Food Science & Technology · Postgraduate & Research · also Research Methodology, Technical Writing, Research Ethics
Planning and reporting research: framing a question, reviewing literature, designing experiments, managing data, writing papers and theses, peer review, research ethics including human sensory trials, and intellectual property.
In the exams: PG and PhD coursework; ICAR AICE-JRF/SRF research-orientation questions
Research turns questions into reliable knowledge. In food science that knowledge decides how processes are designed, which ingredients are approved, what health claims are allowed and which practices protect consumers, so research that is poorly planned, badly analysed or dishonestly reported can do real harm. Postgraduate students and early-career researchers therefore need, alongside their technical training, a working command of how research is planned, carried out, recorded, reported, reviewed, protected and funded.
This subject covers framing research questions and hypotheses; searching and reviewing the literature systematically; planning experiments with proper replication; managing data so that work is reproducible; the ethics of studies with human participants, including sensory and nutrition trials; animal ethics and alternatives to animal use; scientific writing for papers, theses and reviews; presenting research as talks and posters; peer review and publication ethics; intellectual property and technology transfer; and research funding and proposal writing. Statistical design and analysis are taught in detail in statistics-experimental-design-food, and practical project skills at undergraduate level in industrial-training-project; this subject brings them together at the level of a master's or doctoral researcher.
Research methodology is part of postgraduate and doctoral coursework in most universities and is examined in research-fellowship and teaching-eligibility examinations, which include questions on research design, hypotheses, sampling, ethics, publication and intellectual property. The practice questions follow that pattern and include short calculations of sample size and statistical power, and scenarios on ethics, authorship and patents.
Research is classified in several overlapping ways. Basic research seeks understanding without immediate application, such as the mechanism of starch retrogradation; applied research seeks solutions to practical problems, such as a process that slows the staling of bread; and development turns results into products and processes. Descriptive research describes what exists, for example the composition of traditional cheeses in a region; analytical research explains why, testing relationships between variables. Observational studies measure without intervening, as in surveys of food consumption or cohort studies of diet and disease; experimental studies deliberately change a factor and measure the effect, as in a trial of a new preservative or a randomised controlled trial of a dietary intervention. Quantitative research measures and counts; qualitative research explores meanings and experiences through interviews and focus groups, as in studies of why consumers reject a new food; mixed methods combine both.
A good research question is clear, focused and answerable with the available time and resources. The FINER criteria summarise what makes it worth asking: feasible, interesting, novel, ethical and relevant. For nutrition and health studies, the PICO framework structures the question: population (for example adults with type 2 diabetes), intervention (a high-fibre breakfast cereal), comparison (a low-fibre cereal) and outcome (post-prandial blood glucose). For food processing studies, a similar logic identifies the material, the treatment, the control and the response variables, such as 'in pasteurised orange juice, does high-pressure processing at 600 MPa retain more vitamin C over eight weeks of chilled storage than thermal pasteurisation?'. Writing the question precisely is often the hardest and most valuable step of a project, because it decides the design, the measurements and the analysis.
A hypothesis is a testable prediction derived from theory or previous evidence. In statistical testing, the null hypothesis (H₀) states that there is no effect or difference, and the alternative hypothesis (H₁) that there is. For the juice example, H₀ is that mean vitamin C retention is the same after both treatments, and H₁ that it differs (two-sided) or that it is higher after high-pressure processing (one-sided). Scientific hypotheses must be falsifiable, in Popper's sense: they must make predictions that observations could contradict. A study can reject the null hypothesis or fail to reject it; it does not 'prove' the alternative, and failing to find a difference is not proof that none exists.
Variables are classified as independent (the factors manipulated or compared, such as treatment), dependent (the responses measured, such as vitamin C), and controlled or confounding variables (other factors that could influence the response, such as storage temperature, raw-material batch or analyst). A confounder is associated with both the factor of interest and the outcome, so it can create a false association or hide a real one; designs control confounders by holding them constant, randomising, blocking or measuring and adjusting for them. Two kinds of error are possible in hypothesis testing. A type I error is rejecting a true null hypothesis, a false positive, with probability α, the significance level, conventionally 0.05. A type II error is failing to reject a false null hypothesis, a false negative, with probability β. Statistical power, 1 − β, is the probability of detecting a real effect of a given size; studies are commonly designed for a power of 0.8 or 0.9. Under-powered studies waste resources and often produce misleading results, which is why sample size is planned before data are collected (chapter 3).
