Science & Scientific Method Reference Guide
	schmied/reid©2009 All rights reserved

Science is many things to many people!

The process of science is used to discover how to cure everything from toenail fungus to skin cancer! Think of the problems we are using scientific methods to solve today. For example:

• Finding a cure for Human Immunodeficiency Virus (AIDS)

• Discovering a cause and cure for Mad Cow disease

• Restoring the wild Chinook (King) Salmon runs in the Pacific Northwest

• Finding new ways to manage waste without ruining our environment

• Creating alternate energy resources to use before our petroleum resources run out.

Science is extremely versatile! Science can be used to solve many different problems in our society. But, what is science? That might be a helpful to start. Let's look at the big picture of Science first.

 

So just what is Science all about? ref: Dr. Bruce Railsback, Department of Geology, University of Georgia. http://www.gly.uga.edu/railsback/1122science2.html

    Science is the concerted human effort to understand, or to understand better, the history of the natural world and how the natural world works,

….. with observable physical evidence as the basis of that understanding.

Science is done through observation of natural phenomena, and/or through experimentation that tries to simulate natural processes under controlled conditions.

So what types of Science are there?

There are two types of science, Observational and Experimental. Some branches of science mainly use one type, some use some of both. Here are examples of each.

Observational - scientists doing observational science are mainly focusing on collecting data with human senses and tools to answer scientific questions. For example:

• An ecologist observing the territorial behaviors of bluebirds

• A geologist examining the distribution of fossils in an outcrop

• An astrophysicist, photographing distant galaxies to determine the type galaxies and objects within galaxies.

• Lastly a climatologist sifting data from weather balloons on weather patterns

In each of these examples the scientists are making observations in order to find patterns in natural phenomena.

Experimental Science - On the other hand, scientists doing experiment are using senses and tools during an experiment to answer scientific questions. Examples:

• A chemist observes the rates of one chemical reaction at a variety of temperatures to see what patterns emerge.

• A nuclear physicist records the results of bombarding a particular kind of matter with neutrons in order to see what patterns are present.

• Finally, a Biologist observing the reaction of a particular tissue to various stimulants is likewise experimenting to find patterns.

These scientists are experimenting in order to detect patterns in nature.

The really important things all scientists have in common across the different fields of science is that:

1. Making and recording observations of nature, or of simulations of nature, in order to learn more about how nature, in the broadest sense, works.

2. One of the main goals is to show that old ideas (the ideas of scientists a century ago or perhaps just a year ago) are wrong and that, instead, new ideas may better explain nature

Four Things Science Isn't - Another way to look at Science is to see what it isn't.

1. Science Isn't Art - Art is the attempt to express an individual's feelings or ideas about something in a way that others find beautiful, graceful, or at least aesthetically satisfying. Thus art is very individualistic.

On the other hand, science is the attempt to reach demonstrable, replicable, conclusions about the natural world (and social science is the corresponding attempt to reach demonstrable conclusions about the social or human world).

2. Science is not Technology - Science doesn't make things. Scientists generate knowledge. Engineers use scientific knowledge to generate technology.

3. Science isn't Truth and Science isn't certainty - Most scientists seek Truth; they don't know or generate Truth. Scientists use evidence to propose and test theories, knowing that future evidence may cause refinement, revision, or even rejection of today's theories

4. Science isn't a religion or a belief system - Science and belief systems are very different, in what they try to do and in the approaches they use to accomplish their goals. Science seeks to explain the origin, nature, and processes of the physically detectable universe. Science is an activity where the underlying assumptions are tested and retested.

Science Asks Three Basic Questions http://evolution.berkeley.edu/evosite/nature/I3basicquestions.shtml

a. What's there? The astronaut picking up rocks on the moon, the nuclear physicist bombarding atoms, the marine biologist describing a newly discovered species, the paleontologist digging in promising strata, are all seeking to find out, "What's there?"

b. How does it work? A geologist comparing the effects of time on moon rocks to the effects of time on earth rocks, the nuclear physicist observing the behavior of particles, the marine biologist observing whales swimming, and the paleontologist studying the locomotion of an extinct dinosaur, "How does it work?"

c. How did it come to be this way? Each of these scientists tries to reconstruct the histories of their objects of study. Whether these objects are rocks, elementary particles, marine organisms, or fossils, scientists are asking, "How did it come to be this way?"

Seven Characteristics of Science (SHOERAP)

(Many thanks to Professor Roger Olstad, University of Washington for his inspiring thoughts for this section!)

Science is many things, but people who use science agree there are seven characteristics of scientific things. Science is:

• a Structured way of thinking. Science uses a process to solve problems. This process is the Scientific Method! In it scientists figure out how they know what they know and how to show it.

• a Human activity. Science is done by people.

