EXERCISE NO.

OVERVIEW OF STATISTICAL METHODS FOR ECOLOGICAL RESEARCH

Ecology relies on both qualitative and quantitative observations. Many activities

in ecology involve counting and measuring variables (e.g., length, light intensity,
distance, humidity) in order to quantitatively characterize organisms, populations,
ecosystems, and their environment. These sets of measurements allow us to objectively
make comparisons and deduce relationships between spatial and physical entities (e.g.,
sites, populations, habitats) and between temporal points (e.g., changes in the same site
over time) through statistical analyses.

This exercise will be divided into three activities: first on Descriptive Statistics,
second on Inferential Statistics, and the third exercise will deal with establishing
relationships among variables.

Exercise 1.1 Descriptive Statistics: Using Data Summaries and Visual Tools to
Present Results

Introduction

Prior to any statistical analyses, a simple descriptive statistics allows us to

summarize the data in order to draw up a good picture of the samples and populations
we are studying. Measurements such as the mean and standard deviation give us an
idea how uniform or how spread-out our measurements are. Data summaries, presented
as graphs and tables allow readers and evaluators to get a quick grasp of the
characteristics of the results of the study.

Objectives
At the end of this exercise, students are expected to be able to:
1. Carry out a quick survey using measurements of various ecological
parameters.
2. Summarize data sets using common statistical measures.
3. Present data summaries in tabular and graphical form.
Materials
Transect line, meter stick or any measuring tape at least 1 m in length
Plastic ruler
Stakes
String
Light meter
Cutter or scissors
Camera
Gloves
Garbage bags

Procedure
1. Find a fairly large (i.e., between 50 and 100 m2) vegetated space where you will find
at least three (3) species of plants growing. Using your transect line, estimate as best
as you can the total area cover (in m2) of your study site.
2. Randomly select 10 areas and, using the transect line, stakes and string, mark out
square or rectangular areas. Note that the marked out areas need not have the same
dimensions. Record the dimensions of each area.
3. In all 10 areas, count the number of individual plants belonging to the 3 species.
4. Using a plastic ruler, measure the height of each plant counted.
5. Measure and record at least 3 light readings in each of the 10 areas. Note that this
should be carried out in the shortest time span possible so as to make sure that all
readings from all 10 areas are comparable (i.e., measured at approximately the same
time). Record the time when the measurements were taken for each area.
6. Fill out the following table with your raw data.

Dimension
Light Counts Height (cm)
Sampling Length Width measurements
sp sp
Point (m) (m) (Time taken: sp 1 sp 2 sp 3 sp 3
1 2
________)
1st reading

2nd reading

3rd reading

7. Fill out the following table with your data summaries.

Sampling Mean counts per sq

Mean Mean height sd*
Points Computed meter sd*
Light
Area
Intensity sp 1 sp 2 sp 3 sp 1 sp 2
sp 3
1

3
 4

10
*sample standard deviation: !!!
!!!
7. Construct bar graphs of (1) the mean density and (2) the mean height for all 10 areas.
Use the sample standard deviation values for your error bars. Assign a specific color,
shading or stippling pattern for each species. A sample is shown below.

Mean Density
12
Mean Density (counts m-2)

10

6 sp1

4 sp2

2 sp3

0
1 2 3 4 5 6 7 8 9 10
Area

8. Construct scatterplots of (1) mean density vs mean light intensity and (2) mean
heights vs mean light intensity. See sample below.
 Mean Counts vs Light Intensity
10

Mean Density (counts m-2) 8

6
sp1
4 sp2

2 sp3

0
100 120 140 160 180 200 220 240
Light Intensity (lux)

Guide Questions

1. Why is sampling important in ecology? How important is random sampling?

How important is the size and number of your sample areas in sufficiently
representing the study site that you chose for this exercise?
2. What is the value of summarizing your data as opposed to merely presenting
raw data?
3. How easily can trends or patterns be shown in your data even without the
benefit of inferential statistical analyses? Compare how trends and patterns
are discerned in the raw data or in the summarized data. What pieces of
information can potentially be lost when we summarize data?
4. If your counts were not normalized with respect to unit area (i.e., the counts
represent different area coverage), how would this have affected the
presentation of results?
5. How important are tables and graphs in presenting results? How important is
it to select the appropriate tables and graphs in presenting results? How
differently would the presentation of your data be had you used other types
of graphs?

Suggested Reference:
Brower, J., J Zar, CN von Ende. 1997. Field and Laboratory Methods for General
Ecology. 4th ed. McGraw-Hill Education, 288pp.
Exercise 1.2 Inferential Statistics: Comparing Between or Among Groups

Introduction

Ecology is not merely concerned with the description of a particular site or

community in terms of the prevailing environmental conditions, the species composition,
or the relative abundances of organisms in a site. What makes ecology interesting is that
it allows us to make inferences, drawing from our meticulous observations, about the
different communities exposed to unique sets of environmental variables, about the
response of organisms to a particular natural (or artificially introduced or modified)
environmental cue or introduction of a species, or about the temporal shifts in the
structure or function of any ecological entity over time. However, these inferences need
to be drawn using established methods that can assure objectivity and randomness.
Statistical methods that can provide comparisons between two (2) groups or among
three (3) or more groups are an essential tool for the ecologist.