Reviews of the literature range from narrative reviews, in which an expert summarises a field selectively, to systematic reviews, which use explicit, reproducible methods to identify, select, appraise and synthesise all relevant studies on a defined question. Scoping reviews map the extent and nature of research on a broad topic, and rapid reviews simplify systematic methods for urgent decisions. Systematic reviews and meta-analyses sit near the top of evidence hierarchies for questions of effect, because they reduce the bias of selective citation, and they underpin many regulatory and guideline decisions, such as evaluations of health claims and dietary recommendations.
A systematic review begins with a protocol that states the question, inclusion and exclusion criteria, databases, search strategy, screening procedure, data to be extracted, risk-of-bias assessment and synthesis methods; protocols for health-related reviews are often registered publicly in advance, which discourages changing methods after seeing results. The search combines subject headings and free-text terms for each concept, joined by Boolean operators (OR within a concept, AND between concepts), with truncation and phrase searching, run in several databases chosen for the topic: multidisciplinary citation indexes, food-science and agricultural abstracts databases, biomedical databases, and others. Grey literature, such as theses, reports of government agencies and international organisations, conference abstracts and trial registries, reduces publication bias, the tendency for studies with positive results to be published more often. The full search strategy is reported so that others can repeat it.
Records are de-duplicated and screened, first by title and abstract and then by full text, usually by two reviewers working independently, with disagreements resolved by discussion or a third reviewer. The PRISMA 2020 statement (Preferred Reporting Items for Systematic reviews and Meta-Analyses) provides a checklist of items to report and a flow diagram showing how many records were identified, screened, excluded (with reasons) and included. Data are extracted with standard forms. Each included study is appraised for risk of bias with tools suited to the design: for randomised trials, domains such as the randomisation process, deviations from intended interventions, missing outcome data, measurement of the outcome and selection of the reported result; for observational and laboratory studies, other tools and criteria.
Synthesis may be narrative, describing and comparing results systematically, or quantitative. Meta-analysis combines effect estimates from comparable studies, weighting each by its precision, usually the inverse of its variance. A fixed-effect model assumes one true effect common to all studies; a random-effects model assumes the true effects vary between studies and estimates their average, and is usually more appropriate when studies differ in populations, foods or methods. Heterogeneity is examined with statistics such as I², the percentage of total variation across studies due to heterogeneity rather than chance (values around 25, 50 and 75 per cent are often described as low, moderate and high), and explored by subgroup analysis or meta-regression. Funnel plots, which plot effect size against precision, can suggest publication bias when small studies with null results are missing. Approaches such as GRADE rate the overall certainty of evidence for each outcome as high, moderate, low or very low, considering risk of bias, inconsistency, indirectness, imprecision and publication bias. A good review states its limitations and avoids conclusions stronger than the evidence allows.
The experimental unit is the smallest unit to which a treatment is independently applied: a batch of product, a vat of cheese, a pot of soil, an animal or a person. Replication means applying each treatment to several independent experimental units; it allows the variability between units to be estimated, which is the basis of statistical tests. Repeated measurements on the same unit, or subsamples taken from it, are not replicates of the treatment; treating them as replicates (pseudoreplication) underestimates variability and inflates the apparent significance of results, one of the most common errors in food research. For example, making one batch of bread with each of three flours and measuring the volume of ten loaves from each batch gives one replicate per flour, not ten; true replication needs separate doughs prepared independently.
Controls give a reference: a negative control without the treatment, a positive control with a known effect (showing that the method can detect an effect), and, in sensory and nutrition studies, a placebo or comparator product. Randomisation allocates treatments to units and determines the order of processing and measurement by a random procedure, so that unknown factors are spread evenly and cannot systematically favour one treatment. Blocking groups units that are similar, such as batches made on the same day or with the same lot of raw material, and applies all treatments within each block, removing that source of variation from the error. Blinding, keeping the people who measure or assess the outcomes, and in human studies the participants, unaware of which treatment each unit received, prevents conscious and unconscious bias, and is standard practice in sensory tests and clinical trials. Factorial designs study several factors and their interactions at once, and response-surface designs find optimum conditions (see statistics-experimental-design-food).
