Statistics 1 summary for final exam:
Measurement Requirements Examples
level
Nominal Categorical, but no ranking Gender, hair color, type of
accommodation, genotype, religious
preference.
Ordinal Categorical and ranked School ranking, time of day, likert scale,
degree of satisfaction.
Interval Categorical, ranked and width Temperature, IQ, Time.
of intervals known
Ratio Categorical, ranked, width of Age, weight, height, sales figures,
intervals known, and natural 0- number of children, years of education.
point.
Population = the group that you wish to describe
Sample = the group for which you have data
You wish to make inferences about the population with help from the sample.
Parameter = numerical property of the population
Statistic = numerical property of a sample
Sampling error = difference between the value of a parameter and statistic computed to
estimate that parameter
1. Variability: repeated sampling from the same population results in different values for
statistic.
2. Sampling bias: procedures which favor the inclusion in your sample of elements from
the population with certain characteristics.
3. Non Sampling error: all sources of error that are unrelated to sampling.
Reducing sampling error:
- Increasing ‘n’, to decrease variability
- Design sampling procedure, to decrease sampling bias
- Validity, accuracy, precision of variables, prevent coding errors, prevent interpretation
errors, to decrease nonsampling error.
Central limit theorem: even if X is not normally distributed in the population, we may assume
that sampling distribution of the mean is approximately normal with standard error.
Types of sampling:
- Systematic: Population is already ordered in a particular way, like student numbers
which indicate time on uni.
- Cluster: work on the assumption that you have groups in the population that are more
or less similar, it doesn't really matter which one you pick.
Measurement Requirements Examples
level
Nominal Categorical, but no ranking Gender, hair color, type of
accommodation, genotype, religious
preference.
Ordinal Categorical and ranked School ranking, time of day, likert scale,
degree of satisfaction.
Interval Categorical, ranked and width Temperature, IQ, Time.
of intervals known
Ratio Categorical, ranked, width of Age, weight, height, sales figures,
intervals known, and natural 0- number of children, years of education.
point.
Population = the group that you wish to describe
Sample = the group for which you have data
You wish to make inferences about the population with help from the sample.
Parameter = numerical property of the population
Statistic = numerical property of a sample
Sampling error = difference between the value of a parameter and statistic computed to
estimate that parameter
1. Variability: repeated sampling from the same population results in different values for
statistic.
2. Sampling bias: procedures which favor the inclusion in your sample of elements from
the population with certain characteristics.
3. Non Sampling error: all sources of error that are unrelated to sampling.
Reducing sampling error:
- Increasing ‘n’, to decrease variability
- Design sampling procedure, to decrease sampling bias
- Validity, accuracy, precision of variables, prevent coding errors, prevent interpretation
errors, to decrease nonsampling error.
Central limit theorem: even if X is not normally distributed in the population, we may assume
that sampling distribution of the mean is approximately normal with standard error.
Types of sampling:
- Systematic: Population is already ordered in a particular way, like student numbers
which indicate time on uni.
- Cluster: work on the assumption that you have groups in the population that are more
or less similar, it doesn't really matter which one you pick.
