Statistical Analysis of Bean Growth Shank 1

Statistical Analysis of Bean Growth

Introduction
The observation that generated the concept for this experiment was that pinto beans are able to
germinate between eight to fourteen days when given proper amounts of water and sunlight. The
question that came from this initial observation was “If different types of water brands were
distributed to a certain bean, what would the effect be on the rate of germination in beans?” From
this question, an alternative and null hypothesis were able to be made. The null hypothesis was “I
hypothesize that the nutrients in the water do not have an influence on the germination rate of
beans.” The alternative hypothesis was “I hypothesize that the type of water introduced to the
bean will have an impact on the rate of germination due to the different nutrients in each water
sample.” I predict that the introduction of water with additives, such as minerals, antioxidants,
and preservatives, would decrease the rate of germination due to the fact that it has different
amounts of minerals than found in natural water.

Methods
The experiment that was performed was a manipulative experiment that used different water
treatments of the same quantity to determine if there was a relationship between water type and
the overall growth of a pinto bean plant. It was a snapshot study that collected data from the bean
plants at the same time over a period of 15 days. For this experiment, the greenhouse located in
the Arts and Science Center of Stockton University was utilized in order to try to combat
confounding factors in the environment. Some confounding factors that could not be fully
eliminated would be if the initial health and size of the beans as well as the slight difference of
germination rates in each of the beans. In order to decrease the impact of these confounding
factors, Random.org was utilized to create five randomized blocks. These five blocks were the
location of the pots on the shelf in the greenhouse. There was a block in the front right, front left,
back right, back left, and the middle in the very back of all of the pots.

For this experiment, there were seven different water types which were tonic water, antioxidant
water, tap water, filtered water, seltzer, Core water, Essentia water, and S. Pellegrino sparkling
natural mineral water. 25 milliliters of water were added to each plant depending on which water
treatment it was assigned. The water types would be the independent variable and the overall
growth of the plant was the dependent variable. Three main things were considered at the end of
the experiment to determine if significant growth was present: root abundance, beanstalk height,
and the total surface area of leaves. The water treatment was categorical and the three plant
growth factors were continuous, which would mean that an ANOVA statistical analysis test
should be used. After inputting the data into Excel, graphs were made in order to compare the
treatments and the growth factors. When running the data through RStudio 1.2.1, three separate
single-factor ANOVA tests were used to test significance. For this experiment, significance was
determined if there was a p-value that was less than or equal to 0.05.

The experiment was started on October 15th and continued until October 23rd. Each of these
days, the plants were examined to determine if the soil was dry and needed more water. October
17th was the first day that height measurements were taken. Prior to this, qualitative observations
Statistical Analysis of Bean Growth Shank 2

were made about the presence of growth. After the 17th, data was collected on the growth of the
beans on the 18th, 21st, and 23rd. This data was collected using a ruler. At the end of the
experiment, the plants were removed from their pot in order to collect data on the three factors of
growth. In order to collect data on root abundance, the highest and lowest height of root growth
was collected as it held its form from when it was in the pot. Using the average of these numbers
as well as the diameter of the pot, a formula was used in order to determine the volume of a
cylinder. The beanstalk height was found by measuring the stalk from where the stalk met the
roots and where the top leaves resided. This measurement was taken in centimeters. The total
leaf surface area was found by tracing the leaves onto graph paper which had grids that were
squares of one centimeter by one centimeter. If the leaf took up a full block, that would be one
centimeter, but if it took up less than that it would be placed into either a more than half or less
than half category. The sum of these categories for each plant was taken and then added to the
whole boxes. If there were more than one leaf, the total sum of boxes from each leaf would be
added together. This measurement of the total leaf surface area is in square centimeters.