The number of replicates should be chosen before the experiment, from the size of difference that matters, the expected variability, the significance level and the power. For comparing two means with a two-sided test, an approximate number per group is n = 2(z₁₋α/₂ + z₁₋β)²σ²/Δ², where σ is the standard deviation and Δ the difference to be detected. With α = 0.05 (z = 1.96), power 0.80 (z = 0.84), σ = 10 units and Δ = 10 units, n = 2 × (2.80)² × 1 ≈ 15.7, so 16 units per group are needed; to detect a difference half as large would need four times as many. Software and tables give more exact values for different designs. Where no estimate of variability exists, a pilot study provides one and also tests methods, logistics and measurements before the main experiment.
Validity has two aspects. Internal validity is the degree to which the design supports a causal conclusion about the factor studied, protected by randomisation, controls, blinding and the control of confounders. External validity (generalisability) is the degree to which results apply beyond the study conditions: results from one cultivar, one season, laboratory-scale equipment or a narrow group of volunteers may not hold for others, so the range of conditions should match the intended application, and limitations should be stated. Scale-up from bench to pilot to industrial scale often changes heat and mass transfer, mixing and residence times, so process results need confirmation at larger scale. Planning also includes the measurements and their methods, validated and calibrated; the statistical analysis, specified in advance (a statistical analysis plan for trials); the schedule and resources; risk assessment; and ethical approval where needed.
Research data include raw instrument outputs, laboratory notebooks, images, spectra, sensory and survey responses, code and the processed data behind each figure and table. Many funders and institutions now require a data management plan describing what data will be collected, in what formats, with what documentation and metadata, how they will be stored and backed up, who will have access, how personal data will be protected, how long data will be retained, and how they will be shared at the end of the project. The FAIR principles, published in 2016, state that data should be findable (with persistent identifiers and rich metadata), accessible (retrievable by standard protocols, with access conditions stated), interoperable (in common formats and vocabularies) and reusable (with clear licences and provenance). FAIR does not mean that all data must be open; personal and commercially confidential data may be shared under controlled access or not at all, but their metadata can still be findable.
Good everyday practice prevents loss and confusion. Files are named consistently, with dates in a sortable format and version numbers; folders are organised by project and type; raw data are kept unchanged, in read-only form, and all processing is done on copies; laboratory notebooks, paper or electronic, record dates, methods, observations and deviations at the time; and metadata describe what each variable means, its units and how it was measured. Backups follow rules such as the 3-2-1 rule: at least three copies, on two different types of media, with one copy off-site or in an institutional cloud service. Data are retained for the period set by the institution or funder, often several years after publication.
A result is reproducible when the same data and analysis give the same result, and replicable when a new study, with new data, finds a consistent result. Surveys across many fields have found that a substantial proportion of published results could not be replicated, a concern often called the reproducibility crisis, attributed to small samples, flexible analysis, selective reporting, publication bias and poor documentation. Food science shares these problems, for example in studies of bioactive compounds and health effects with small, short trials.
Practices that improve reproducibility include pre-registration of hypotheses, designs and analysis plans before data collection, which separates confirmatory from exploratory analyses; scripted analysis in languages such as R or Python instead of manual steps in spreadsheets, so that every step from raw data to figure is recorded and can be rerun; version control systems that track changes to code and documents; reporting guidelines that require full description of methods; sharing of data, code and materials in recognised repositories, which give persistent identifiers such as digital object identifiers; and publishing all results, including null results. For laboratory methods, detailed protocols, sometimes published separately, and the use of reference materials and inter-laboratory comparisons make results comparable between groups. These practices are increasingly required by journals and funders, and they also protect researchers: a well-documented project survives the departure of a student, the failure of a hard disk or a query from a reviewer years later.
