~ The Scientific Method ~
|
Watch the inimitable Neil DeGrasse Tyson present an engaging (and comical) example of science and the scientific method for a general audience, like only he can!
Science distinguishes itself from non science (and bad science) by way of a strict set of criteria that is then subjected to the scientific method. When we speak of something that is scientific, whether it is a medical intervention, a new discovery, or a particular theory about something previously discovered or known, it must first and foremost be what is referred to as falsifiable. Falsifiability, a concept originally proposed by Dutch physicist Christiaan Huygens (1629-1695) but brought into public awareness and endorsed by the famous British philosopher of |
science Karl Popper. Falsifiability simply states that for something to be considered scientific it must be able to be subjected to the scientific method, and the scientific method can only apply to a theory or hypothesis that can be shown to be false. So what does that mean?
CHECK OUT THE VIDEO BELOW: The Need For Rigorous Science...
If something is falsifiable, it means we can show that it either does not exist, or is incorrect. A typical example would be the belief about the existence of God. The existence of God cannot be proved to be true because we can't disprove it or prove that it is false. People the world over have been quick to demonstrate certain 'signs' that are meant to be interpreted as proof of the existence of God. But even if that were true, it would not reach the threshold of science. How do we research it? Whether God does or does not exist can neither be proven nor disproved, and is therefore relegated to that of a belief which cannot be subjected to the laws of science. It is important to understand that in science, we never set about to prove something is true. To the contrary, scientific research sets about to prove that something either is or is not false - these are quite different things. And we set out to do this by something called the Null Hypothesis symbolized as H0. This is discussed below, but first let's take a look at the steps of the scientific method.
The scientific method is a set of hierarchical steps that must be followed in order to conduct research. Research that is not conducted in this fashion (or at all) is not considered scientifically tested and therefore not scientific. Scientific research is based upon an accumulation of evidence that can be replicated, is falsifiable as stated above, is systematically observable, is free from bias, uses a control group in which to contrast from the experimental group, contains a large enough sample size (number of study participants per group measured), follows a standardized methodology for the hypothesis under investigation, submits the information to the appropriate set of statistics for data analysis, interprets the analysis, draws conclusions, and communicates the results. Here is the standard that is followed for any all types of scientific research:
The Scientific Method
1) Make an observation about something you are curious about
2) Ask a question about the observation
3) Conduct a thorough, exhaustive literature review on any and all of the known variables in your question
4) Construct an hypothesis based upon the background literature review by deciding on the variables to use (i.e., decide what
will be the independent variable, and what constitutes the dependent variable. This is what automatically determines the
experimental design to use
5) Test the hypothesis using standardized statistics (this will include the sample size or 'n' you need to use, and at what 'level
of statistical significance', i.e., at the .05 or .01 level, do you need to achieve
6) Analyze the data set as it is, without drawing any conclusions or making any inferences
7) Draw your conclusions based upon the totality of the analyzed data, making knowledgeable, informed inferences
8) Communicate the results according to the standardized methodological format in a peer-reviewed journal.
Testing The Null Hypothesis (H0)
When conducting scientific research we are interested in finding out whether or not a statistical difference (AKA an 'effect') was found. If we are interested in testing a new drug, ultimately we want to know whether or not the drug in question will be effective. In research we discuss this in terms of whether or not there was a statistical difference in the H0. The H0 means that there was no statistical difference, that our new drug is not effective. If the drug turns out to be statistically significant, that it in fact had an effect, we refer to this as the alternate hypothesis, symbolized as Ha or H1. But remember, we never test to prove effectiveness, we test against the H0.
Scientific research is a complex, multidimensional set of statistics and statistical analysis, sampling techniques, probability theory, methodological understanding, knowing the different types of research design such as prospective, retrospective, and ambidirectional cohort studies, case or surveillance studies, case-control and cross-sectional studies, experimental, quasi-experimental or clinical research studies. In addition, there is the need to understand how to statistically calculate the sample size called the N, what are the different variables and how one has structured and differentiated the dependent and independent variables, and the types and levels of error (which does not mean 'mistake') such as distinguishing between a Type 1 or Type II error, and the directionality of the hypothesis such as a one-tailed or two-tailed study. Then there is understanding randomization, power, and statistical levels of significance, and so very much more. All of this is listed here, not to confuse, but to illustrate the magnitude, complexity and depth that is involved in conducting scientific research. Hopefully this will help you to understand that science and scientific research is incredibly rigorous, and not something to take for granted or pass off as minor, unimportant, or trivial.
