Unit 2 Data Management and Analysis

Question 1

What is parametric statistics

Answer 1

Make assumptions about the
distribution of the data. Usually data need to follow normal distributions. Powerful to detect Type I errors.

Question 2

What is non-parametric statistics

Answer 2

Do not make assumptions. Data
usually transformed into ranks (qualitative, ordinal), i.e.
distribution free. Less powerful, more conservative.

Question 3

What is the central tendency in descriptive statistics

Answer 3

Single value that attempts to describe a set of data by
identifying the central position within that set of data. Also
known as measures of central location.
- The mean or average is the most common.

Question 4

Whats the mode

Answer 4

The most frequently recorded value in a set of data.
There could be more than one mode, e.g. mussel population structure data

Question 5

Whats a sample mean

Answer 5

Calculating sample means is the best estimate of the true mean of a population.
More sampling, better estimation

Question 6

What are measures of variability

Answer 6

numbers that describe the diversity
or dispersion in the distribution of a variable

Question 7

Commonly used variation measures:

Answer 7

Range: crude measure of variability: minimum-maximum
values—Good to know if we’ve made mistakes entering
the data
Interquartile range: uses medians (ie boxplots)
◼ 1st quartile 25th percentile (first 25% values)
◼ 2nd quartile (Median) or 50th percentile (half the values)
◼ 3rd quartile (upper quartile) or 75th percentile

Question 8

Ways variability (error) is calculated

Answer 8

Sum of Squares (SS) = represents a measure of deviation from the mean
Variance proper (sample) (S^2) is a measure of average
variation.
Standard Deviation (S): squared root of S^2 (measured in the
same units as your variable)
Standard Error of the mean (SEOM):
Coeficcient of variation (CV)

Question 9

What are confidence intervals

Answer 9

Derived from the standard error of the mean (SE) Principle: if a sample from a population is very large the true mean of the population has 95% probability of lying within 1.96*SE.
Larger CIs, larger uncertainty….

Question 10

Graphs & tests when want to test for normality

Answer 10

Histograms & Density plots, Q-Q Plots, Skewness & Kurtosis.
Kolmogorov-Smirnov Test, Shapiro-Wilk Test (: more appropriate for sample sizes <50)

Question 11

When do you know whether to use non parametric or parametric tests

Answer 11

Parametric Analysis: Null hyp (Ho) - No sig diff from a normal distribution.
Non-Parametric Analysis: Alt hyp (Ha) - There is a sig diff from normal distribution.