LabiIntroduction to Experimental Design

Joan Stevens, Arcadia High School

AbstractYou will learn the process of designing a controlled experiment using radishes as a testorganism. You will assess the impact of increasing salt concentration on seed growth, andanalyze your data using a t-test and learn issues of statistical probability.

PurposeUnderstanding experimental design is fundamental to knowing how science works. In this labyou will learn the process of designing an experiment to assess impacts of salinity on seedgrowth and the implications for agriculture. You will analyze their data using a t-test to determinethe statistical significance of your results.

Objectives• Design a controlled experiment.

• Use appropriate experimental design vocabulary.

• Assess the impact of salinity on growth of radish seeds.

• Use statistics to interpret and describe experimental results.

BackgroundScience moves forward through the committed action of scientific research. After

observing something of interest a scientist forms a hypothesis that attempts to understand thephenomena. Research and experimentation are designed to test hypotheses. Some researchis exploratory and descriptive in nature (describing the life cycle of a star for example). Otherresearch is based on experiments. If you had observed that plants with in gardens seemed togrow faster than plants in the wild, you could design an experiment based on your hypothesisthat it's fertilizer that helps plants grow. To be valid and accurate an experiment needs to createa basis of comparison between what is being tested and the normal population. In order to dothis, scientists strive to keep conditions constant, or the same, among the groups they areobserving. Ideally, the only things that differ are the experimental treatment and the results atthe end. These are called variable. Controlled experiments are designed to compare

Introduction to Experimental Design

experimental/treatment groups (subjects that have been exposed to a variable of interest) andcontrol groups. The control groups represent the normal conditions. At the end of theexperiment the groups are compared based on a parameter called the dependent variable (theoutcome depends on which group the subjects were in). The dependent variable shouldmeasure something the experimental variable impacts. One example would be measuring theheight of sunflower stems in a group of plants that were given fertilizer and a group that didn'thave fertilizer. Keep in mind that you need to use at least four organisms (and more is usuallybetter!) in each group as a minimum sample size you are testing or you might confuse individualvariation for an experimental effect. Without a large enough sample size you might accidentallyhave chosen seeds that came from especially tall sunflowers to compare to seeds that camefrom very short plants.

In the sunflower example above, we designed the experiment in order to see if thefertilizer would have an effect on the sunflowers. Fertilizer would be considered theindependent variable because it is what the experiment was designed to test.

In this experiment we will use radish seeds as our subjects or test organism. Ourindependent variable will be salt water. We will investigate if the independent variable(concentration of salt in the water) has an effect on the dependent variable. The dependentvariable for this lab is the germination (growth) of the radish seeds. Germination can be definedas the radicle (embryonic root) breaking through the seed coat and emerging as a healthy whitefilament. This experiment is designed to test if different concentrations of salt water have an .effect on the germination of radish seeds. I

In order to understand whether the difference between the dependent variable for acontrol group and an experimental group is big enough to warrant greater attention, scientistsuse statistics. A very simple statistical test is called the t-test. A t-test uses the averagesbetween two groups and the variability between raw scores in a group to calculate how likely theobserved differences in the averages are random or due to chance. A value greater than 0.05indicates that there is more than 5% chance that the difference between the control andtreatment groups is random. Another way of saying this is that the experimental variable had noeffect. If the value is less than 0.05 that implies that the results are most likely (less than a 5%probability) not due to chance but rather due to the effect of your experimental variable.Another way to say this is that the results were significantly different. The use of the wordsignificant in statistics means that there is less than 5% likelihood that the observed results weredue to chance.

This experiment also addresses an issue that is relevant to agriculture today. With ourincreasing misuse of water resources and applications of fertilizers, salts accumulate in soils.This is called salinization. Many farmers find that their harvests decrease each year as saltsbuild up in their soils. How much salt does it take before plants show the effects? We will usevarying concentrations of salt water to address this question.

Labi

Materials and EquipmentFive zip-lock plastic bags and 5 or 10 (depending on thickness) paper towels

Radish seed-50 per group

Salt solutions - 2.5%, 5%, 7.5%, and 10% - To make the 2.5% solution, add 2.5 g of table saltto a 100 mL graduated cylinder and fill to the 100 mL mark. Add the appropriate mass of saltto make each concentration of salt solution.

20 mL Graduated cylinder

Water

Excel or other software program that can perform a t-test. If you have never used this statisticaltest before please use the following websites to familiarize yourself with a t-test.http://www.socialresearchmethods.net/kb/stat_t.phphttp://www.columbia.ed u/ccnmtl/projects/qmss/t_about.html

Procedure1. Read the objectives and the introduction to the lab. Write down a hypothesis about what

effect you think saline solution will have on the germination of radish seeds.2. Gather your materials. Label the five bags as follows: Control, 2.5%, 5%, 7.5%, and

10%. Add your group names or identifier.3. Fold the paper towel to make a rectangular nest for the seeds (one or two paper towels

depending on the thickness of the towel). Fold the edges in so they won't fall out. Putten seeds of approximately equal size in each paper towel nest.

4. Put one nest into each plastic bag. Pour 15 mL of water into the bag labeled "control".Pour 15 mL of the appropriate salt solution into each bag making sure to get the papertowel evenly moist.

5. Place a tie or loose rubber band around your bags to keep them together (avoidsqueezing or compressing the paper towels). Put them in the location identified by yourteacher - preferably somewhere cool and dark.

6. Leave the seeds in the bags for 2-4 days. Germination time will vary depending on theage of the seed and the temperature of the room.

