What is science, what is research? Basic terms
General definitions
Science
The word ‘science’ comes from the Latin word scientia, which means knowledge. Science refers to systematic and organized knowledge in any area of investigation that has been obtained using the ‘the scientific method’. The scientific method is the best method we have, to obtain reliable data about the world, which helps both to explain and predict different phenomena. Science is based on observable and measurable things/phenomena. However, there is no absolute scientific truth; it is just that some knowledge is less likely to be wrong than others (Nayak & Singh 2015). Statements produced through scientific research must be testable, and research by itself must be reproducible (a good scientific paper is one which enables the method to be replicated).
Research
Research is defined as a scientific and systematic search for relevant information on a particular issue. In that case the term ‘research’ refers to the systematic method that includes articulating the problem, formulating a hypothesis, gathering the facts or data, analysing them, and drawing certain conclusions, either as a solution(s) to the investigated problem or as generalisations for some theoretical formulation. Research is termed ‘scientific research’ if it contributes to the pool of science and follows the scientific method.
Generally, research can be divided into two groups:
Basic research: the main goal is to acquire an organized body of scientific knowledge and not necessarily to generate results with direct practical impact. Basic research is about fundamental properties of objects, their relationship and their behaviour, which includes theoretical and experimental research.
Applied research: the main goal is to solve practical problems and the goal of contributing to the pool of scientific knowledge is secondary. Applied research is focused on the usefulness of objects and their behaviour, and improvements of technology.
Research vocabulary
Variables and levels of measurements
A variable is a measurable characteristic of an abstract construct. A variable is something that can have more than one value and can vary from negative to positive, from low to high, etc. It is the opposite of a constant. The values of a variable can be words (e.g. gender) or numbers (e.g. temperature). Constructs by themselves cannot be measured directly; therefore, scientists need to find substitute measures called variables. For example, water quality is often measured as nitrate and orthophosphate concentrations and chemical oxygen demand, which are different parameters gained from analytical laboratory procedures done on a water sample. In this case, water quality is a construct, and nitrate and orthophosphate concentrations and chemical oxygen demand are the variables that measure it.
Variables that describe other variables are termed independent variables, while variables that are described by other variables are dependent variables. In a research experiment there may be other variables that are not relevant for studying a selected dependent variable but which could have some impact on it. These variables must be controlled throughout the experiment and are termed control variables (e.g. pH and oxygen concentration in the case of water quality). In research we want to select specific variables and search for relations among them; moreover, we aim to understand if and how variation in one variable affects variation in another.
Different variables have different levels of measurement in ascending order: nominal, ordinal, interval, and ratio. For research it is important to always select variables with the highest level of measurement (Nayak & Singh 2015):
Nominal level of measurement: the values at this level include a list of names/words. Naming values is a qualitative measurement (e.g. vegetable species or varieties, colour of leaves). It is also possible to substitute the names of values with numbers (e.g. 1 for Boston Bibb, 2 for Red Leaf, 3 for Iceberg etc.); however, in this case the numbers only mean a different type of name, and do not make the variable quantitative. Giving numbers to characteristics facilitates statistical analyses of qualitative data. The statistical analysis of central tendency of nominal measurements is mode; mean or median cannot be defined (it is not possible to calculate an average sex or colour). Appropriate statistical analyses are chi- square and frequency distribution, and a one-to-one (equality) transformation (e.g. 1=green, 2=yellow, 3=red).
Ordinal level of measurement: the values at this level can be ordered in ranks. All variables measured as high, medium, or low (e.g. yellowing of plant leaves), or as scales of opinion (strongly agree / agree / neutral / disagree / strongly disagree) are ordinal. Ordinal scales provide data about less and more – e.g. strongly agree is more than agree; however what ordinal variables do not tell us is how much more. The central tendency measure of an ordinal scale can be defined as median or mode, while mean cannot be interpreted. Appropriate statistical analyses are percentiles and non-parametric analysis, and monotonically increasing transformation (which retains the ranking); nevertheless, more sophisticated analyses like correlation, regression, and analysis of variance, are not suitable.
Interval level of measurement: the values at this level have all the properties of nominal and ordinal variables; additionally, the distances between the observations are meaningful. Interval level of measurement is quantitative measurement. The measured values are not only ordered in ranks, but the distance between adjacent attributes on a scale is always the same; for example, the temperature scale in Celsius, where the difference between 30 and 40 degrees is the same as that between 80 and 90 degrees. Interval scale enables us to describe how much more, or how much less, one measurement is compared to another, which is not the case with nominal or ordinal scales. The central tendency measures can be mean, median, or mode. Measures of dispersion, such as range and standard deviation, are also possible. Appropriate statistical analyses include all of the methods suitable for nominal and ordinal scales, as well as correlation, regression, and analysis of variance. Scale transformation should be positive linear.
Ratio level of measurement: in addition to having equal intervals, the observations can have a value of zero as well, meaning the absence of the phenomenon being measured. Ratio scales have all the characteristics of nominal, ordinal, and interval scales, as well as a ‘true zero’ point. Most measurements in the natural sciences and engineering, such as mass, volume, concentrations of compounds, and electric charge, are ratio scales. All statistical methods and transformations are suitable.
Figure 1: Levels of measurement
Validity, Reliability, Accuracy, and Precision
Validity is the quality of being legally or officially binding or acceptable. The validity of instruments, data, and findings is the most important requirement in research. It refers to their accuracy and trustworthiness. The validity of the data depends on the validity of the instruments; however, assuming that the instruments and data are valid, the validity of the findings and conclusions can still be questioned (Nayak & Singh 2015).
Reliability is the quality of performing consistently well. Reliability shows if it is possible to get the same result by using an instrument to measure a variable more than once. Instruments can be laboratory devices, scales, or they can be questions given to a group of people.
Precision refers to the number of decimals in a numerical result of a measurement.
Accuracy is the degree to which the result of a measurement, calculation, or specification conforms to the correct value or a standard. Accuracy refers to the level of precision of the scale.
Copyright © Partners of the Aqu@teach Project. Aqu@teach is an Erasmus+ Strategic Partnership in Higher Education (2017-2020) led by the University of Greenwich, in collaboration with the Zurich University of Applied Sciences (Switzerland), the Technical University of Madrid (Spain), the University of Ljubljana and the Biotechnical Centre Naklo (Slovenia).