Research with human participants is governed by ethical principles developed after abuses in the twentieth century. The Nuremberg Code (1947) established that voluntary consent is essential. The World Medical Association's Declaration of Helsinki, first adopted in 1964 and revised several times (most recently in 2024), sets out principles for medical research involving human participants, including review by an independent research ethics committee, informed consent, the priority of participants' wellbeing over the interests of science and society, special protection for vulnerable groups, and registration and publication of research. The Belmont Report (1979) summarised three principles: respect for persons (autonomy and informed consent, with protection for those with diminished autonomy), beneficence (maximising benefits and minimising harms) and justice (fair distribution of the burdens and benefits of research). National laws and institutional rules put these principles into practice, usually requiring approval by a research ethics committee or institutional review board before any study with human participants begins.
Food research with people includes sensory evaluation and consumer tests, questionnaires and interviews, dietary surveys, feeding studies and nutrition intervention trials, and clinical trials of foods or ingredients with health effects. The ethics committee considers the scientific value of the study, the risks and burdens to participants, how they are recruited, the information given and how consent is obtained, protection of privacy and data, payment and its potential to coerce, and arrangements for adverse events. Low-risk studies such as many sensory tests may go through expedited review, but they still need approval and consent; studies that involve novel foods, allergens, vulnerable groups such as children or pregnant women, or invasive measurements receive full review.
Informed consent requires that participants receive, in language they understand, information about the purpose of the study, what they will be asked to do, how long it will take, foreseeable risks and discomforts, possible benefits, how their data will be used and protected, payment, who to contact, and their right to refuse or withdraw at any time without penalty; consent is voluntary and documented, usually by signature. For children, parents or guardians give permission and the child gives assent where able. Participants are not deceived about risks, and any necessary deception, rare in food research, must be justified and followed by debriefing.
Safety is central in food studies. Samples served to participants are prepared hygienically from ingredients permitted for food use, or with specific approval for novel ingredients, and participants are screened for allergies, intolerances and dietary restrictions relevant to the samples; ingredient lists are available. Personal data are collected only as needed, stored securely and handled according to data-protection law. Nutrition intervention trials that test health effects are increasingly expected to follow the standards of clinical research: registration in a public trial registry before the first participant is enrolled (international registries are linked through the WHO International Clinical Trials Registry Platform), a protocol and statistical analysis plan, randomisation and blinding where feasible, monitoring and reporting of adverse events, good clinical practice guidelines, and reporting according to the CONSORT statement for randomised trials or STROBE for observational studies. Health claims on foods in many jurisdictions are assessed against evidence from well-conducted human studies, so poor design or conduct can make an expensive trial useless for its purpose. Fair recruitment, avoiding exploitation of students or employees who may feel pressure to take part, and sharing results with participants where appropriate complete ethical practice.
Animals have been used in food and nutrition research to test the safety of additives, contaminants and novel foods, to measure protein quality and nutrient bioavailability, and to study diet and disease. Their use raises ethical questions because animals can suffer, and it is regulated in most countries. The framework accepted internationally is the 3Rs, proposed by Russell and Burch in 1959: replacement (using methods that avoid or replace animals, such as in-vitro, in-silico or human studies), reduction (using methods that obtain comparable information from fewer animals, or more information from the same number, through good design and statistics), and refinement (methods that minimise pain, suffering and distress and improve welfare, such as better housing, handling, anaesthesia, analgesia and humane endpoints). Legislation such as the European Union's directive on the protection of animals used for scientific purposes, and national laws and guidelines in other regions, require that projects be authorised after ethical evaluation, that staff be trained and competent, that facilities meet standards, and that the 3Rs be applied.
Institutional animal ethics committees (called animal care and use committees, animal welfare and ethical review bodies or similar) review proposals before work begins, weighing the expected scientific benefits against the harms to animals (a harm-benefit analysis), checking that alternatives have been considered, that the number of animals is justified by a sample-size calculation, that procedures and humane endpoints are defined, and that the species chosen is appropriate. Regulatory safety studies of food chemicals, where animal studies are still required, follow internationally harmonised test guidelines, such as those of the OECD, which reduce duplication because data accepted in one country are accepted in others.
