Data Analysis & Measurement in Psychology: Scientific Method Foundations

Posted on Sep 25, 2025 in Statistics

Data Analysis and Measurement in Psychology: The Scientific Method

The objective of scientific method studies is to conduct procedures that are systematic (with established steps) and verifiable (with data that can be replicated or refuted by any researcher). However, the scientific method is just one component of the scientific research process, which consists of three levels (Arnaud):

1. Theoretical and Conceptual Level

   1. Defining the problem and hypotheses
   2. Deduction of testable predictions

2. Theoretical-Methodological Level

   3. Establishing a data collection procedure (choosing a research design)

3. Statistics-Analytic Level

   4. Data analysis (to test hypotheses)

Interpreting the results obtained (5. Discussion of results and conclusions, 6. Elaboration of the research report) leads back to the first level of the process, creating a cyclic structure.

You should always work with a set of data that describes the study you are conducting.

Understanding Statistics and Data Analysis Levels

Statistics is the science that collects, orders, and analyzes data from a sample to understand a true population. From this, with the calculation of probabilities, it makes inferences about the population (it is a fundamental branch of mathematics for empirical science). Data analysis can be done at the following levels:

1. Descriptive Analysis

   Natural behaviors are observed, and descriptive data measurement is performed, identifying frequencies and percentages, calculating correlations, and representing data graphically.

2. Inferential Analysis

   A sample represents the entire population. Based on descriptive results, it aims to determine any errors or probabilities related to the population.

3. Experimental Analysis

   Experimental conditions are created by manipulating the independent variable (IV) to obtain values. It requires assigning groups to compare the effects of experimental conditions. Depending on the degree of manipulation of the IV, we have:

   • Observational Design: Minimal or no manipulation of variables (e.g., naturalistic observation or field studies).
   • Quasi-experimental Design: Uses non-equivalent groups.
   • Correlational or Selective Design: Studies the relationships between variables by selecting subjects based on certain characteristics.
   • Experimental Design: Maximum extent of manipulation (e.g., lab experiment).

Once the problem is defined, a model is proposed that allows, through mathematical development or formal languages, to elaborate predictions that can be empirically tested. Statistics is involved in contrasting the model’s predictions with the obtained results (acting as a bridge between mathematical models and real phenomena).

The adequacy of experimental results and the model’s predictions consistently confirms the explanations; discrepancies necessitate the reformulation of the model and a new start to the process.

Also, we can distinguish:

• Theoretical Statistics: Focuses on theoretical aspects and formal regulations.
• Applied Statistics: Application to a particular area, in our case, statistics applied to data analysis in psychology.

Data Analysis in Psychology

Methodological and statistical tools needed for research in psychology.

General Concepts in Statistical Research

Key terms include: Population, Sample, Parameter, Statistic, Feature, and Mode.

The research objective is to characterize the properties of a Population (PB): the entire set of elements (statistical entities) that share one or more characteristics, which can be finite or infinite.

This characterization is done using numerical values (e.g., percentages, means) that summarize the population: Parameters (describing the population’s properties).

Although generally, the total population is inaccessible to the researcher, a Sample must be used: a representative subset of elements from the population. It is fundamental for the sample to be representative, and statistics includes sampling theory, which studies methods for representative sample extraction.

From that sample, researchers calculate summary values: Statistics (describing the sample’s properties), and thereby infer results to determine the population’s properties.

Although the true values of parameters are unknown, they are symbolized by Greek letters (e.g., ?, ?, ?). Statistics, which can have many values depending on the sample taken, are symbolized by Latin letters (e.g., , P ).

The properties of a population are present in all its constituent elements, although not to the same degree (e.g., marital status: widowed, divorced, single, etc.).

• Feature: A property shared by a population.
• Mode: Each of the variants representing a feature.

In psychological studies, characteristics such as personality or memory adopt modalities, which, to be statistically treated, must be expressed numerically.

Measurement and Measurement Scales

Measurement: The process of assigning numbers to objects or features according to determined rules. Ideally, measurement relates the real world of objects to the numerical world, preserving the exact relationships observed in the empirical world. Only those numerical relationships that can be empirically verified would be valid.

For this, we use Measurement Scales: a procedure that biuniquely relates the form of a property with its modalities to a set of numbers, for each mode. There are 4 types of scales (Stevens):

1. Nominal Scale

   If two observations belong to the same modality, they are assigned the same number, and if they belong to different modalities, different numbers. The assigned numbers must reflect the equality or difference that the modalities have for the characteristic.

   Example: sex (male, female), personality (introvert, extrovert, etc.). Eligible transformation: Any transformation that preserves the equality or difference empirically observed between modes.

2. Ordinal Scale

   Can indicate that objects for a certain characteristic are greater than or less than some other degree or mode. The assigned numbers reflect this magnitude.

   Example: Grades (0-Fail, 1-Pass, 2-Good, 3-Very Good, 4-Excellent). Eligible transformation: Any transformation that preserves the order of magnitude, increasing or decreasing, of objects possessing the characteristic.