Tools Used in Scientific Research

Since science is a particular way of knowing about the world based on observation and measurement, the Science disciplines, both pure and applied, have some similarities to a foreign language. There is a culture -- beliefs, values, and very specific ways of doing things -- that seems different, even strange, to those who don't speak the language. The language includes unique names, a different way of doing things, and a different way of communicating. Acronyms and abbreviations can be particularly troubling. Scientists assume that those in their field will understand the abbreviation because it is common to them. "Precise technical terms and their definitions are formally recognized, documented, and taught by educators in the field." Scientists need to communicate with precision and as few words as needed. Note too that the same term may have different meaning in different scientific fields.

Select a scholarly scientific periodical from the current periodicals shelf or a website. Browse in a random article. How easily is the content understood? How many words or phrases would need to be defined?

Names and Naming

Each academic discipline and profession has its own unique vocabulary and this creates problems for laypeople or those in other fields. The vocabulary of scientific fields is doubly difficult because of the use of a dedicated vocabulary for both the discipline and also for the applied mathematics used in measurement and testing. For those outside science or outside the discipline or profession, the technical or specialized vocabulary "becomes an impenetrable forest of abstract or unknown terms." "Jargon" is often used in commenting about the overly technical and complex language used in some scientific communitation.

A good example of scientific naming are the taxonomies developed for naming living organisms. Perhaps the best known is that of Carl Linnaeus which goes from the general to the specific:

1. Kingdom
2. Phylum
3. Class
4. Order
5. Genus
6. Species.

Modern man = animalia, chordata, mannalia, primates, hominidae, homo, sapiens. Of course, the scientific name is often notably different from the popular name which may vary from place to place even within the same country. Standardized names allow scientists to unambiguously identify an item regardless of language or locality.

There are many examples of words and phrases that have specific scientific meaning and a specific relationship to other words. Here are a few:

• A law is an generalized explanation of a natural world event or relationship under stated circumstances. It is accepted to be true based upon testing and considerable experience but might not be supported at some time in the future.
• A hypothesis is a reasonable assumption of a natural world event that can be tested. However, it has not been tested rigorously.
• A theory is a complex explanation of natural world events based upon considerable testing and verification.
• Significant is a finding that can be reasonably generalized to a larger population. It is not a synonym for important.

A good way to become more familiar with the vocabulary of a particular science or a cluster of sciences is to browse in scientific journals [especially the abstracts], talk with those involved in science [scientists, assistants, students], and read and view authoritative but popular accounts of scientific research.

Numbers and Mathematics

While there is some controversy about whether or not mathematics is a science [perhaps a science of patterns], there is  no question that it is the single most important tool in science. Science is the process of measured observation. Measurement requires the use of numbers or applied mathematics. Mathematics focuses upon quantity, structure, space, and change. Each of these elements is fundamental to scientific inquiry. As with the sciences, math may be pure [math for its own sake] or applied. Interestingly, the Greek root of "mathematika" means study, learning, and science so it really means science. As with the scientific disciplines, math has its own technical vocabulary as does its sub discipline of statistics. 

The science information professional does not need to be a mathematician, but he does need to be comfortable in understanding and to being able to collect and analyze scientific data using fairly complex math. A course or two in statistics would be helpful. Again, browsing abstracts in scholarly scientific articles should quickly identify the most frequently used statistical tests and procedures. You can then learn a bit more about each via the web, a colleague or a standard textbook.

Before 500 CE, Babylonians used math to describe and predict astronomical data. By 1500, it was common for scientists to believe that "mathematics is the door and key to the sciences" [Roger Bacon]. In 1900, David Hilbert said that "mathematics is the foundation of all exact knowledge of natural phenomena."

Some have said the math is the language of science. Is math a language that you can speak or read? If not, how would you gain competence?

Statistics and Statistical Tests

Statistics is the best known of the applied mathematics. Observation of natural phenomena requires counting and the ability to summarize collected data. When attempting to discover causal relationships, observed relationships between variables must be measured.

One of the challenges of scientific research is to discover if findings are generalizable to larger populations. Results that are generalizable are "significant." Statistical tests hope to avoid two errors: [1] finding a relationship in a sample not present in the population and [2] not finding a relationship in the sample present in the population. Statistical tests or statistical hypothesis testing is applied by statistical software packages such as SAS. The software does the statistical analysis so you do not need to how how to compute. However, and this is crucial, you do need to know which tests are appropriate for your data and why.

How would you do this?

Understanding research publication requires some understanding of the language of statistical tests. For example, you need to know that data comes in different flavors. Nominal data have no numerical attributes other than naming a value and the mode is the average here. Ordinal data may be ranked and the median is the appropriate average. Parametric means that interval data were used. Non-parametric means that ordinal or nominal data were used. Interval data have arithmetic properties as seen in the mean. As is true of pure and applied mathematics, statistics has its own vocabulary so you would need to be able to define in your own words such terms as "standard deviation," "margin of error," and "confidence level."