• Observable. Things done in science are things that are able to be observed. After all, the process of science demands evidence. Often scientists need tools to help make observations to get evidence, like the Hubble Space Telescope or a Scanning Tunneling Electron Microscope.

• Experimental. People do science by choosing a question about something in our world. Then people make observations or experiments to try to answer the question.

• Repeatable. If an experiment is done properly anyone ought to be able to repeat the experiment and obtain the same data/results. Every time!

• Assumes the Universe can be ordered. Scientists feel there is an underlying order to how things happen in the universe. They also think there are unifying rules that produce this order, and that these rules are able to be known. Scientists develop better and better ways to order things (like classifying plants, animals, stars, and minerals) in an effort to find a concept or system that will help order the entire universe. Awesome!

• Public and Verifiable. Science is a publicly verified process. All science work is reviewed, and must be verified (discussed and agreed upon as being true based up on the evidence present at the time) by other scientists to be published and used by the world

 

The Scientific Method

The scientific method can simply be described as "Organized common sense". But there is a lot more to it than that. In fact, There is no such thing as "THE Scientific Method." If you go to science fairs or read scientific journals, you may get the impression that science is nothing more than "question-hypothesis-procedure-data-conclusions." But this is seldom the way scientists actually do their work. Most scientific thinking, whether done while jogging, in the shower, in a lab, or while excavating a fossil, involves continuous observations, questions, multiple hypotheses, and more observations. It seldom "concludes" and never "proves."

Putting all of science in the "Scientific Method" box, with its implication of a white-coated scientist and bubbling flasks, misrepresents much of what scientists spend their time doing. In particular, those who are involved in historical sciences work in a very different way&emdash;one in which questioning, investigating, and hypothesizing can occur in any order. http://evolution.berkeley.edu/evosite/nature/IIprocess3.shtml

I'd like you to keep the above ideas firmly in mind while you work through the rest of this handbook. The handbook will explore the Scientific Method, knowing that this is just a starting place for learning about science. As your experience deepens you will learn science isn't always a step by step process, but a whirlpool of ideas, experience, intuition, evidence, and serendipity.

First, lets take a look at the different types of investigations scientists actuallydo.

What most people learn in Science class is usually a mix of all five, with emphasis on the Controlled Experiment. However the controlled experiment is not all there is to science! The other types of investigation are every bit as valid, but focused differently. We are now going to look really hard at the overall process of science and keep in mind that there are different types of investigation.

Take a look at this big picture view of the "textbook" scientific method. Examine the process in detail.

 

The Scientific Method is a process. A process is a framework for doing science. Embedded in this process is a whole lot of creative thinking, hard work, determination, perseverance, dedication, and common sense. As scientists find, with experience and time, there's some incredible flexibility in the process. But if you are reading this, you are probably just starting off, so it ís best to learn the basics. Lets begin by learning the process and by some standard ways of "showing what you know" with figures, data tables, graphs and words to describe what you accomplished in your experiments. So take a deep breath and lets take an in depth look at this famous process!

 

Step One: Stating a Problem Starting at the beginning assumes a researcher has no prior knowledge of the topic. This is uncommon for scientists, who have years of experience, but for students, it's a good place to begin..

A sense of awe, curiosity, and wonder….

One time, long ago, every scientist was a curious young girl or boy, enthralled by a dragonfly, a frog, the stars, or a cool looking rock. Somewhere along the way each grew up and became an adult. However that captivation about the natural world never left. Today, as for centuries past, in some way, shape, or form all scientists are dedicated to the study of the natural world. As we discussed earlier, each scientist is different; some are specialized in atomics, ecology, geology or other fields. But they all possess something very special, that's a sense of awe and wonder in the natural world. This sense frames their vision and clarifies their thoughts. This perspective allows scientists to see problems in new ways and see new problems surfacing. It isn't so surprising, after all, that most scientists have incredibly strong attachments for the natural world and in sustaining our environment.

Observe and gather Information about a problem.

Normally, when faced with a brand new situation one has to start off by observing and gathering data about the situation. Once you gather enough data you can frame or describe a problem. That's what scientists do, but it isn't as simple as it looks. Observing is a skill that takes time to learn and collecting data is too. Observers recognize or note facts and occurances in the natural world using tools and senses. (touch, seeing, hearing, smelling, and tasting.) However, quite often what we see and what's really there are two different things, so taking numerical data helps sorth things out. Generally, scientific data falls into many categories and types. Overall, data is factual information from investigations that end up being recording as numbers, or using descriptive words. It is worth noting that evidence includes observations, including the act of measuring, and data, which is experimental information gathered as numbers.

 

Scientists are trained observers. They are tuned in to receiving data and creating questions about the natural world. There are many other types of trained observers in the world, for example a Coast Guard Deck Officer conning a ship is a trained observer, as is a Police Patrolman, or Nurse. Trained observers see things a bit differently than the rest of us. A trained observer sees beyond the surface facts to possible causes.