Food science has developed many alternatives that replace or reduce animal use. Standardised in-vitro digestion models simulate the mouth, stomach and small intestine with defined enzymes, pH, electrolytes and times; the internationally harmonised static protocol developed by the INFOGEST network is widely used to study protein, lipid and starch digestion and the release (bioaccessibility) of nutrients and bioactive compounds, and dynamic models add mechanical action and gastric emptying. Cell cultures model absorption and effects at the intestinal barrier, for example Caco-2 cells, a human colon cancer line that differentiates into enterocyte-like cells, used for transport and bioavailability studies, often combined with in-vitro digestion. Organoids and organ-on-a-chip systems give more realistic tissue models. In-silico methods, such as quantitative structure-activity relationship models, read-across from related substances, and physiologically based kinetic modelling, support safety assessment. For protein quality, chemical scoring corrected for digestibility (PDCAAS and DIAAS, see general-biochemistry) has largely replaced the rat growth assay of the protein efficiency ratio for human nutrition purposes, although digestibility values still rely partly on animal or human studies. Human volunteer studies, ethically conducted, give the most relevant information for many nutrition questions.
When animal studies are still needed, they are reported fully so that they can be evaluated and need not be repeated. The ARRIVE guidelines (Animal Research: Reporting of In Vivo Experiments, version 2.0, 2020) list essential items such as study design, sample size, inclusion and exclusion criteria, randomisation, blinding, outcome measures, statistical methods, and details of the animals, housing and procedures. Poorly designed and reported animal studies waste animals and produce unreliable results, so good methodology is itself an ethical requirement.
A research paper reports original work so that others can evaluate and build on it. Most follow the IMRaD structure: introduction (why the work was done, what was known, the gap and the aims or hypotheses), methods (what was done, in enough detail to repeat it), results (what was found, in text, tables and figures, without interpretation) and discussion (what the results mean, how they compare with previous work, their limitations and implications), preceded by a title and abstract and followed by conclusions, acknowledgements, funding and conflict-of-interest statements, data-availability statements and references. The title should be specific and informative; the abstract, often structured into background, methods, results and conclusions, should stand alone, because many readers read nothing else; keywords help indexing.
Reporting guidelines specify what must be reported for particular designs: CONSORT for randomised trials, STROBE for observational studies, PRISMA for systematic reviews, ARRIVE for animal studies, STARD for diagnostic accuracy studies, and others collected by networks that promote them. Clear scientific prose is precise, concise and logically ordered: one main idea per paragraph, the active voice where it is clearer, consistent terminology, defined abbreviations, SI units and correct significant figures, and claims that match the evidence. Tables present exact values; figures show patterns; neither duplicates the other. Writing is iterative: an outline, a first draft written quickly, revision for structure and argument, then editing for clarity and language, with feedback from co-authors and colleagues.
Authorship carries credit and responsibility. Widely used criteria from the International Committee of Medical Journal Editors require that each author has made substantial contributions to the conception or design of the work, or to the acquisition, analysis or interpretation of data; drafted the work or revised it critically for important intellectual content; approved the final version; and agreed to be accountable for all aspects of the work. All four criteria should be met; people who contributed in other ways, such as providing funding, general supervision, technical help or writing assistance, are acknowledged rather than listed as authors. Gift authorship (adding people who did not contribute) and ghost authorship (omitting those who did) are unethical. Contributor role taxonomies such as CRediT describe each person's role. The order of authors is agreed in advance, following the conventions of the field.
Choosing a journal involves its scope, audience, quality of peer review, speed, open-access options and indexing. Open access may be gold (published openly by the journal, often with an article processing charge), green (a version deposited in a repository) or diamond (open without charges to authors or readers). Predatory journals, which charge fees but provide little or no genuine peer review and editorial service, are avoided by checking indexing, editorial boards, review processes and publisher reputation. Preprints, posted publicly before peer review, speed dissemination but must be read with caution. A thesis is longer than a paper and demonstrates the candidate's mastery of the field: typically an introduction and literature review, aims, methods, results and discussion chapters (sometimes as a series of papers), a general discussion and conclusions; it must follow the university's regulations on format, originality and length. Review articles synthesise a field; systematic reviews follow PRISMA, while narrative reviews should still be balanced, critical and comprehensive enough to be useful.