7. Open the bags. Record your observations (mold, odor, condition of the seeds, length ofradicle, etc). Count and record the number of successfully germinated seeds (radicleemerged and white tipped). Dispose of lab supplies as instructed by your teacher.

Introduction to Experimental Design

Data CollectionData Table 1 .

Salineconcentration

# seeds

germinated

Control 2.5% 5% 7.5% 10%

Get the germination data from all the groups in your class. The following instructions are for

Excel. If you are using a different software program for entering data and running statistics

please follow your teachers' instructions. In Excel, enter the lab group numbers in the firstcolumn. Enter the data in the next two columns.

Use column two for the control data (you will needto copy this data over for each lab group in your

class). Record the germination values for the

different salt solution groups in the third column.Use the t-test function to determine if the

difference between the control and experimental

trials is statistically significant.

Grp#

Grp 1

Grp2

Grp 3

Grp 4

Grp 5

Grp#

Grp1

Grp 2

Grp 3

Grp 4

Grp 5

Grp#

Grp 1

Grp 2

Grp 3

Grp 4

Grp 5

Grp#

Grp1

Grp 2

Grp 3

Grp 4

Grp 5

control

10

9

7

10

8

control

control

control

2.50%

7

6

8

10

9

5%

7

8

7

7

7

7.50%

3

7

6s*6

4

10%

2

5

5

1

3

T -value

0.432

T -value

0.13

T -value

0.02

T -value

0.01

Choose a cell in the fourth column and choose

function from the insert menu. In the function

window select statistical and TTEST. The next boxwill ask you to enter the values you want to

compare. In the box that says "Array 1" either

highlight the cells with the control group data or

enter the cell range by hand (B2:B5 for example).For Array 2, highlight the cells (or enter the range

by hand) for that salt concentration. Enter "2" fortails and "1" for type. It should give you a T-value

in the cell you selected.

Here is an example of what your data might look

like in an excel spreadsheet.

Labi

Data AnalysisTo understand your results you must compare the T-test values. If any comparisonsbetween the control and the saline groups have a value that is smaller than 0.05 it meansthat salinity affected germination. Please describe the results of your T-tests. Which

comparisons were significant (<0.05)? Which were not? Make sure to use real T-testvalues in your description.

2. Make two graphs showing average germination results for your group and for the class.Here is an example of what one of your graphs might look like:

Mean Germination of Radishes in Saline Solutions

Jcontrol 2.50% 5% 7.50%

Saline concentration10%

Introduction to Experimental Design

3. Describe any overall trends you observe in your individual group data as well as rn thecombined class data. Does your group follow the same trend as the class or are theydifferent?

4. Error Analysis. Please address any problems your group encountered in completing thelab. Did you follow the procedure as it was meant to be done? What about other potentialsources of error? Were the bags with the seeds in a controlled environment? Could somebags have experienced different temperatures than others? Could some have beenexposed to contaminants more than others? Are there any other inadvertent variables thatmight have influenced your results?

ConclusionPlease explain your results.

1. Which of your saline groups was significantly different than the control group?

2. Were there any trends observed?

3. Did your group data agree with your class data?

4. Was the purpose of the experiment fulfilled?

Labi

5. Was your hypothesis accurate? Why or why not?

6. What was the purpose of the control group?•

7. Do you have a greater understanding of how to design experiments?

Suggestions for Further InvestigationNow that you have a handle on how to design experiments how could you take this lab

to the next level? What other dependent variables could you measure to determine the effect of

saline solutions on radishes? What other independent variables could you test using the radish

lab protocol? Could you do an experiment like this on a bigger scale? How could you avoidsome of the sources of error you encountered in this lab? Would other plants react like the

radishes reacted? How could you investigate? Consider these questions then write a

hypothesis about the effect of saline solution on the growth of radish seedlings. Use the

following format to make a draft of a second experiment.

Experimental Design DiagramTitle: The effect of the IV on the DV. Which concentration of salt water will you choose? Tokeep the experiment simple just choose one concentration of saline. Which will you choose andwhy? What parameterwill you measure to assess the impact that saline has on growth? Stem

height? Leaf area? Plant mass?

Hypothesis: Relate what you think will occur to the DV as you change the level of IV. E.g. If the

IV does this then the DV will do this. Your educated guess as to the outcome of the experiment.

Follow the example of your earlier experiment with radish germination to make a hypothesis

about one level of saline solution and it's impact on some aspect of radish growth.

Introduction to Experimental Design

Independent Variable: Name the variable that you purposefully change during the experiment,include units. Indicate the levels of IV in the columns below.

Repeated Trials: How many data sets were collected; how many times was the experimentdone.

Control: Note the experimental group used for comparison purposes.

Dependent Variable: Name the variable that responds to changes in the IV, include units.

.

Constants: List everything that was kept the same in the experiment. Remember-only the levelof IV should change.

VocabularyControl group—the experimental subjects that represent the "normal" conditions.

Dependent variable—what is evaluated at the end of the experiment to determine if theexperimental and treatment groups are different.

Experimental/treatment group—the experimental subjects that are exposed to the variable ofinterest.

Germination—a seed breaking its seed coat and beginning to grow.

Hypothesis—a proposed explanation for observed phenomenon.

Independent variable—what the experiment is designed to test the effect of.

Radicle—the embryonic root that emerges when a seed germinates.

Salinization—the build up of salts in an area - commonly associated with desertification

Sample size—the number of experimental subjects in an experiment (bigger sample size isusually better).

Variable—aspects of a scientific experiment that are designed to differ