Transcript Breeding and Non-breeding Survival of Lesser Prairie

MANAGEMENT AND ANALYSIS OF
WILDLIFE BIOLOGY DATA
Bret A. Collier1 and T. Wayne Schwertner2
1Institute
of Renewable Natural Resources, Texas A&M
University, College Station, TX 77845, USA
2Department of Animal Sciences and Wildlife
Management, Tarleton State University, Stephenville,
TX 76402
Introduction

In wildlife biology, data analysis underlies nearly all the
research that is conducted

The range of statistical methods available is extensive

Ultimately, good questions, study designs, and analysis
are complementary topics
First Thoughts

When designing a study: Talk to a professional

No amount of statistical exorcism can fix a bad study
design

Methods are rapidly advancing, staying in front is tough

Again: When designing a study: Talk to a professional
Study Design

In scientific research, results hinge on study design

Define population of interest





Ecological populations
Inferential populations
Target populations
Sampled populations
Population inference requires data representing
population of interest
Data Collection

Conceptual framework for ‘how’ to collect
 1. Outline study question.
 2. Define response variable (e.g., nest survival).
 3. Define explanatory and/or descriptive variables that might affect
response (e.g., vegetation cover).
 4. Define steps for minimizing missing data.
 5. Outline data collection approach.
 6. Design initial data collection instrument specific to response or
explanatory variables.
 7. Conduct field test of protocols and data instruments.
 8. Evaluate efficiency of data instruments.
 9. Repeat steps 2–8 if necessary due to logistical difficulties.
 10. Initiate data collection.
Data Management

Data types
 Qualitative
 Quantitative

Data measurement scales





Nominal
Ordinal
Interval
Ratio
Data files
 Files containing all data in rows and columns
 Commonly put into spreadsheets
 More advantageous-database management system
Data Presentation

Tables and Graphs

Variety of uses
Bar Graphs

Bar Plots
Point Graphs

Point Plots
Dot Graphs

Dot Plots
Scatter Graphs

Scatter Plots
Hypothesis Development

Good questions come from good hypotheses about how
a process occurs

Statistical models can help evaluate strength, or lack
thereof, of how a process occurs

Models should inform the ecological question, not drive
the question
Hypothesis Development

Good questions come from good hypotheses about how
a process occurs

Statistical models can help evaluate strength, or lack
thereof, of how a process occurs

Models should inform the ecological question, not drive
the question
Inference

Descriptive Statistics
 Mean
 Mode
 Median
 Variance
 Standard Deviation
 Standard Error
 Confidence Intervals
Comparative Analyses

Chi-square tests

T-tests

F-tests (Analysis of Variance)

Correlation
Regression Analyses

Linear Regression

Multiple Regression

Generalized Linear Models
Community Analysis

Wildlife research has traditionally focused on the
population level.

Some study questions, however, address how wildlife
communities:
 Respond to management activities or other perturbations
 Biodiversity is affected by various activities
 Change across space and time
Species Richness

Number of species in a community.

Strongly influenced by sample size.
 Makes comparisons difficult.
Complete Enumeration
 Provides the minimum number of species present.
 Works for simple communities.
 Rarely possible.
Richness Indices
 Margalef’s index
►Not
an estimate.
►Cannot
be compared with other indices or richness
estimates.
►Strongly
influenced by sample size.
Richness Estimates
 Estimate the actual number of species in the community
 Data collected as a single sample
► Rarefaction
 Used for standardizing sample sizes, and the resulting estimates of species
richness, among samples.
► Chao
1 Method
 Especially useful when a sample is dominated by rare species.
 Requires species abundance data.
 Data collected as a series of samples.
► Chao
2 Method
 Modified Chao 1
 Can be used with presence-absence data
► Jackknife and
Bootstrap estimates
 Involve systematically resampling the original dataset.
Species Heterogeneity
 Measures the degree to which individuals in a
community are distributed among the species present.
►Shannon-Weiner
Function
 Based on information theory
 Measures the amount of uncertainty associated with predicting the
species of the next individual to be collected.
►Simpson
Index
 The probability that 2 individuals drawn randomly from a community
will be same species.