For example: A scientist, when hiking along a Alaskan streams that no longer have salmon runs, might notice that health of the forest edge near the water is not as healthy as the forest edge of Alaskan streams that still have healthy salmon runs. The scientist is observing trees in one place, but also thinking about what the tress look like in another. A normal hiker might just be thinking, "What a fine fall day." The scientist, on the other hand, is gathering evidence by observing, taking and comparing data based upon experience.

How to create an inference. Basically, creating an inference is linking evidence (including observations) together in your brain and creating a preliminary conclusion. Inferences are supported by evidence, but not necessarily fact.

Let's use the salmon example from the paragraph above. Here are two observations that we can use to make an inference in this situation.

a. The forest edge near streams that no longer have salmon runs is unhealthy.

b. The forest edge near streams with healthy salmon runs is healthy.

Now let's create an inference from the data. Use reasoning we can draw a preliminary conclusion, or inference, from this data.

Inference: Salmon provide nutrients to the streambanks when they die. So the stream banks in areas with no salmon runs have less nutrients, there are less healthy plants on these streambanks.

The process of creating inferences is called Inductive Reasoning, or inferring.

"On any new problem inductive inference is not as simple and certain as deduction, because it involves reaching out into the unknown."

Science, Strong Inference.- Proper Scientific Method (The New Baconians) 16 October 1964, Science Magazine, Volume 146, Number 3642 by John R. Platt

Developing a Scientific (Testable) Question

Thanks to the: Guide to Scientific Questions http://www.science-house.org/nesdis/gulf/guide.html

 

A Question in science isn't like a problem on Rosie, or Oprah, it's entirely different. A scientific question is about something natural in our universe. It is stated in a way that frames, or describes, a problem that can be tested.

Scientific questions evolve through three main types of questions. Each builds up more knowledge of the matter to be tested until the researcher creates an Experimental question .

1. Verification questions are basic data collecting questions. (Is it cold today? What is the temperature/color? Is this how this works?). They build up knowledge.

2. Significant questions require an explanation and prior knowledge. (Is it important that this is done first? Do clouds have to be in the sky before it will rain? Is it significant that…. Etc) They increase knowledge of the subject.

3. Experimental questions require explanations and are testable. (If salt is added to the copper sulfate solution, would the solution still boil? or If SPF 45 suntan lotion is put on ultraviolet detecting beads, will they still change color?) They are what researchers use!!

Creating scientific questions is a skill that can be learned, just like sewing. We master this skill when we are able to ask and identify all three types of scientific questions. Lets take a look at some good and not as good scientific questions.

 

There are a couple guidelines to creating strong scientific questions.

1. A good scientific question is one that can have an answer.

"Why is that a rock?" is not as good as "What are rocks made of?"

2. A good scientific question can be tested by some experiment or measurement that you can do.

In this case "Where does the Sun come from?" is not as good as, "What adaptations do some birds have that allow them to fly??"

3. A good scientific question builds on what you already know.

"Will fertilizer make grass grow greener?" is not as good as, "What is the source of the genetic mutations that cause birth defects?"

4. A good scientific question, when answered, leads to other good questions.

"What is HIV?" does not lead to as many other questions as, "How does the HIV virus cause the immune system to malfunction?"

The good questions above ask What and How in a way that focuses you in on the specific problem to be studied. These questions frame a problem in a way that can be tested.

For example, a question for the salmon situation above might be:

"What is causing the forest bordering the streams that no longer support salmon runs to be unhealthy?"

 

Do research about the question. Even the best scientists need to do research to learn more about their question.

First, research is done to see if the question has been answered already.

Second, research will show if the question, or a similar question, has been attempted by another researcher. If there is data on how to answer a similar question, the researcher might be able to adapt the procedure and technique used. If research shows other scientists have been unable to answer the question, then it will be important to read their work closely. This is key because the original researcher will report the methods, materials, and experimental design they used. New researchers can try the original set up again to see if there were flaws, or errors, in the procedure. Or they can change the procedure to see if a new procedure will give them better results.

Finally, if no one has attempted the problem, a scientist knows there is a fresh open field ahead for them to work in.

Creating strong Scientific Questions:

We use a specific format to create scientific questions that a student can use anytime and still come up with a strong question that science can answer. Here is how this is done.

Things needed to start the process. To develop a strong scientific question a student needs to know three things and form the question in a way that it can be answered scientifically:

Three things to know:

• Study Subject (SS)

• Manipulated Variable (MV)

• Responding Variable (RV)

Three other guidelines:

• Is phrased to ask for an answer that provides details vs a simple yes or no answer.

• Is a definite statement of what is proposed. ex: increasing the amount of fertilizer

• Has no Pronouns. Pronouns like "it" are vague, easy to misunderstand. Use proper nouns instead.