Objectives

At the end of this exercise, students are expected to:

1. Perform comparative statistical methods such as the t-test or ANOVA
(Analysis of Variance)
2. Detect statistically significant differences among groups of measurements
and weigh if the difference is also biologically or ecologically significant.
3. Explain differences (or non-differences) that they are seeing in their data.

Materials
Data from Exercise 1.1
Calculator or Laptop with statistical software

Procedures

1. Collect the class data from the previous exercise. Each group should have a
copy of the data from all other groups.
2. Find two sets of data (i.e., data from two groups in class) that had the same
plant as their most dominant species. Fill out the table below:
 Mean
Mean counts per Mean height
Area Light
sq meter
Intensity

Group A

Group B

3. With n=10 for each group, perform a t-test for the mean light intensity, mean
counts and mean height. Determine the appropriate t-test and state your
hypotheses for each test.
4. Using an alpha level of 0.05, interpret your t-test results and provide a logical
and empirically supported explanation for each of your conclusions.
 5. Find the species of plant that was common among the most number of
groups in class. Use this criterion in selecting the data sets that will be used
for the next part of this exercise.
6. Create a table similar to that in step 2. The data sets (n=10 per set) should be
describing only one plant species.
7. Perform a single factor analyses of variance on using light, mean counts and
mean height as your dependent variable and group as your independent
variable ( = 0.05). Provide the hypotheses for each analysis. Test the
assumptions of the ANOVA to determine if your results are valid or not.
8. In cases where the assumptions are not met, perform data transformations
until the assumptions are satisfied. If attempts still fail, resort to a
nonparametric counterpart of ANOVA (e.g., Kruskal-Wallis test).
9. In analyses where the null hypothesis is rejected, perform a multiple
comparisons test to determine which group/s is/are different.
10. Interpret and discuss your results. Additionally, present your comparisons
using appropriate graphs.
Guide Questions
1. Why are statistical analyses important in ecology?
2. How important is objectivity in drawing your conclusions?
3. How do we determine if statistically significant results are
biologically/ecologically significant as well?
4. In trying to explain significant differences among the groups, how can one
explain the observed differences better? Is there a need for more
measurements of the same parameter or more measurements of other
parameters?
5. How important is accuracy (i.e., how close your measurements are to the true
value) and precision (i.e., how consistently close your individual
measurements are to each other) in your data sets when trying to compare
among groups?

Suggested References:

Brower, J., J Zar, CN von Ende. 1997. Field and Laboratory Methods for General
Ecology. 4th ed. McGraw-Hill Education, 288pp.

van Emden, HF. 2008. Statistics for Terrified Biologists. Blackwell Publishing: UK.
343pp.
 Exercise 1.3 Establishing Relationships among Variables

Introduction

An important concept in ecology is that all things are connected. Biotic and
abiotic components of ecological units are constantly interacting, thus, can be influenced
by changes in other components. Aside from describing and comparing various entities,
an equally crucial aspect in ecological study is the ability to explain the measurements
we are getting and why one group is different from another. When manipulative
experimentation is not feasible, we can establish relationships among different variables
in order to help us understand what we are observing. The simplest way to establish
relationship is to determine if variables have a significant straight line relationship or not
by correlation and regression analyses.

Objectives

Materials

Class data from Exercise 1.1

Calculator or laptop with statistical software

Procedure

1. Fill up the table below with the class data from Exercise 1.1.
Mean counts per sq Mean height (cm) Present (1) or Absent
Mean meter (0)
Area Light
Group
(m2) Intensity sp sp sp sp sp sp sp sp sp
(lux) k
1 2 k 1 2 k 1 2

1
Note that data columns for mean counts, mean height, and presence/absence should
contain data for a particular species only. Thus if there were 12 species recorded for the
entire class, there should be 12 columns (i.e., sp1, s2,sp12) for mean counts, mean
height and presence/absence.
2. Run a simple correlation analysis comparing all variables pairwise. Determine
significant correlations.
3. Run multiple regression analyses on variables that you think can be explained by
other variables. (Ex. The mean counts of species 3 can be predicted by a
combination of light intensity, the presence or absence of species 1 and the mean
counts of species).
4. Take note of regression parameters that are generated (e.g., r2, constant,
regression coefficient, etc.) that can be used in interpreting the analysis.
5. Interpret and discuss your results.
Guide Questions

1. Why is it important to establish mathematical relationships among ecological

variables?
2. In what instances would linear regression fail to establish a relationship between
variables?
3. How will autocorrelated variables influence your regression results?