HOW TO DO A SCIENCE OR MATHEMATICS RESEARCH PROJECT

1. First, make sure you have a research notebook. A spiral notebook with pockets is a good suggestion. Suggestions on what to put in a notebook or how to organize it are listed in the “Research Notebook” section.

2. Research starts because you have a question about something that interests you. This may come from reading, preliminary experimentation or observing something in nature. It may come from a question that arose from something you studied or discussed in a science course or it might be related to some interest you have from outside of school. Start your project by examining your interests, even things that bother you. Among your interests will be some questions that you would like answers to.  

3. You may have several interesting questions. Use your research notebook to list each of them. For each question, determine a preliminary plan of attack by answering the following questions based on the scientific method:

(a) What is the hypothesis or objective? This should address what you want to test and what you think your experimental outcome will be. This should make you start to think about experimental process. 

(b) What measurements need to be made? Do I have the equipment necessary for making these measurements accurately?

(c) Do I have enough time to properly complete the experimental testing?

(d) How can I control all variables except the one being tested? The greater the number of variables, the more carefully you must design your study to eliminate the possible effects of these variables on the significance of your results.

(e) How many subjects or tests will be necessary to validate my conclusions? Plan to use a sufficient number of subjects or test runs to make the results of your study significant. If you are in doubt, check with your science or math teachers and/or consult some references on statistics.

 (f) What kind of statistical analysis is appropriate to show significant results in my study?

4. Determine which question you are going to test.  It might be helpful to discuss your choices with a teacher, mentor or colleague. If you have a problem that requires experts in a specific area, the State KJAS Director might be able to put you in contact with someone with expertise in your research area.

5. Start gathering background information from scientific journals (online or from the library), books or credible web sites.

6. Begin writing the introduction to your scientific paper. (See section on “Writing a Good Scientific Paper”)

7. Start by roughly outlining or making a bulleted list of experimental instructions in your notebook. This will eventually be written into paragraph form and become the Methods section in your scientific paper. You will need to edit the outline or list as necessary to accurately reflect what you did as you  went through the testing phase of your experimental process. When complete, this should specifically describe what you will do in your experiment. It is like a recipe for cooking. Your experiment has to be reproducible and if someone can’t duplicate your work by reading your methods, your research won’t be valid.

KEEP THE FOLLOWING IN MIND WHEN DESIGNING YOUR EXPERIMENT:

a. Change one variable at a time. Don't change both temperature and concentration in the same experiment. Do two experiments.

b. Run a blank or control. A blank involves non-living components while a control involves living components. They are used to determine what happens normally.

For example, iodine is being measured in spinach and several reagents were added to make this measurement. The results may be in error if there is iodine already in the reagents. Therefore you must determine if there is iodine in the reagent and if so the amount of iodine present in the reagent without the spinach. This is a BLANK.

Ten people are to be tested for the effects of caffeine on reducing headaches. Five are allowed to drink 24 ounces of soda containing caffeine (the test) and five are given caffeine-free soda (the control). The time for the headaches to leave is then measured and the test is compared to what would occur normally. This exemplifies a CONTROL.

c. Run a blind study. Have someone make unknowns for you to test without telling you what they are. This eliminates any personal bias that may influence the results.

For example, have someone other than you prepare and code the samples before the ten people above are tested. When you then measure the times, you cannot be biased knowing which people have the test drug. This is a BLIND study.

 8. Start gathering the equipment you need. Find out what is available, what you might be able to borrow, what you need to purchase and what you might have to build.

9. Begin your testing. Make sure you have isolated all the possible variables except the one being tested, that you have planned for enough trials to make your research valid and that you have replications of your experimental units . Replication helps to reduce random results that may be caused by a fluke or accidental error in your experimental method.  For example, if you are testing what amount of fertilizer is most effective on plant growth, you need three replicates of each variable (3 plants with no fertilizer to act as the control, 3 plants with 5 ml of fertilizer, 3 plants with 10 ml of fertilizer and 3 plants with 20 ml of fertilizer). You also need to do at least 3 trials. So you would have at least 36 experimental plant units in this experiment.

10. Do a preliminary analysis of your experimental results. (NOTE: The whole reason for the research is to answer this question: DID MY EXPERIMENTAL UNITS SHOW SIGNIFICANT DIFFERENCES FROM MY CONTROL UNITS AFTER BEING TESTED). A significant difference is determined by applying the appropriate statistical analysis. You cannot say your experiment showed a significant difference just by observing alone. Don’t use any statistics you don’t understand or can’t explain to someone else. Prepare any graphs or data tables that will let you better see and understand your results. Discuss any results which appear to be inconsistent or just don’t seem to fit with your teacher, mentor or colleagues. You might find you need to reevaluate your testing methods. (Using the plant experiment, you might find that all plants died. This might mean there was residue of something in the container you planted in or watered in so you will have to start over.)

11. If you feel your results are acceptable, create final graphs and tables and begin writing the results section of your paper. (See section on “Writing a Good Scientific Paper”) Do not report too many significant figures. Discuss with your teacher the correct number of significant figures to use. If you did the experiment again, could you get the same number?

12. Rewrite your Methods into a paragraph form.

13. Make conclusions about your research results and write the conclusions section of your paper. BE CAREFUL. Experienced researchers are usually cautious in the conclusions they draw until they have repeated the work several times. When writing your conclusions, do not say that your project "proved" something. Scientists conclude only that their studies supported or rejected their hypothesis--and only under the conditions that the research was carried out. (See section on “Writing a Good Scientific Paper”)

14. Finish writing your paper. CAREFULLY READ OVER IT WHEN DONE. It is obvious to other readers when you have not carefully read over your own paper. Also have at least one other person read and edit for misspellings, grammar and formatting mistakes.

15. Finally, write an abstract. (See section on “Writing a Good Scientific Paper”)