Now let's use an example experiment to demonstrate one way to form a scientific question.

Here's the situation:

Greenlander Tulips and Light

In this experiment students exposed 12 Tulips (Greenlander type) to sunlight from dawn to dusk .

Then another group of 12 Tulips were grown at the same time, in a similar sized container, but were given 14 hours of sunlight each day, which is a couple hours more than the other trials, on average.

Let's say this experiment went on for a two months and that ever week students measured and recorded the height of each Tulip.

So let's be sure we can answer these three questions:

What is the Study Subject?	What is the Manipulated Variable?	What is the Responding Variable?
Greenlander Tulips	garden fertilizer (10:10:10)	the height of the Tulips.

Now that we have outlined the experiment, we can put together a question in this format:

Example question format:

How will ...?... Manipulated Variable .... Study Subject.. (either order) affect the Responding Variable ?

(The Responding Variable should be last.)

Example Question for the Tulip problem would look something like this:

How will increasing the amount of sunlight on Greenlander tulips affect the height of the tulips?

As you can see, this question has an answer that can be obtained using scientific methods. Also:

a. The question asks for an answer that provide details vs a yes or no answer.

b. The question is a definite statement of what is proposed.

c. There are no pronouns in the question.

 

 

Step Two: Forming A Hypothesis

Creating a Prediction.. (From I think to an If, Then prediction)

A prediction is a conditional statement forecasting future events. When you make a Scientific prediction, you need data to support it. With data it's easy to set up a prediction using an IF, Then phrase. (If this occurs,... then I predict this will/might happen...) Predictions phrased like this help develop hypothesis and the experiments to test them.

To help us let's use this question:

Can Ultraviolet detecting beads, placed in the sun, be kept from changing color?

The problem is framed as an Experimental question. Now let's use other information we already know to help predict ways to solve it. We know ….

A. The sun emits ultraviolet

B. Suntan lotion prevents humans from getting sunburn by blocking Ultraviolet radiation

Applying this knowledge to the question, its possible to come up with this prediction

If SPF45 sunscreen is put on Ultraviolet detecting beads and the beads are placed in the sun,

Then UV detecting beads with SPF45 sunscreen will have less color change compared to normal UV beads.

One can see how the use of an If, Then statement is very powerful. It focuses one's thinking on a potential area of study; the ability of sunscreen to block ultraviolet radiation. A prediction like this makes it easy to set up an experiment later!

 

Forming a Hypothesis

Thanks to: http://www.gpc.edu/~bbrown/psyc1501/methods/hypothesis.htm

The "centerpiece" of research is the hypothesis.  The hypothesis reflects the researcher's beliefs about the answers to the question he or she is studying, and the research methods must be designed to adequately and appropriately test the hypothesis.

A Hypothesis makes a specific prediction about the outcome of the research. You can also think of the hypothesis as an "educated guess" or prediction about what the research will find.

For research purposes, hypotheses are usually made as testable statements about the relationships between or among variables.  (Henslin, 2000/ Shaughnessy, 1994, Cades, 1999)

A Hypothesis is a structured educated guess which identifies a possible test of the prediction's strength. Some people call a hypothesis, "A prediction stated in a way that can be tested." or a "Prediction with a reason."

Hypotheses are put together in many ways. One good way is to use the If, Then, Because format. Thus, we start with our prediction and tack on a potential reason why it might work. Then we can figure out what to test in our experiment.

Lets take the UV bead prediction and revise it into a Hypothesis.

If SPF45 sunscreen is put on Ultraviolet detecting beads and the beads are placed in the sun,

(Note the SS & MV are in the IF)

Then UV detecting beads with SPF45 sunscreen will have no color change compared to normal UV beads.

(Note the experimental trial is stated first, then there is a definite prediction of what will happen and the prediction is compared to the Control!)

Because SPF 45 sunscreen prevents UV rays from penetrating to the surface of the UV beads. Therefore the UV beads with sunscreen will not change color.

(Note the SS, MV, and the RV are in the Because!)

 

Truth and Hypothesis: Now we don't know if this hypothesis is true or not. For one thing, it may be quite a stretch to think that sunscreen designed for humans will work on UV Detecting Beads, but this is what experimenting is all about, to test out the research hypothesis.

For another thing, hypotheses are never proven to be true! Hypothesis can only be tested and found to be "Accepted" or "Rejected". (Note: A hypothesis can be weak. This means the data is inconclusive, doesn't support OR disprove your conclusion at the moment and needs further testing. This situation doesn't usually last very long.)

 

Multiple (Alternate) Hypotheses. a twist.. scientists often do two or more things at once!!

Scientists often create more than one hypothesis about the same problem and test them out at the same time! This technique goes by a couple names.... The Alternate Hypothesis or Multiple Hypothesis. Why create multiple hypotheses?