CHECK OUT THE VIDEO BELOW: The Need For Rigorous Science...
If something is falsifiable, it means we can show that it either does not exist, or is incorrect. A typical example would be the belief about the existence of God. The existence of God cannot be proved to be true because we can't disprove it or prove that it is false. People the world over have been quick to demonstrate certain 'signs' that are meant to be interpreted as proof of the existence of God. But even if that were true, it would not reach the threshold of science. How do we research it? Whether God does or does not exist can neither be proven nor disproved, and is therefore relegated to that of a belief which cannot be subjected to the laws of science. It is important to understand that in science, we never set about to prove something is true. To the contrary, scientific research sets about to prove that something either is or is not false - these are quite different things. And we set out to do this by something called the Null Hypothesis symbolized as H0. This is discussed below, but first let's take a look at the steps of the scientific method.
The scientific method is a set of hierarchical steps that must be followed in order to conduct research. Research that is not conducted in this fashion (or at all) is not considered scientifically tested and therefore not scientific. Scientific research is based upon an accumulation of evidence that can be replicated, is falsifiable as stated above, is systematically observable, is free from bias, uses a control group in which to contrast from the experimental group, contains a large enough sample size (number of study participants per group measured), follows a standardized methodology for the hypothesis under investigation, submits the information to the appropriate set of statistics for data analysis, interprets the analysis, draws conclusions, and communicates the results. Here is the standard that is followed for any all types of scientific research:
The Scientific Method
1) Make an observation about something you are curious about
2) Ask a question about the observation
3) Conduct a thorough, exhaustive literature review on any and all of the known variables in your question
4) Construct an hypothesis based upon the background literature review by deciding on the variables to use (i.e., decide what
will be the independent variable, and what constitutes the dependent variable. This is what automatically determines the
experimental design to use
5) Test the hypothesis using standardized statistics (this will include the sample size or 'n' you need to use, and at what 'level
of statistical significance', i.e., at the .05 or .01 level, do you need to achieve
6) Analyze the data set as it is, without drawing any conclusions or making any inferences
7) Draw your conclusions based upon the totality of the analyzed data, making knowledgeable, informed inferences
8) Communicate the results according to the standardized methodological format in a peer-reviewed journal.
Testing The Null Hypothesis (H0)
When conducting scientific research we are interested in finding out whether or not a statistical difference (AKA an 'effect') was found. If we are interested in testing a new drug, ultimately we want to know whether or not the drug in question will be effective. In research we discuss this in terms of whether or not there was a statistical difference in the H0. The H0 means that there was no statistical difference, that our new drug is not effective. If the drug turns out to be statistically significant, that it in fact had an effect, we refer to this as the alternate hypothesis, symbolized as Ha or H1. But remember, we never test to prove effectiveness, we test against the H0.
Scientific research is a complex, multidimensional set of statistics and statistical analysis, sampling techniques, probability theory, methodological understanding, knowing the different types of research design such as prospective, retrospective, and ambidirectional cohort studies, case or surveillance studies, case-control and cross-sectional studies, experimental, quasi-experimental or clinical research studies. In addition, there is the need to understand how to statistically calculate the sample size called the N, what are the different variables and how one has structured and differentiated the dependent and independent variables, and the types and levels of error (which does not mean 'mistake') such as distinguishing between a Type 1 or Type II error, and the directionality of the hypothesis such as a one-tailed or two-tailed study. Then there is understanding randomization, power, and statistical levels of significance, and so very much more. All of this is listed here, not to confuse, but to illustrate the magnitude, complexity and depth that is involved in conducting scientific research. Hopefully this will help you to understand that science and scientific research is incredibly rigorous, and not something to take for granted or pass off as minor, unimportant, or trivial.
|