Oral presentations at seminars, conferences and examinations reach audiences that may never read the paper. A good talk tells a story with one clear message: why the question matters, what was done, what was found and what it means. Time is the main constraint: a 12- or 15-minute conference slot allows the background in two or three minutes, methods in two or three, results in five or six, and conclusions in one or two, with time for questions. The audience usually includes non-specialists, so jargon is explained, the problem is set in context, and only the results that support the message are shown; complete data belong in the paper.
Slides support the speaker. Each slide has one main idea, ideally stated as a short headline sentence; text is minimal and large enough to read at the back of the room; figures are simplified for the screen, with large axis labels, few lines and colours that remain distinguishable for people with colour-vision deficiency; animations are used only to reveal information step by step. A common guide is about one slide per minute. Speakers rehearse aloud with a timer, know their opening and closing sentences, face the audience, speak clearly, and point to what matters on each figure. Questions are listened to fully, repeated if the audience may not have heard, and answered briefly and honestly, including saying 'I don't know' where appropriate. Online presentations need attention to camera, sound, screen sharing and keeping the audience engaged without visual cues.
A poster presents a study on one large sheet that visitors read in a few minutes while the author stands nearby. It is designed to be read from a distance: a clear title in large type, authors and affiliations, and short sections (introduction or aim, methods, results, conclusions) arranged in columns that the eye follows naturally, with results dominated by well-designed figures and minimal text. White space, consistent fonts and colours, and a highlighted take-home message help; a quick-response code can link to the full paper or data. The author prepares a two-minute explanation for visitors and a longer one for those who want detail. Short-format presentations, such as flash talks and three-minute thesis competitions, train researchers to state the essence of their work.
Communicating research beyond specialists is increasingly expected by funders and institutions. Food science touches everyone, so researchers talk to industry, regulators, farmers, consumers and journalists, and write plain-language summaries, press releases, blogs and social-media posts. Good public communication is accurate, avoids exaggerating findings (a single laboratory or animal study does not show that a food prevents disease in people), explains uncertainty, and distinguishes the researcher's results from wider evidence. Press releases are reviewed by the researchers to prevent overstatement, which studies have shown often originates in the release rather than the news story. Visual principles for scientific figures, such as showing the data, choosing chart types that match the comparison, starting bar axes at zero, and avoiding three-dimensional effects and unnecessary decoration, apply to both specialist and public audiences.
Peer review is the evaluation of a manuscript by independent experts before publication. An editor first checks whether the manuscript fits the journal's scope and standards (desk review) and may reject it without external review. Otherwise, the editor invites usually two or three reviewers, who assess the importance of the question, the soundness of the design and methods, the validity of the analysis and interpretation, the clarity of presentation and the adequacy of references, and recommend acceptance, minor or major revision, or rejection. Models differ: in single-blind review, reviewers know the authors' identities but not vice versa; in double-blind review, neither knows the other's identity; in open review, identities, and sometimes reports, are published. The editor makes the decision, weighing the reviewers' reports.
Reviewers have obligations: to accept only manuscripts they are competent to review and can review on time, to declare conflicts of interest (such as close collaboration or competition with the authors, or financial interests), to keep the manuscript confidential and not use its contents for their own work, and to write reports that are specific, constructive and courteous, distinguishing essential problems from minor suggestions. Authors respond to reviews point by point, explaining changes made or, politely and with evidence, why they disagree. Peer review improves papers and filters out many flawed studies, but it cannot detect all errors or fraud, and published papers continue to be scrutinised by readers after publication.
Research misconduct is commonly defined as fabrication (making up data or results), falsification (manipulating materials, equipment, processes, data or images so that the research is not accurately represented) and plagiarism (using others' ideas, words or results without appropriate credit). Beyond these, questionable research practices undermine reliability: p-hacking (trying many analyses until one gives a significant result), HARKing (hypothesising after the results are known and presenting the hypothesis as if it had been planned), selective reporting of outcomes, omitting inconvenient data without justification, salami slicing (splitting one study into many minimal publications), duplicate publication of the same results, inappropriate image manipulation, citation manipulation, and undisclosed conflicts of interest. Self-plagiarism, reusing one's own published text or data without disclosure, also breaches publication ethics.