First of all, the use of multiple hypotheses reflects reality. There are often many possible reasons something occurs the way it does. So why not test as many of the leading possibilities as practical?

Second, it forces the researcher to really think about the entire situation and not focus on only one answer before the experimental results are analyzed. (this could lead to bias)

Finally, it saves time and money to investigate multiple hypotheses at once. For example two hypotheses for the question

"What is causing the forest bordering streams that no longer support salmon runs to be unhealthy?"

might be

1. "If salmon are no longer present in a stream, then the nearby forest will lose nutrients and suffer a loss in health, because when salmon decompose their nutrients are returned to the forest edge." could also be

2. "If hunters kill all the bears around a stream, then the nearby forest will lose the nutrients the bears leave and suffer a loss in health, because the bears would normally excrete waste that are nutrients for the forest."

...Interestingly enough, both are sometimes supported by the experiment!

(Ok! Hopefully by now you have a good idea on how to observe, gather information, create a question, make a prediction, and form a Hypothesis. (If not, please go back and reread the previous sections, then if you still don't know how to do these things, ask an educated classmate or your instructor for help.)

 

Step Three Performing the Experiment, or "Testing the Hypothesis"

The entire purpose of doing an experiment is to "test the hypothesis". In other words, to see if the experiment's results accept or reject the Hypothesis.

Technically, a properly designed hypothesis sets up the design of the experimental trials. In the UV bead and SPF sunscreen example above, properly framing the problem as a scientific question led right into a prediction and then into a hypothesis.

We've found that if a scientific question and hypothesis are written properly, students can actually visualize the experiment before they even start an experiment!

Then tests can be easily designed to measure a hypothesis's strength by seeing if the things in the BECAUSE phrase actually happen! For example, one can actually visualize how to design of a series of test trials to see if SPF 45 sunscreen, applied to Ultraviolet beads, actually blocks the sun's ultraviolet radiation.

Since the experimental design has many elements, it's important to be systematic. I'll list key the experimental elements below, and go over each not covered above in the text following the list.

Experiments:

• seek to answer a question • are well researched • make a prediction

• have a clearly stated hypothesis and test that hypothesis

• have a clear, complete procedure, including all safety procedures

• minimizes all possible errors that could occur (6 general types)

• identifies all materials used and shows/tells how the materials are set up,

• changes only one variable at a time (manipulated variable)

• always measures one or more key variables (the responding variable)

• controls as many other variables possible. ( both controlled variables and uncontrolled variables)

• have two types of trials, Control Trials and Experimental Trials.

• measure the changing in the responding variable (either quantitatively or qualitatively)

• documents the results in data tables and displays results on graphs and figures

• has an understandable analysis of the trials

• conclusion tells if the original hypothesis is accepted or rejected, and uses evidence and facts from the experiment to support the conclusion. For example Highs, Lows, Averages, Ranges, Differences etc

• repeats and recheck results by using 3 or more trials (Reliability).

• shows the Validity of the experiment by doing other trials that demonstrate another way that a change in the ManipulatedVariable actually caused a change in the Responding Variable

• communicates the results via a peer review, publishing the results, defends the results and allows the results to be applied.

Writing an experimental procedure

A scientific procedure creates a step by step picture of the experiment's methods, materials, safety precautions, and materials.

When writing a procedure, the researcher has to start from the beginning,and carefully put the procedure together step by step, .

Generally procedures include

things to know before starting	materials*	safety precautions
a logical sequence of steps to follow	figures	definitions

*Making a List of Materials there always is a section in the procedure that contains a list of materials used in an experiment. Be sure to include safety equipment! Its a good way to start an experiment safely.

Procedure writing isn't as easy as it looks and requires practice. New procedures should be tested in real life situations, then revised to fix flaws before use. The rule of threes applies here...after the third time through you are ready to go!

Peanut Butter and Jelly and Procedure making: Many students have done the exercise in which each is asked to craft a procedure to make a peanut butter and jelly sandwich. Then another person reads the procedure while the original writer makes an actual sandwich. The results are often hilarious! In reality this is a great way to test out a procedure.... It's the write, test, and revise method.

Once a procedure is written, the old saying, "When all else fails, read the instructions!" applies.

Error: It's important to write a experimental procedure and do the experiment in such a way that all sources of error are minimized. However error is ever present in experiments. In fact, errors can be described as a

Over the years many millions of dollars of experimental data has been rejected due to different types of errors creeping into experimental results and making the final data sets unuseable.

In any event researchers are always on the lookout for different types of error and analyze sources of error in the final conclusions of an experiment.

So what is error?

Error can be defined to be a mistake in perception, measurement or process. What this definition describes are human caused errors. The errors end up in the final results of an experiment.... and the final results are the numbers a researcher analyzes and uses as a basis or their conclusions. That's where statistics comes into the picture.

