briefly the information in the table. Each column has to have a heading. Immediately beneath the column headings, the precise units of measurement of the data should be shown, if applicable (e.g. “%” or “years”). Units of measurement should be enclosed in parentheses. Longer units (e.g., “number of infant deaths per 1,000 live births and stillbirths”) should be put in a footnote to the table. In these cases, footnotes should be placed beneath the lower boundary of the table. However, the footnote should be at least two double-spaced lines above the following text to distinguish it clearly from the text.

If possible, the column headings should be concise so that they can be written horizontally. They may contain abbreviations. Although different table layouts can be chosen, with lines separating columns, rows, or headings, the same style should be maintained throughout the document.

As a rule, tables are presented vertically on the page, although wide tables may be presented sideways (landscape). If a vertical table is too long to fit on one page, it should be continued on the next page. At the bottom of the first page, the word “continued” should appear. The second should begin with the words “Table X continued,” and the column headings should be repeated. Such larger tables should probably appear in an appendix because they will contain more detail than is necessary for the points made in the text.

Figures

Various types of illustrations may accompany the research proposal, such as line drawings, graphs, maps, and photographs. Drawings should be presented clearly with india ink on the same white paper used for the text. Graphs that are developed by computer software should be printed with ink jet or laser printers. Photographs should be printed on glossy paper. All figures should be separated from the text and include the following elements:

» a heading, consisting of a sequence number in arabic numerals (preceded by “Figure”) and the subject heading or title;

» the figure itself;

» where necessary, footnotes or explanatory notes.

Citation of literature

In the text, references should be cited by author and year of publication. For more than two authors, use “et al.” For multiple citations in one year from the same author(s), use 1992a, 1992b, etc. If two or more references have authors with the same last name, use first and middle initials to identify them.

Common formats for citations in the text include the following:

Miller et al. (1992) investigated...

Other investigators (McColm et al, 1989; Rahmadalan et al., 1992) found...

A combination of iron status indices offers better accuracy in detection of iron deficiency than the use of a single assay (Cook et al., 1976b).

Styles for different types of citations in the literature cited section are as follows:

Article in journal

Weiser, N.W., Sumhadi, W. & Chen, R. (1992) Title of article. Journal 11:111-113.

Book

King, J.C. (1992) Title of Book. Publisher, City, State (or Country).

Chapter in book

Brown, K. & Dewey K.G. (1992) Title of chapter. In: Title of Book (Finley, D.A., ed.), vol. 2, pp. 1-20. Publisher, City, State (or Country).

Abstract or letter to the editor

Lembke, J.G. (1992) Title of abstract. Journal 11:111 (Abs. or Letter).

Thesis

Haskell, M. (1992) Title of thesis. Doctoral thesis, University Name, City, State (or Country).

Once the list of literature citations has been established, it should be rechecked against the citations in the text and vice versa. This check will not only confirm the completeness of the list but also identify references that should be deleted because they are not cited in the text.

It is advisable to use a computer programme to manage the references. These programmes store details of references and generate bibliographic citations, footnotes, or endnotes from the text file in over 1,000 different publishing styles. The programmes allow references cited in the research proposal to be included in the subsequent scientific article without retyping if the citation style is different from that described above. In addition, these programmes will sort references and copy information from other bibliographic sources.

Selected readings

Beach DP, Alvager TKE. Handbook for scientific and technical research. Englewood Cliffs, NJ, USA: Prentice-Hall, 1992.

Beghin I, Cap M, Dujardin. A guide to nutritional assessment. Geneva: World Health Organization, 1988.

CBE Style Manual Committee. Scientific style and format: the CBE manual for authors, editors, and publishers. 6th ed. New York: Cambridge University Press, 1994.

CIOMS. International guidelines for ethical review of epi-demiological studies. Geneva: Council for International Organizations of Medical Sciences, 1991.

Kirkwood BR. Essentials of medical statistics. Oxford: Blackwell Scientific Publications, 1988.

Lefèvre P, Beghin I. Guide to comprehensive evaluation of the nutritional aspects of projects and programmes. Health and Community Working Paper No. 27. Antwerp, Belgium: Institut de Médicine Tropicale “Prince Leopold,” 1991.

Locke LF, Spirduso WW, Silverman SJ. Proposals that work: a guide for planning dissertations and grant proposals. Newbury Park, Calif, USA: Sage Publications, 1987.

Martin SW, Meek AH, Willeberg P. Veterinary epidemiology: principles and methods. Ames, la, USA: Iowa State University Press, 1987.

Mauch IE, Birch JW. Guide to a successful thesis and dissertation: conception to publication: a handbook for students and faculty. 2nd ed. New York: Marcel Dekker, 1989.

Meador R. Guidelines for preparing proposals. 2nd ed. Chelsea, Mich, USA: Lewis Publishers, 1991.

Sultz HA, Sherwin FS. Grant writing for health professionals. Boston, Mass, USA: Little, Brown & Co, 1981.

Appendix. Metaplan technique

If SHARP is implemented by a team, the Metaplan technique should be used. The key approach of this technique is participation, visualization, and documentation. The objective of this technique is to allow input from all individuals on the team.

During a brainstorming time of 10 to 20 minutes, every participant writes ideas on several cards, one idea per card. During this time no discussion is allowed and everybody works by themselves. The specific steps of the process are as follows:

» Write an idea on a card;
» Formulate only one idea per card;
» Use a thick pen to write on the card;
» Write in printed letters as large as possible;
» Write no more than three lines on each card. After this session:
» All cards are collected;
» The written text of the first card is read aloud to all participants;
» After it has been read, the card is pinned to a board;
» The second card is read and pinned to the board, and so forth.