Sample size and sampling are important considerations in selecting statistical tests so it's important to be able to evaluate the adequacy and typicality of a given sample. This also involves some consideration of how the data was collected. Data is valid if it really measures what is supposed to be measured. Data is reliable if the same response would be gathered by another or at another time.

Integrity in Scientific Journal Publications

After the leading journal Science published two papers on cloning and stem cells by Woo Suk Hwang that included manipulated and exagerated evidence, a variety of scientists and others noted that journals needed to adopt strong procedures to detect and filter out work that is misleading or falisified. The quality control system traditionally used [editors, and subject reviewers], may no longer be effective, especially when the stakes for successful publishing can be substantial in  recognition and monetary reward. In particular, it is suggested that journals need to do a better job of "risk assessment" and ask authors to provide more particulars about their work, including procedures and primary data. 

Another aspect of integrity involves externally funded research. The funder may restrict sharing of method and results or only allow selective reporting of results which may create false generalizations.

We know that research misconduct is a problem, but no one seems to know how serious it is. It includes mistreatment of research subjects, falsification and fabrication of data, and piracy and plagiarism. Correcting the literature after misconduct is discovered can also be a problem.

If you were the editor of a scientific periodical, how would you reduce the risk of publishing bad science?

How To Share Results

Regardless of the format scientific research reporting requires:

1. Clarity [report must be clear and unambiguous throughout]
2. Conciseness [use as few words and illustrations as possible while helping the audience to understand]
3. Continuity [content must flow logically and in an expected manner]
4. Ojectivity [the evidence speaks without exaggeration or manipulation] [fact clearly separated from interpretation/opinion].

Note that this must be used in terms of the likely audience, i.e. clarity for the chemist or clarity for a manager or a budget officer, but not every man.

Oral Presentations

Results, whether tentative or final, may be shared in various ways. Presentations may be oral. Oral presentations may be published as part of conference proceedings. Oral presentations may also be based on published papers. Normally, oral scientific presentations are based on PowerPoint slides or [in the older days] transparencies. Such presentations may be given:

1. as part of the proposal process
2. when the proposal has been funded
3. as a progress report
4. as tentative or early draft findings
5. as primary disclose with the final findings.

Presentations may be invited or the author may ask to present. Presentation at scholarly meetings is an important way to gain recognition and status in the field.

There is a substantial literature on how to effectively use PP slides to communicate clearly and interestingly. This literature suggests that it is best to avoid bulleted lists and substitute short phrases and sentences connected to appropriate, interesting illustrations. Some one on the research team or in a support position needs to be highly competent in preparing slide sets. One could argue that PP presentations with their standard format is a language of sorts.

Poster Sessions

Another presentation medium is the poster session. A poster, limited to a standard size, is prepared and presented to an audience walking through a room or halls. The presenter stands near her poster, answering questions and being approachable. The poster must be limited to a few key elements with a highly visible title, abstract, and appealing illustrations or images. Usually, the poster is arranged in the same order and with the same elements as the research article, but with brief summaries only for each element. Newer, less seasoned researchers often use poster sessions to gain recognition for their work and help with networking.

Proposals

Proposals are persuasive documents intended to secure funding for scientific research. They may be aimed at an internal audience such as the dean of a college or an external audience such as the National Science Foundation. Almost always, there is an announced request for proposals [RFP] with clearly specified deadlines and format elements and requirements. The proposal provides a strategy for solving a notable problem of interest to the funding agency. The strategy includes clear statements of the needed resources, the time and effort needed, the skill and experience of those involved, and details on expenditures including cost-sharing.

Research Reports

As with oral presentations, research reports [may be called technical report] may come at different times in the research process. Accountability is a major issue in research funding and reports are a device to insure that the right things are done on time.

Typically, there are interim reports that provide evidence of progress and final reports that discuss what was done and what was learned. Reports may also be informal or formal, but here we are only concerned with the formal ones. Reports are usually related to funding or oversight so that the style and format are prescribed by whoever funded or requested the research. The major elements are those that appear in the research article but with much more emphasis on accountability. For example, reports may ask for detailed reports on time, expenditures, and effort by those involved in the research, as well as objectives and deadlines met and those not met. Interim reports are normally proprietary, but final reports are often not.

Because they are published in-house, reports are normally longer and contain much more detail than a published article. At the same time, they do not go through a quality control process [no external subject expert readers]. Reports are much more likely to contain extensive back matter or appendices with details on method and less important findings.

An executive summary is a short highlights version of a longer research report. Typically, it is oriented toward decision-makers rather than scientists so results are often somewhat simplified and related to larger corporate or governmental goals and objectives.

Solid tips on writing a scientific technical report may be found at <http://grcpublishing.grc.nasa.gov/editing/CHP1.CFM>. NASA typically requires these report elements:

1. Summary
2. Introduction
3. Analysis
4. Apparatus [or Apparatus and Procedure to include the next two steps]
5. Test Specimens
6. Test Procedure
7. Results and Discussion
8. Summary of Results
9. Conclusions
10. Concluding Remarks
11. Appendix
12. References.

