2.2 Process for using the Scientific Method

Observation and Research

Scientific research starts with scientists making an initial observation that they cannot explain with an existing theory. It can also be a more open-ended question, such as ‘How can I treat this particular disease?’ The scientist will study any previous data – that is, review the existing literature on the topic area- to determine which facts are relevant to the problem at hand. This research may reveal that another scientist has already answered the question, and, in this case, a new question will be asked. If the question hasn’t been answered, the scientist will use the findings of their research to help them design a good question that they can test.

Hypothesis and Aim

Research and imaginative and creative thinking help scientists formulate hypotheses that they can test to explain an observation or answer a question. A hypothesis is an assumption based on knowledge obtained during the observation stage – it is an ‘educated guess.’

Data generated during an experiment either supports or fails to support the hypotheses. Hypotheses may be specific (e.g. chronic ingestion of artificial sweeteners causes weight gain in humans) or broad (e.g. viruses cause disease in animals). Hypotheses are always subject to modification.

However, it is important to note that hypotheses may be proven untrue. This occurs when the data do not support the hypothesis, in which case, the hypothesis must be rejected or refined.

A hypothesis is tightly linked to the aim of the study, which is the objective or goal of an experiment or series of experiments. The aim is written in a similar manner to the hypothesis. A hypothesis is a prediction about the outcome of an experiment; an aim is usually written in the form of an instruction:

Experiment

Physiologists conduct experiments to test whether the real world behaves as predicted by the hypothesis. The purpose of the experiments is to find out if the resulting observations of the real world agree or conflict with the prediction – if they agree, we can be more confident that the hypothesis is true, but if they disagree, we become less confident.

Experiments are a valuable scientific tool because when experimenting, we are able to control and manipulate phenomena and events. However, this also means that we are creating an ‘artificial environment’ for our experiments. This may not end up reflecting what is happening in the real – messy – world.

Experiments will only help us to support or not support hypotheses if they are carried out properly. This means using care and the right equipment and settings with the goal of minimizing error. If we conduct careless experiments that introduce errors in the results, the observations (or data) are useless when we try to evaluate our hypothesis.

Part of understanding how science works includes becoming aware of the methods and tools of science. Each scientific discipline has its own specialized techniques and tools that it uses to observe the natural world; student scientists will be exposed to some of these methods in practical classes during their studies. Experimental methods are not perfect, even when scientists try their best, and they can be updated or replaced over time. Advances in technology or scientific understanding may mean that previously accepted methods are rejected and replaced with new ones. This may also mean that experimental results obtained using old methods are ignored or viewed as less relevant because our way of thinking about the topic has changed.

It is not always easy to conduct well-designed experiments and to get reliable results. New experiments may take months or even years to perfect. Scientists need to work hard to collect high-quality results, including:

• developing their theoretical and practical knowledge of an area and a technique
• learning through trial and error
• mastering the available technology.

Researchers must provide details on how their experiments were conducted because other researchers must be able to repeat the results. If results cannot be repeated, the hypothesis is no longer supported, and the conclusion may be incorrect. When the results of experiments are shared with the wider scientific community, the research will be judged on many things, including the methods employed.

Also, it is very rare for a researcher to generate their experimental methods from scratch; it is more common to use methods that have been shown by other researchers to be successful at measuring particular variables. This is another reason it is important to provide details of your methodology.

Results

Scientists collect the data generated from their experiments and analyze them to determine whether the hypothesis was supported or not. In the analysis stage, the scientist combines the results from repeated experiments and carries out statistical analysis to test the hypothesis – for example, to determine whether differences exist between different experimental groups. Scientists will present their results as text, tables and graphs.

Results should be reproducible – that is, another scientist should be able to follow the same method and obtain the same results. If they are not able to do this, it may mean that the original results were an error. For this reason, scientists will usually conduct a single experiment multiple times.

Scientists rarely attempt to replicate the findings of other scientists despite the importance of reproducibility. There are few incentives for scientists to dedicate time to replicating existing studies, and it can be difficult to replicate studies because the methods may be unclear, and poor or incorrect study designs may have been used in the original research (Belluz et al., 2016).

Recently, some scientists have found that when they do try to replicate studies, they have not been able to obtain the same results. This has led to a ‘crisis of reproducibility.’ Scientists who have investigated reproducibility have found that few findings published in top psychology journals as well as highly cited papers on psychiatric treatment (Tajika et al., 2015) could be replicated (Open Science Collaboration, 2015). This issue is gaining more attention and must be addressed by scientists from all disciplines.

Conclusion

Conclusions are made about the results of the experiment in light of the hypothesis. Simply, the results do or do not support the hypothesis. The conclusions below are the two possibilities arising from the example hypothesis and aim shown previously:

The study hypothesis, aim and conclusion are therefore highly connected.

Scientific conclusions are subject to change when new findings dictate that they need to be changed. Quite often in science, new studies, which might use new techniques and equipment, tell us when conclusions need to be modified or changed entirely.

Share results

Results of experiments should be shared with the wider scientific community, so they can contribute to the pool of knowledge on a topic and lead to better understanding for everyone.

Scientists can communicate with their peers in different ways:

• make an oral presentation at a conference
• present a poster at a conference
• publish a journal article in a peer-reviewed academic journal

The publication of results makes information widely available, and the results will be used by others if they are seen as interesting. Results that are not deemed interesting will fade from sight; sadly, this is the fate of most published articles.

The Scientific Method in Action

Observation: A physiologist determines that available evidence indicates that, say, drug X – which Australia’s Therapeutic Goods Administration does not yet approve – increases the metabolic rate and may, therefore, be a promising weight loss drug.

Hypothesis: The physiologist hypothesizes that ‘drug X will cause weight loss in adult rats.’

Aim: The physiologist plans a study with the aim to ‘determine if drug X will cause weight loss in adult rats’.

Experiment: The physiologist randomly assigns 20 adult rats to a control and treatment group – 10 per group. The control group is administered a daily dose of the placebo, and the treatment group is administered a daily dose of the drug for four months. The control group receives a placebo instead of a drug, and all other conditions are identical to the treatment group, with the exception of the drug. They receive the same kind of food and water. Body weight is recorded daily.

Results: At the end of the 4-month treatment period, the physiologist performs a statistical analysis on the body weight data. The data reveal a statistical difference between treatment groups, with the treatment rats weighing less than the control rats.

Conclusion: Drug X caused weight loss in adult rats.

Share results: The physiologist will share their work with peers through a poster or oral presentation at a conference and/or a journal article.

 

“Science and the Scientific Method” from How To Do Science by University of Southern Queensland is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.—Modifications: Used sections Observation and research, hypothesis and aim, Experiment, Results, & Conclusion, edited; Used section Share Results, edited, removed Box 1.5 and associated content.