Errors and numbers: Since experimental data is mostly quantitative (data using numbers) researchers use statistics to analyze and evaluate the hypothesis. Statistically speaking, an error is the difference between the true value and the measured value.

It can be said that there are six general sources of error that can occur in an experiment.

a. Experimental Design error:

This is error is written into the procedure. For example: The procedure overlooks a key element, fails to control variables, skips or reverses steps, doesn't list necessary materials, or isn't clear..etc

b. Operator Error.

These are errors made by the people performing the experiment, which are numerous!

c. Observation Error:

Observers are human and people make mistakes by reading scales incorrectly, seeing things that don't actually happen the way one thinks is so etc !

d. Recording Error:

Human recorders and recording devices make mistakes when recording data. Some of these mistakes are hard to find, especially when a researcher is using tools to record data. At first, the last thing a new researcher will look for when hunting down the source of an error is their equipment. Later, its one of the first!

e. Calculation Error:

Most everyone who has used a calculator, or added a long string of data is familiar with this source of error! (ex 2.4 + 2.4 = 5.3!) Gosh, I've seen people saying some mighty mean things to their computer with the Excel spreadsheet isn't coming up with reasonable calculations.

f. Measuring tool limitation.

This occurs when the scale, size or design of a tool is lacking. This means the researcher is not using the correct tool for the job.

As a result, a tool may not measure to the degree of accuracy needed, the measurements may not be as precise as needed, or perhaps the tool has a faulty design and doesn't stay in calibration

 

Variables Overview:

A Proper Experimental Design:

a. Changes only one variable at a time to test the hypothesis. This is called the manipulated variable in an experiment.

b. Always measures at least one key variable's response. The measured variable is called the responding variable, or independent variable.

c. Controls all other variables possible so all trials experience similar conditions.

 

Variable conditions that could change during an experiment, ie. rainfall, sunlight, temperature, humidity, pH, food, position of samples, etc, etc, etc...

a. There are two types of variables, Controlled and Uncontrolled variables.

b. There are two kinds of Controlled Variables,

i. The Manipulated variable is the variable changed for the purpose of testing the hypothesis.

ii. "All other "controlled variables that are controlled in an experiment to ensure all trials are experience the same conditions and to minimize potential error.

c. There are also two kinds of Uncontrolled Variables.

i. The Responding variable, which is the variable you measure as a result of the experiment. You measure this variable in ALL OF YOUR TRIALS. Actually there could be more than one Responding variable. However, in basic science classes, there is usually only one Responding variable.

ii. "All other" uncontrolled variables that occur during an experiment. Researchers try very hard not to have any uncontrolled variables as uncontrolled variables can wreck an experiment's results!

There are two types of "All Other" uncontrolled variables. These are:

1. Uncontrolled Variables that influence an experiment's results and cause error.

2. Uncontrolled Variables that don't influence an experiment's results.

The graphic below shows another way to look at the World of Variables and may be easier to understand.

Controlling "variables". Scientists try very hard to ensure that all the experiment's trials experience the same conditions. Otherwise their results may be worthless.

There are two General ways control variables!

a. One way is to create a carefully controlled climate for all the trials.

b. A Second way is to expose the experimental and control trials to the same changing conditions.

For example if you perform an experiment outdoors, all trials will experience changes in the weather. This is ok as all trials are getting the same change in temperature, humidity, etc. This can be tricky...

NOTE: If the experimental and control trials are exposed to different conditions, the experiment's results will be of no use because of faulty methods, creating an invalid comparison.

Some ways to control variables would be to water and feed subjects on regular schedules and use identical amounts of water and food.

Also a researcher could expose all test subjects to the same temperatures, same sunlight, bedding, food, rainfall, pH, etc....

For example, in an experiment with mice

a. One might check and fill the mice's water every three days.

b. There would be a certain amount of food present every day, and

c. There would be a timer to turn off and on the lighting on a regular schedule etc, etc, etc

The Manipulated Variable, the Variable changed to test the Hypothesis.

The variable you vary on purpose in an experiment is called the manipulated variable (MV)(also independent variable). It is the one variable changed to test the hypothesis.

For example, one do a test to see if different amounts of sunlight than normal would produce bigger beefsteak tomatoes. (Yum!)

a. In this experiment the amount of sunlight would be the Manipulated variable (MV).

b. The size of tomatoes, or what is measured, is the Responding variable (RV).

In this example the Experimental trial beefsteak tomato plants would receive different amounts of sunlight (MV) than the normal amount of sunlight each day.

The Control beefsteak tomato plants would receive the normal amount of sunlight (RV) for the area.