Note that the elements above differ somewhat from those found in the research periodical article.

The Glenn Research Center at NASA suggests that technical reports be evaluated considering these points:

1. Clear purpose and scope statement? Purpose met?
2. Does the introduction provide the background needed to understand the study methods, findings, and discussion?
3. Will major findings be understood by the reader?
4. Study limitations clearly stated and easily understood?
5. Facts clearly separated from interpretation/opinion?
6. Are references properly constructed and complete?

Although such reports are much more widely available today than before [via websites], bibliographic control of technical reports may be poor. NTIS and OSTI are good examples of agencies that aggregate technical reports prepared by government agencies. Going directly to the funding agency may also provide access to research resports. Try a subject search via USA.gov. 

Research Article

Introduction

The most common way of sharing research results is via the scientific paper published in a scholarly journal or a technical report. There is no single standard style [structure, use of illustrations, language] and format [arrangement of content elements] for either pure or applied science. The most common approach is to use the style and format seen in  the discipline or profession-specific journal where, it is hoped, the article will be published. Reviewing typical recent issues and detailed examination of instruction to author guidelines should yield an article that is publishable [if the science is good and the results represent an original contribution].

The scientific researcher must consider structure, audience, and style. Structure, such as use of standard headings and subheadings, and style are normally determined by others [journal editors and editors of technical report series]. The publication outlet also makes the audience or typical reader clear so that the writer knows, for example, that most readers will be serious researchers rather than managers or senior administrators. The relatively standard arrangement of elements as mentioned below makes it easy for scientists to follow and compare research study results.

When non-scientists need to read and understand scientific research findings, they may need help with terminology, including abbreviations and initialisms.

Title and Abstract

The title must be succinct, inform, and provide the key words  [descriptors] needed for intellectual access. A main title and a subtitle may be used. Both titles and abstracts must contain no unnecessary words.

The abstract is a succinct summary of the research paper. It allows the reader to learn:

• why the research was done and what it intended to do
• how the study was conducted
• what was discovered
• so what.

Many readers will only read the abstract and virtually all will read the abstract first so it is genuinely important. Abstracts are  normally written by the researcher after the paper has gone through a few drafts and the results are well ordered and clearly understood.

Introduction

The introduction section, sometimes called introduction and problem statement, answers two questions:

1. What is the problem being investigated, including its larger context?
2. Why is the problem important?

In some cases, this section will also include the literature review. In other cases, the literature review will be a separate section. The literature review answers these questions:

1. What do we already know about the problem?
2. What are the strengths of previous research?
3. What are the gaps in previous research?
4. What methods have been successful and most commonly used?

The literature review may also be used to validate the importance of the problem and the need for further research. At the end of the literature review, the author states her research objectives, usually as a hypothesis but sometimes as a research question [we hypothesized that mutilation would decline when the cost of photo duplication is reduced]. This is the transition to the next section.

Methods

Since scientific research must be able to be replicated or falsified, the methods section and the ability to contact the author to gain more methods specifics is essential. This section clearly indicates in a narrative how the research was conducted and the various steps taken. Standard methods require minimal explanation, especially if used by others who have examined this problem. We assume that readers will also be scientists so they should be familiar with the research processs, including statistical tests. For example, you don't need to explain a "null hypothesis."

Results

This section may also be called "findings." Research results are normally divided into major and minor findings. Major findings are reported. They are not interpreted or evaluated here. There is no "so what" since that normally is in the next section. Summary level data may be presented in narrative text, as a table or chart or as an illustration [figure, diagram, graph or photograph]. The same findings should not be duplicated in the text and then in an illustration. Both tables and illustrations should make relationships easier to understand. Each table and illustration is numbered, includes a succinct, informative title and should be understood without reference to the surrounding text. Similarly, the text should be understood without reference to the illustrations.

Discussion

The purpose of the this section is to interpret the results and clearly indicate what they mean and why they are important. Since most scientific research is build upon a hypothesis, this is where the researcher discusses the degree to which the hypothesis was supported and what that means. The present research needs to be related to the earlier research for similarities and differences in findings.

This is also the place to discuss problems encountered in the research [and how they impact your ability to generalize] , what might be done differently next time, and what research should follow.

Literature Cited

The scientist knows that research is built upon the work of others. Ideally, the researcher would read, digest, and evaluate all relevant research available on the problem area. Citation means that the author thoughtfully reviewed the work, understood it, and found it to be useful. Thus, each citation is a vote of confidence in the utility of a given item. We also assume that the author conducted a rigorous search for relevant literature and did not merely rely on literature that was convenient and easily found.

Resources and Examples

CHEMINFO at Indiana University has a good list of Science Writing Aids with some emphasis on chemistry. For example, sixteen sources are listed as "chemincal nomenclature aids."

The Council of Science Editors created an excellent white paper on "promoting integrity in scientific journal publications.