Results
For the treatment with tap water, there was no growth of the beans. In the treatment with the
antioxidant water, only one of the beans had growth. The treatment with the filtered water there
was growth in three of the beans, but only two of these beans had growth in all three categories.
The Core treatment had two beans that had growth. The treatment with Essentia water had three
beans experience growth, but two that fulfilled all three of the categories. For the sparkling
natural mineral water treatment, four of the beans experienced all three growth factors. Finally,
the tonic water had two beans which developed roots.

Figure 1: A bar graph comparing the different influences water treatment had on stalk height,
root abundance, and the total leaf surface area all of which were measured in centimeters.
Statistical Analysis of Bean Growth Shank 3

When conducting the single-factor ANOVA analysis test of the relationship between root
abundance and water treatment, it was found that the p-value is 0.388. The second single-factor
ANOVA analysis test was between water treatment and stalk height showed that the p-value
would be 0.229. The final single-factor ANOVA analysis test had resulted in a p-value of 0.376
between total leaf surface area and water treatment.

Discussion
The results of this experiment showed that there was no significant relationship between the
water treatments and the growth of the pinto beans. The original question that sprouted this
experiment was “If different types of water brands were distributed to a certain bean, what would
the effect be on the rate of germination in beans?” The results answered this question because it
showed that there was no significant effect on the germination of the pinto beans. With the
p-value of the relationship between root abundance and water treatment showed in the results
section, since it is more than the established p-value of 0.05, the null hypothesis fails to be
rejected and thus the relationship is insignificant. With the p-value between stalk height in
centimeters and water treatment, the p-value was found to be above the established acceptable
p-value of 0.05 and thus yields the same result as the first single-factor ANOVA test. The third
single-factor ANOVA test between the total leaf surface area and water treatment was found to
have a p-value of more than the accepted p-value of 0.05 which would lead to the same
conclusion to be drawn from the other two single-factor ANOVA tests. When failing to reject the
null hypotheses for all three of these relationships, it also means that the alternative hypothesis is
failed to be accepted.

With the results, it is shown that waters that are for human consumption all play similar roles on
plant growth. The data makes sense because these waters that are sold for human consumption
have to pass testing and tap water is known to be drinkable by the average person. This would
mean that vegetation is likely to accept water sources in a similar manner. For example, some of
the waters used in this experiment held different pH values which also translates to the
introduction of acidic rain. Since there was no significant relationship found between the
different pH values in the different waters and the overall plant growth, this would then translate
to the natural world and the relationship between the growth of vegetation and the acidity of rain
in the sense that the introduction would not be significant enough to alter the biomass of an
ecosystem.

Some confounding factors that could have impacted the results of this experiment is the original
bean mass, the original bean size, and the original bean quality. Some limitations of this
experiment that could have impacted the results would be too few replications. It might have
been beneficial to the experiment to test the impact on different types of seeds to see if the same
relationship translated to other plant families. If conducting this experiment again, I would
suggest tested multiple different plant species and have more replications in order to reduce the
impact of confounding factors. For example, I would want to try to see the relationship between
the three growth factors and the water treatment with a few types of flowers, fruits, vegetables,
grasses, mosses, and maybe even trees. In the future, this could also lead the way to test native
plants and alien plant species to see if foreign plants have more susceptibility to being more
Statistical Analysis of Bean Growth Shank 4

vulnerable to change from its average growth. This would then help researchers to determine if
alien plant species are something to try to protect against or if their presence there is so fragile
that their presence is not significant in the quality and health of an ecosystem.

Appendix
The single factor ANOVA statistical testing of significance between root abundance (cm) and the
water treatment:
Df Sum Sq Mean Sq F value Pr(>F)

Treatment 1 9669 9669 0.764 0.388

The single factor ANOVA statistical testing of significance between stalk height (cm) and the
water treatment:
Df Sum Sq Mean Sq F value Pr(>F)

Treatment 1 83.5 83.47 1.503 0.229

The single factor ANOVA statistical testing of significance between leaf surface area (cm^2) and
the water treatment:
Df Sum Sq Mean Sq F value Pr(>F)

Treatment 1 4944 4944 0.805 0.376