During this experiment the tomatoes from the Control plants would be compared to the tomatoes from the Experimental trial plants. The researcher would look at the final data set to see if there was a difference in size between the Experimental Trial tomatoes and the Control tomatoes

Of course in this experiment the researchers would make sure all trials used the same type of soil, the same amount of water, fertilizer, humidity, etc,

 

The Responding (Measured or Dependent) Variable is the variable(s) that is measured in an experiment. Researchers measure one or more variables in an experiment. These variable(s) are called the Responding variable(s).

For example, Joe performs an experiment with Ultraviolet radiation detecting beads. His question is: Will SPF45 sunscreen block the Ultraviolet radiation on Ultraviolet radiation detecting beads?

a. In this experiment applying SPF45 sunscreen on the UV beads is the Manipulated Variable.

b. The change in color over time of the UV beads is the Responding Variable.

The control trials will have no sunscreen, the Experimental Trials will have samples with SPF45 sunscreen. Joe would put both in the sun and see if the experimental trials changed color when compared to the control trials. If they didn't the results would show that sunscreen blocks UV radiation on UV detecting beads.

 

Qualitative or Quantitative measurements. There are two ways a responding variable can be measured.

a. With numbers (ûC, meter/sec etc). This is quantitative measurement

b. By using descriptions (soft, bitter, bigger, sweet). These are qualitative measurements

If a responding variable can be measured with some type of number, that's better for the researcher. However, it's hard to measure some responding variable with numbers. This poses a special problem when analyzing data. Usually a researcher creates some type of standard numerical rating scale to give value or relative weight to the observations.

 

Experimental Trials & Controls. An experiment always has two sets of trials.

a. Experimental trials have one variable changed, the manipulated variable.

b. Controls in which no variables are changed. Controls simulate the "normal situation". This is often, but not always, the condition found in nature. Control trials are done at the same time and place as Experimental trials.

Why have Control Trials? Scientists need something to use as a standard to compare the experimental trial results to, so Controls are used. Then scientists can tell that if the results from the control and experimental trials are similar, there was no effect caused by changing the manipulated variable.

For Example, using the Ultraviolet detecting beads experiment from 2H above…

• Experimental trials = UV detecting beads with SPF45 sunscreen.

• Control trials = UV detecting beads with no sunscreen (the natural situation)

(You might even add other experimental trials that have different SPF sunscreens applied to them in same experiment. Later, all the experimental trials will be compared to the Control trials!)

Say in the experiment above the color of the experimental trial UV beads (with sunscreen on them.) doesn't change after placing them in the Sun. However, the control trial's beads all change color when placed in the sun, and you've done many repeated trials. From this data you can conclude that, SPF45 sunscreen is effective in blocking the ultraviolet radiation from reaching the surface of UV detecting beads. If there are no significant differences between the control and experimental trials , one can show that SPF45 sunscreen is ineffective in blocking ultraviolet radiation from reaching the surface of the UV detecting beads.

 

Reliability. Researchers need to be sure their data is repeatable. So experiments are designed to use multiple trials with many samples in each.

For our purposes, obtaining the results from at least 3 trials is the standard we use to assure good reliability. So if there are less trials, the reliability of the data is questionable. More than 3 would assure good reliability.

 

Validity. An experiment with strong validity means its data is of high quality.

Strong validity means there is clear evidence that a change in Manipulated Variable actually caused the change in Responding Variable

A way to assure strong validity is to do a similar experiment to see if a change in the experiment's manipulated variable actually caused the measured change in the responding variable.

If the results from the second experiment show clearly shows similar results, then the validity is strong. Finally, having strong validity assures the question under investigation is being answered with confidence.

 

Step Four: Record and Analyze Results

Observing and recording data properly and accurately will cut down two major sources of error (Observer & recorder error). Both are simple tasks that are easy to mess up.

Analyzing the results of an experiment can be very simple, like averaging the data, or exceedingly complex. It requires focus and strong deductive thinking skills. Your Data Analysis prepares you to create your final conclusion. The analysis should include:

• A comparison of the experimental trial data to the control trial data

• Search for trends in the trials, data… etc

• Find instances of key data from the experiment to use to compare to the control trial's data (examples, high, low, averages and range)

• A survey of the experimental design, data, and comparisons for errors

• An assessment of the Reliability of the experiment.

• Data tables, graphs, figures, and or photos to support your final results

A powerful analysis will set you up for a strong conclusion.

Magnificent experiments can lead to spectacular embarrassments if errors go undetected in your analysis. Scientific work is required to be open to public scrutiny and have to be able to show others their data upon request! So be meticulous in recording and analyzing your data. Thoroughly think through all possible scenarios before you create your final conclusion.

 Step Five Creating a Conclusion

A conclusion is sort of like the conclusion at the end of a mystery novel. It tells who dun-it, but in the accepted scientific way. Conclusions involve using deductive thinking, a logical method of thinking that compares the final results to the hypothesis.