Journals, publishers and institutions handle these issues through policies and procedures. The Committee on Publication Ethics (COPE) provides guidance and flowcharts for editors on suspected misconduct, authorship disputes, conflicts of interest and corrections. Problems discovered after publication lead to corrections (for honest errors that do not invalidate conclusions), expressions of concern, or retraction (when findings are unreliable because of misconduct or major error). Funding sources and competing interests, including support from food companies or trade associations, must be disclosed, because funding has been associated with more favourable conclusions in some areas of nutrition research; disclosure lets readers judge, and good design and transparency protect the credibility of industry-funded research. Institutions have research integrity officers and procedures for investigating allegations fairly, protecting both whistle-blowers and the accused. National and international codes, such as statements on research integrity agreed at world conferences, set out principles of honesty, accountability, professional courtesy and good stewardship that apply to all researchers.
Intellectual property (IP) rights give creators control over the use of their creations for a period, encouraging investment in innovation. Patents protect inventions, such as new processes, products, compositions and uses, that are new (novel), involve an inventive step (are not obvious to a person skilled in the field) and are capable of industrial application; in return for publishing the invention in full, the owner gets an exclusive right, usually for 20 years from the filing date, in each country where the patent is granted. Discoveries, scientific theories, natural substances as found in nature and, in many jurisdictions, methods of medical treatment are not patentable as such. Novelty is destroyed by prior public disclosure anywhere, including the inventor's own conference talks, posters, theses and papers; a few jurisdictions allow a grace period (for example 12 months in the United States) for the inventor's own disclosures, but most do not, so researchers must file before they publish. The Patent Cooperation Treaty allows a single international application that preserves the right to seek protection in many countries, with national phase entry usually at about 30 or 31 months from the earliest filing.
Other forms are important in food. Trade secrets protect confidential information, such as recipes, process conditions and supplier lists, for as long as it remains secret and reasonable steps are taken to keep it so; famous beverage and seasoning formulas are protected this way rather than by patents, which would expire and require disclosure. Trademarks protect brand names, logos and other signs that distinguish goods, renewable indefinitely while in use. Copyright protects original expression, such as text, images, software and databases, automatically. Industrial designs protect the appearance of products and packaging. Plant variety protection, under the UPOV convention in member countries, protects new plant varieties for breeders. Geographical indications protect names of products whose qualities are linked to their origin, such as many cheeses, wines, teas and rices from particular regions.
Technology transfer moves knowledge and inventions from research organisations to users who can develop them into products and processes. Universities and research institutes usually have technology transfer offices that assess disclosures of inventions by researchers, decide whether to seek patent protection, market technologies to companies, negotiate licences and support spin-off companies. In many countries, laws or policies give institutions ownership of inventions made with public research funds, provided they seek to commercialise them, a model associated with the United States law of 1980 often called the Bayh-Dole Act and adopted in various forms elsewhere; institutions usually share income with inventors. Researchers should read their institution's IP policy and their funding and collaboration agreements, which determine who owns results.
Licensing grants another party the right to use IP under conditions: exclusive licences give one licensee sole rights (sometimes in a field or territory), non-exclusive licences allow several; terms include upfront fees, milestone payments, royalties on sales, diligence obligations and sublicensing rights. Collaborative research agreements with companies define background IP (what each party brings), foreground IP (what is created), publication rights (often allowing a short delay for patent filing) and confidentiality; material transfer agreements govern the exchange of research materials such as microbial strains or cell lines; and non-disclosure agreements protect confidential information in discussions. Before commercialising, a freedom-to-operate analysis checks that the product or process does not infringe others' patents. Technology readiness levels, from 1 (basic principles observed) to 9 (actual system proven in operation), describe how mature a technology is and help match research to funding and industrial partners. For food innovations, regulatory approval (for novel foods, additives or health claims), scale-up and consumer acceptance are often harder obstacles than IP, so commercial planning must address them early.
Research is funded by national research councils and science agencies, government ministries (agriculture, health, industry), international and regional programmes (such as the European Union's framework programmes for research and innovation, international agricultural research partnerships and development agencies), philanthropic foundations, industry and industry associations, and the researcher's own institution. Each funder has priorities, eligibility rules, budgets and evaluation criteria, published in calls for proposals; reading the call carefully and matching the proposal to it is the first step. Many funders expect collaboration across institutions, disciplines or countries, involvement of users such as companies, farmers or public bodies, and plans for impact and open access.
