Data processing and statistical analysis

Research process

• Formulation of the research topic

• Overview of the literature

• Objectives and hypotheses

• Development of the research plan and methodology

• Data collection, data processing, and statistical analysis

• Interpretation of results and comparison with data from the literature

• Evaluation of hypotheses

• Conclusions

Hypothesis

A statement or assumption concerning one or more populations. The validity of a hypothesis is tested based on data derived from a sample (or samples).

Qualitative-Categorical

Nominal scale - Eye color, gender, marital status
Ordinal scale - Exam scores, ranking, cancer stage

Quantitative-Numeric

Interval scale - Number of injuries, temperature
Ratio scale - Body weight, age, range of motion

Nominal scale

symbols and numbers used solely for identification (gender, place of residence); they do not possess any of the characteristics of real numbers and cannot be ranked.

Ordinal scale:

numbers and symbols suitable for establishing an order among the objects under study. Differences in the values of the variable are not interpretable (grades, educational attainment, product ratings).

Interval scale:

examines measurable data, has a defined unit of measurement, so differences are meaningful (age, date of birth, intelligence); an absolute zero point is not meaningful.

Ratio scale:

a numerical value that expresses the characteristics of objects. It can be ordered, and the differences in their ratios are meaningful (body weight, height), and it includes an absolute zero point.

Continuous variables

Measures of central tendency
Measures of dispersion

Measures of central tendency

mean median mode

Measures of dispersion

• standard deviation

• IQR

• Range

Categorical variables

Count (frequency) n (%)

Parametric tests

• Greater reliability
• More rigorous testing
• They have strict usability requirements

use bar charts

Non-Parametric tests

• Weaker tests
• No strict usability criteria
• If parametric tests are not applicable, then use them

use box plot