You are on page 1of 4

Module #1: Descriptive Statistics and Estimating with Uncertainty

Lecture 1: Populations, Variables and Samples


(Reading: Whitlock and Schluter Chapters 1, 3 & 4)

I. Statistical Populations
Population ≡ the set of all “subjects” relevant to the scientific hypothesis under
examination (p. 5)
- statistical population
 If the entire population is examined, the collection of data is known as a census.
However, we rarely have this type of information about a population

 Quantities describing a population are called parameters. For example, means


and measures of variation are parameter of the population. These are denoted
by Greek letters.

 The first step in the research process is to carefully define your statistical
population of interest

 For the population under examination, I may be interested in certain variables.


Variables are characteristics that differ among individuals (p.11). Variables can
be of different types (p.11-13):

Categorical Numerical
- qualitative
-characteristics
- don't correspond Continuous Discrete
to degree of diff. on a
numerical scale

NUMERICAL = Quantitative measurements that have magnitude on a numerical scale.


Are numbers always numerical?
- in many cases numbers are not numerical

1. Mohs mineral hardness scale (1- softest mineral, 10 - hardest mineral)


- used number as - no real numeric component - ordinal difference
place holder - it is not quantitative
- scale is not
numeric

1
Are numbers with decimal places always numerical?
- not necessary - just an half rank

Are numerical variables with decimal places always continuous?


- No, it does not necessary
- money continous? it is discrete
- test score- like 0.5 is it continuous? No

We can then measure these variables on different scales.


 Categorical variables
o Nominal scale
- No ordering
like phenotype, eye colour, dead or alive, left/right handed
- categorical - describe charateristics but there is no order

o Ordinal scale
" order"
- rank order
- differences between observations is undefined

 Numerical variables
o Interval scale (celsius, years and species )
- arbitrary zero point
- differences between points are defines
- ratios undefined
- no true

o Ratio scale (mass, length, abundance)

- physically meaningful zero point

- difference and ratio are defined


- ration of measurement
- true zero
2
An example of interval vs ratio:

°C: an interval temp. scale


20 °C - 10 °C = 10 °C (10 degrees difference)
20 °C ÷ 10 °C = 2 (2 times as hot)
- does not have true zero
But what if we change the zero point to where -8 °C is on the current
scale?
28 °C - 18 °C = 10 °C (still 10 deg. difference--great!)
28 °C ÷ 18 °C = 1.56 (1.56 ≠ 2! Oops!)

You as a researcher can often choose the scale of measurement for your variables of
interest. Be careful at the outset of your experimental design. The scale of
measurement you choose, will dictate (can limit) the descriptive statistics and analyses
that you can conduct!!
- ration are the most flexible- most statistical analysis
- nominal are the least flexible - in terms of statistic
- can move variables from ration to nominal but we can't go backwards 55

II. Samples
Sample - the subset of “subjects” selected from a statistical population that are
actually examined during a particular study (p. 5)
 Sample statistics are calculated from the collected sample, and these are used
to estimate the population parameters from the statistical population from
which the sample was drawn. Roman Letters for sample statistics.
- normal letters - a,b,c....

 In most biological studies, we are interested in using the characteristics of a


sample to infer the characteristics of the population from which the sample
was drawn

 Therefore, we want to sample to be as representative of the population as


possible - it's through random sampling

 The way that we ensure we have a good sample is by random sampling (p. 5-9)
3
1. o Equal chance of being selected
high low sample size
2. o Independent sample size
Precise Imprecise

random sampling Fig 1.2-2

Unbiased
Accurate

inaccurate
Biased
non-random sampling

 What are some methods you can use to collect random samples for an
experiment?
 What are some of the concerns with haphazard sampling? Volunteer bias?
In statistics, when a sample mean (𝑌̅) differs from an expected/hypothesized
population mean (µ) we will always have two competing ideas about why they differ:
1) The difference is ‘real’ 2) The difference is due to sampling error alone.
when we have bias in sample

6. All of the following statements about random sampling are correct except:
a) The basic unit for random sampling is a single individual organism.
b) All individuals in the population of interest have an equal probability of
being included in a study.
c) The selection of one member of the population for inclusion in a study has
no influence on the probability of the selection of another member of the
population.
d) Random sampling ensures no systematic discrepancy between the sample
estimate and the true population characteristic. - diff in random way not systematical way
e) Random numbers used to randomly sample should always be generated
with the aid of a computer.

You might also like