All conclusions must clearly

___ Tell the Question being answered

___ State the hypothesis & tell if it was accepted or rejected.

___ Tell why the Hypothesis was accepted or rejected by using supporting data from the experimental trials. This include Facts, Reasons, Examples, & Details.

___ Include Evidence, Use key data from the experiment to compare the experimental trial's data to the control trial's data. (Note the emphasis on Using Key Data vs merely displaying data!)

So what does key data mean? Key data depends on the experiment!

Experimental trial and control averages are often compared to show key differences in the data. For example the average increase in height (over time) of the study subject in a growth experiment would be key data to compare and discuss any differences.

Highs and Lows are used to show the variation in the experimental and control data.

Range is often used as an overall indicator of the variation, or more practically, an indicator of the potential error in an experiment. A large range between the High and Low measurements in a controlled experiment indicates a large variation in results. This might indicate a couple things.

First a large range might indicate errors are present in the experiment (poor design, operations, tools etc).

Second a large range might indicate that there is no relationship between the manipulated variable and the responding variable. In other words, a change in the MV doesn't cause a change in the RV

On the other hand, a very small range between the experiment's High and Low might indicate just the opposite, little error in the experiment...

Trends in data can reveal important relationships in an experiment. Thus the researcher should on the lookout for trends. For example if an increase in pressure in an enclosed contained causes an increase in temperature in all trials, this trend would reveal a relationship between temperature and pressure.

(btw: This example reveals that there is a Direct Relationship between temperature and pressure in an enclosed container.)

___ Evaluate & discuss how to improve the data or experiment.

___ Tell all possible sources of error (The six types of error are: experimental design, operator, observer, recording, calculation, measuring tool limitations) and explain how these errors could be minimized or corrected in the future

___ Evaluate the safety procedures/equipment used. (Adequate / if Inadequate tell why)

___ Determine Reliability (Were trials done at least 3X, look at

___ Determine Validity, or at least propose a way to check the experiment's Validity. (show that a change in the MV actually caused a change in the RV).

Conclusions use DATA from the experiment, supported by clear reasoning, to show that the hypothesis is accepted or rejected.

Good facts and good reasoning are key. Scientists want to know what you found out. In fact, scientists often skip down to the conclusion after they have read the initial summary of your experiment to find out the what happened. Afterwards they decide if the researchers decide if they want to read the rest of your work.

Strange, but true. I know this is kind of like reading the beginning and ending of a book before reading the middle!!! That's because researchers are very busy people and don't like to waste time on work that is poorly thought out, not based upon hard facts, and contains sloppy language.

So pay attention to details when writing up both analysis and conclusions!

 

Step Six: After the conclusion, what about the Hypothesis?

a. If the Hypothesis was Rejected the scientist starts over, creating a new Prediction/Hypothesis

b. If the Hypothesis was Accepted the scientist Repeats and Rechecks the Results of the experiment three times (3X).

 

Step Seven: Communicate Results

We started this discussion by saying that good science is many things, SHOERAP.

Three of these things are that science is repeatable, public, and verifiable. Researchers rarely feel good about work that isn't triple checked and verified by the public, so researchers use a Peer Review process to assure their work is fit to print.

Peer Review steps:

a. Make the work available to peer researchers in your field to check.

Good science is published and adds to our knowledge of the natural world. Before publishing you must have a Peer Review done. This occurs when you submit your work to be published. Then several anonymous peers in your field from around the country or world review your data and rethink your conclusions. The peer reviewers verify you have used accepted procedures to get your results. In this step you will have to defend your procedures and results to your peers.

b. Publish the information.

c. Defend the Results to the general public. Once a scientist gets work published that's not the end of the process. Now the scientist must be ready to defend the conclusions to anyone, scientist or not, anywhere in the world.

d. Apply the information This is the main point of experimenting, to add to the world's information so the information can be put to use to solve problems that, hopefully, will make our world a better place for all living things.

 

Some thoughts on Hypothesis, Theory, and Law. There's a big difference between scientific Hypothesis, Theory, and Law!

Many people use these scientific terms loosely. That's not good form!

i. A Hypothesis makes a specific prediction about the outcome of the research. You can also think of the hypothesis as an "educated guess" or prediction about what the research will find. After a hypothesis is tested and strengthened repeatedly by many experiments it becomes accepted as a theory by the scientific community.

ii. A Theory is an explanation of how and why the natural world is structured the way it is. Theories are used to generate hypotheses and testable predictions about the natural world. Not all hypotheses become theories and theories can be disproved. Theories don't become laws, although a theory may "spin off a law" or two!

iii. A Law is a reliable description of nature based upon many experiments that gave supporting observations. A law can be used to accurately and reliably predict what will happen in many situations. Laws don't explain the how and why of things, just describe what will happen.

Ok, with all this knowledge you should be ready to have fun performing experiments!!

 

Scientific Method Index Project Page Home

© 2009 John Schmied & Mike Reid all rights reserved