Understanding Levels of Data and Inferential Statistics
This page covers the fundamental concepts of data levels and inferential statistics in psychology research. It provides definitions, examples, and explanations of key statistical concepts and procedures.
Levels of Data
There are three main levels of data used in psychological research:
Definition: Nominal data refers to categorical information that can be sorted into distinct groups but has no inherent order.
Example: The number of males and females in a psychology class is an example of nominal data.
Definition: Ordinal data can be ranked or put into a meaningful order, but the intervals between values are not necessarily equal.
Example: Scores on a personality test represent ordinal data, as they can be ranked but the differences between scores may not be consistent.
Definition: Interval data provides both a rank order of scores and equal intervals between values, allowing for more precise measurement.
Example: Time, heart rate, and length are examples of interval data in psychology research.
Inferential Statistics
Definition: Inferential statistics are methods that allow psychologists to draw conclusions about populations based on sample data and assess the probability that results occurred by chance.
Key points about inferential statistics:
- They use probability to determine if results likely arose by chance
- In psychology, a p-value of ≤ 0.05 is commonly used as the significance level
- This p-value strikes a balance between avoiding Type 1 and Type 2 errors
Vocabulary:
- Type 1 error: Rejecting a true null hypothesis falsepositive
- Type 2 error: Accepting a false null hypothesis falsenegative
Statistical Tests
The choice of statistical test depends on the level of data and experimental design:
- Nominal data non−parametric: Chi-squared test
- Ordinal data non−parametric: Mann-Whitney U test, Wilcoxon signed-rank test
- Interval data parametric: Independent t-test, Paired t-test
Highlight: Parametric tests are more powerful than non-parametric tests but require interval data.
