Paper 5) Practical Skills Flashcards
standard error (S.E)
- reliability of the calculated mean
- larger the S.E, less reliable.
standard deviation (s)
- spread of data around the mean
- narrow spread = more reliable
you can then calculate : - 95% CL
- error bars
95% confidence interval
- 95% probability that true mean lies within this range
- to see whether error bars overlap
- to get a measure of how close the calculated mean is likely to be true
error bars
- how close the calculated mean is likely to be true to the actual mean
- overlap in error bars = no sig. differences between 2 means
Describe how you could use belt transects to collect data needed to determine Simpson’s
index of diversity
using a belt transect:
- use of tape / line / string, and quadrat (to create belt transect)
- selecting a start point for the, transect / tape / line / string
or
all transects same length, in plots / plantations / repeats
- same (size) quadrat
- use quadrat at, regular intervals / stated distances along the, tape / line / string / transect
or
continuous belt transect
collecting data:
- method to identify each of the plant (species in the quadrat)
- count / record / note the number of (individuals), each plant species (n) (in each quadrat / at each sampling point)
or
estimate percentage (%) cover of each plant species
- use at least 3 different transects (in each plot)
- named hazard and risk and precaution
percentage change formula
(new - old) / old
x100%
improvements for confidence in results
- more intermediate values (to get better trend/pattern)
- more range of values/ increase sample size
- exclude anomalous results (improve reliability)
- repeat investigation at least 2 (improve reliability)
- calculate standard error/deviation/confidence levels (check for significance)
- carry out in dark room
variables to be standardised
- temperature
- humidity
- light intensity
- CO2 concentration
- mass of soil
- depth of soil
null hypothesis
there is NO significant difference between….
what axis does the independent variable go on
X - axis (horizontal)
what axis does the dependent variable go on
Y - axis (vertical)
why might you not have enough confidence in results
- many anomalies
- no statistical test carried out
- measured to nearest cm/rounded?
hazards
- irritants
- allergies
- burns
how can a sampling method be improved
- sample at different times of the year
- sample at different times of day
- sample at different sites/other than along the path
- use expert/guidebook/app to identify individuals
- carry out statistical analysis
rate
percentage/time
x100%
why calculate percentages ?
- to make valid comparisons
- if number of samples are varied
when is the t-test suitable?
- continuous data
- normal distribution
- comparing 2 means
- s.d apprx same
why use Pearson’s ?
- data is paired/linked
- data is continuous & normally distributed
- scatter diagram suggests linear relationship
- at least 5 paired observations
why use Spearman’s?
- data is paired/linked
- both data are independent of each other
- data is interval/discrete
- scatter graph suggests correlation
- more than 5 paired observations
what does an rs value of 0.85 and rs value of -0.85 mean
0.85 = strong positive correlation
-0.85 = strong negative correlation
what does an rs value of 0.1 and rs value of -0.1 mean
0.1 = weak positive correlation
-0.1 = weak negative correlation
degrees of freedom for t-test
n1 + n2 -2
degrees of freedom for Pearson’s/Spearman’s
n - 2
degrees of freedom for Chi-squared test
n - 1
