Stats

Question 1

Sample

Answer 1

Information obtained from a small group in order to make valid conclusions about a population. Greater the sample size, better representation of the population.

Question 2

What are the four sampling methods?

Answer 2

Stratified sampling, clustered sampling, systematic and random sampling

Question 3

What is stratified sampling plus pros and cons?

Answer 3

Use prior knowledge of a population in the sample; stratify sample size to fit pop.
+ get more precise estimates of population parameters
- every member must fall into a strata

Question 4

Clustered? + & -?

Answer 4

Divides the population into groups and randomly selects a number of those for sampling
+ can save time and cost of obtaining data
- only suitable if groups heterogenous (not if clusters homogenous)

Question 5

Systematic? + & -?

Answer 5

Numbers the individuals from 1 to n then selects the individuals at regular intervals
+ Is easy and sample spread through the population
- could be a regular pattern of variation in pop so sample could be biased (dampen / amplify pattern) and can render CI’s invalid

Question 6

Random Sampling?

Answer 6

Every member of the population is given an equal chance of getting into the sample. If we don’t use this any CI’s calculated may not be valid- critical to include randomness in all sample types

Question 7

Sampling Frame

Answer 7

A limited group from which we select our sample as it is impractical to sample whole population. This may mean inferences aren’t strictly valid in terms of whole pop.

Question 8

What do we look for in preliminary data analysis?

Answer 8

Descriptive stats: Mean = median = normal distribution

• Stdev = tells us the variability in the data (tight and clustered means small)
• Equal variances = if different can’t use parametric testing (large/small = fcrit < fcalc then different df = n-1)

Question 9

PDA: What does a dotplot show?

Answer 9

Look at the shape of distributions if they’re symmetric or skewed (similar spread) Outliers and best used when comparing one data set

Question 10

What does a boxplot show?

Answer 10

Wanting to find difference in distributions
-Shows box with upper and lower quartiles (25th and 75th) and median sample range. Are used to assess and compare sample distributions between populations or locations