Proposals are typically evaluated on three broad groups of criteria: excellence (the importance and novelty of the objectives, the soundness of the concept and methods, and the track record and capability of the team), impact (the expected contributions to knowledge, the economy, society, health or the environment, and the plans for dissemination, exploitation and communication), and quality and efficiency of implementation (the work plan, resources, management structure, risk management and value for money). Reviewers often read many proposals quickly, so clarity, a logical structure and a compelling case in the first pages matter. Success rates in competitive programmes are often low, so rejection is common; reviewers' comments are used to improve resubmissions.
A research proposal usually contains a title and short summary (often written last, but read first); background and rationale, showing the problem, what is known and the gap; aims and specific objectives; hypotheses or research questions; methodology, describing the design, materials, methods, sample sizes with justification, data analysis, and ethical and regulatory considerations; a work plan divided into work packages or tasks, with a timeline (Gantt chart), deliverables (tangible outputs such as a report, a prototype or a dataset) and milestones (control points that show progress); a management plan and risk register, identifying risks such as recruitment delays or equipment failure with mitigation; expected outcomes and impact, with dissemination and data-management plans; a budget with justification; and supporting documents such as CVs, letters of support from partners and ethics approvals. For development-oriented projects, a logical framework (logframe) links activities to outputs, outcomes and goals, with indicators, means of verification and assumptions.
The budget covers direct costs, such as staff (researchers, technicians, students) according to time spent, consumables, equipment, travel, publication and open-access fees, subcontracted services (for example analyses by external laboratories) and participant payments, and indirect costs (overheads) for the institution's infrastructure and administration, calculated by the funder's rules, often as a percentage of direct costs. For example, direct costs of 400 000 currency units with an indirect rate of 25 per cent of direct costs give a total of 500 000. Each item is justified in relation to the work plan, without inflation or omission; reviewers penalise budgets that are unrealistic in either direction. Once funded, the researcher manages the project: monitoring progress against milestones, managing the budget within the funder's rules, reporting periodically, handling changes with approval, and publishing and sharing results. Writing proposals is a skill that improves with practice, feedback and reading successful examples.
A postgraduate student uses this subject to frame a thesis question with FINER and PICO, write a systematic literature search, calculate sample size, design a replicated, randomised experiment, prepare ethics applications for sensory or nutrition work, keep a data management plan and scripted analyses, write papers and a thesis to reporting guidelines, present at conferences and respond to peer review. Examination questions focus on hypotheses, errors and power, replication, ethics principles, authorship and patent criteria.
Five years after a doctorate, a researcher in a university, institute or company leads projects: writing proposals and budgets, supervising students, managing data and reproducibility, reviewing manuscripts, negotiating collaboration and material transfer agreements, disclosing inventions to the technology transfer office, and running human or animal studies under ethics approval.
A senior research leader with twenty-five years' experience sets research strategy, builds funding portfolios and partnerships, chairs ethics or integrity committees, edits journals, negotiates licences and spin-offs, and mentors researchers in good practice. In industry, a research director manages IP strategy and the evidence base for product claims and regulatory submissions.
A researcher at this level shapes the research system itself: developing methods for evidence synthesis and reproducibility in food science, advising funders and regulators on study standards for health claims, contributing to ethical guidelines and reporting standards, and leading large international consortia.
This manual introduces research methodology for food science researchers. It does not replace the statistics taught in statistics-experimental-design-food, the specific ethics rules and forms of any institution or country, or professional legal advice on intellectual property, which varies by jurisdiction. Patent terms, grace periods and procedures are described in general terms; funding programmes and their criteria change and must be read in each call. Qualitative research methods are mentioned but not taught in depth.
Original practice questions written to the syllabus and pattern of the exams named in each question's examTags. They are not past papers and do not reproduce any board's questions.
These notes are free to read here, with no account.
What the Laisha app adds:
Free for ten days, then a subscription. We would rather write that here than have you find it later.