Taller de tesis - 7

Understanding Research Design in Crisis Methodology

Introduction to Research Methodology

• The session begins with an introduction, confirming audio clarity and setting the stage for discussing Chapter 3 of a crisis methodology.

• The focus is on various elements essential for graduate and postgraduate theses, including research approach, design, methods, population, techniques, and instruments.

Types of Research Design

• Research design is categorized into experimental and non-experimental types; this distinction applies only to quantitative research.

• Experimental research can be further divided into pure experimental, quasi-experimental, and pre-experimental designs.

Experimental Research

• Pre-experimental: Minimal manipulation by the researcher; groups remain intact without comparison.

• Quasi-experimental: Involves intact groups where comparisons are made based on existing conditions (e.g., different classes).

• Pure experimental: The researcher intentionally manipulates variables under controlled conditions to measure effects accurately.

Key Characteristics of Experimental Designs

• Pure experiments require careful control over manipulations to ensure validity; multiple groups are necessary for comparison.

• Example provided includes vaccine testing on rats with varying characteristics to observe effects across two distinct groups.

Non-Experimental Research

• Non-experimental research can be classified as transactional (cross-sectional), longitudinal or backward-looking studies.

Transactional vs. Longitudinal Studies

• Transactional (Cross-sectional): Data collection occurs at a single point in time; useful for capturing current profiles or behaviors.

• Longitudinal: Measures changes over time through repeated observations; allows analysis of developmental trends (e.g., children's behavior related to television viewing).

Conclusion on Research Types

Research Design and Methodology Overview

Understanding Non-Experimental Research

• The speaker emphasizes the need to clarify whether research is non-experimental, particularly in transactional contexts. They highlight the importance of justifying the chosen design by addressing "what" and "why."

Research Methods Explained

• The discussion transitions to defining research methods, specifically focusing on theoretical methods that guide how the theoretical part of research is developed.

• Key types of methods mentioned include historical and logical methods, with a focus on analysis, synthesis, induction, deduction, and hypothetical-deductive approaches.

Application of Research Methods

• The speaker explains that researchers should articulate how they have investigated their theoretical framework using various methods like analysis and synthesis.

• A distinction is made between inductive (from specific to general) and deductive (from general to specific) methodologies in qualitative versus quantitative research.

Clarifying Population and Sampling

• The concept of population is introduced as the group from which data will be collected. Populations can consist of subjects or objects relevant to the study.

• Examples are provided regarding populations for studies—such as first-year law students or five-star hotels—emphasizing that populations must share common characteristics without mixing different groups.

Defining Sample Size and Sampling Process

• The speaker discusses finite versus infinite populations, explaining that sampling involves selecting a representative subgroup from a larger population for practical data collection.

Understanding Sampling Methods in Research

Types of Sampling: Probabilistic vs. Non-Probabilistic

• The sampling method can be categorized into probabilistic and non-probabilistic approaches, where probabilistic sampling uses mathematical formulas to determine sample size, while non-probabilistic sampling relies on other criteria without a formula.

• An example illustrates that with a population of 900 students, probabilistic methods are applicable; however, for a smaller population like 5 teachers, non-probabilistic methods are more suitable.

Application of Sampling Methods

• Probabilistic sampling is primarily used for quantitative research, whereas non-probabilistic sampling is typically employed in qualitative research contexts.

Calculating Sample Size

• A specific formula is introduced for calculating sample size when the population is infinite or unknown. The variables include 'n' (sample size), 'z' (confidence level), 'p' (probability of success), 'q' (probability of failure), and 'd' (precision).

• The confidence level can vary between 90% and 95%, influencing the value obtained from the z-table. A common assumption for p and q is a 50/50 probability split.

Understanding Precision and Error Margins

• Precision refers to the maximum allowable error in research findings, generally set between 5% to 10%. An example from presidential elections highlights how significant errors can occur if precision isn't maintained.

Formulas Based on Known Population

• When the population size is known, another formula applies where values are substituted directly into the equation based on known parameters such as total population ('n') and confidence levels ('z').

• For instance, if there are 900 individuals in a study group, calculations involve substituting this number along with predetermined values for z and precision to derive an accurate sample size.

Tools for Sample Size Calculation

• If calculations seem complex, online tools like SurveyMonkey provide direct results for sample sizes based on user inputs regarding population size and desired confidence levels.

• These tools simplify the process by allowing users to input their parameters easily without needing extensive statistical knowledge.

Adjusting Confidence Levels and Margins of Error

• Users can adjust their margin of error or confidence level within these tools to see how it affects their required sample size—demonstrating flexibility in research design decisions.

Understanding Stratified Random Sampling in Research

Introduction to Random Sampling

• The discussion begins with the concept of random sampling, emphasizing that it is not merely about using a calculator but involves applying statistical formulas.

• A sample size of 273 is mentioned from a total population of 900, highlighting the importance of calculating random numbers for effective sampling.

Stratification in Sampling

• The speaker explains stratification, which involves dividing the population into strata (e.g., age groups or gender) to ensure representative samples.

• Examples are provided on how to stratify based on marital status and living conditions (urban vs. rural), illustrating the need for diverse categories in market studies.

Calculating Sample Sizes

• A hypothetical population of 600 is introduced, with a calculated sample size of 130 surveys needed for accurate representation.

• The breakdown between genders shows that out of 130 surveys, adjustments must be made based on the actual composition (450 women and 150 men).

Adjusting Survey Numbers by Stratum

• The calculation process for determining how many surveys should be conducted within each stratum (men and women) is discussed.

• An example calculation reveals that approximately 33 surveys should target men while adjusting the remaining number for women to maintain balance.

Finalizing Sample Distribution

• The final distribution indicates that if there are 33 men surveyed, then 97 women will also need to be surveyed to meet the total sample size requirement.

• Emphasis is placed on ensuring that these calculations reflect true proportions within the population for valid results.

Importance of Probabilistic Sampling

• The speaker contrasts probabilistic sampling with convenience sampling, stressing its significance in research validity.

• Questions regarding clarity indicate an interactive session where participants seek further understanding about probabilistic methods.

Statistical Formulas and Confidence Levels

• Discussion shifts towards statistical formulas used when population sizes are unknown, introducing concepts like z-values derived from z-tables.

• Confidence levels such as 90% or higher are explained alongside their implications for data accuracy and reliability in research findings.

Probability Calculations

• Finally, probabilities related to success and failure rates are addressed; percentages must be converted into decimal form for calculations.

Understanding Non-Probability Sampling in Qualitative Research

Key Concepts of Probability and Non-Probability Sampling

• The discussion begins with the explanation of a formula used for probability sampling, emphasizing that the only variable is the population size (N), which is set at 900.

• In contrast to probability sampling, non-probability sampling is highlighted as being utilized primarily for qualitative research without relying on specific formulas.

Characteristics of Non-Probability Sampling

• Non-probability sampling does not yield representative results; it cannot generalize findings due to its reliance on subjective criteria based on research purposes.

• The selection process involves identifying types or cases relevant to the study, focusing on units of analysis such as volunteers or experts rather than random samples.

Types of Units and Selection Criteria

• Examples are provided where researchers must select knowledgeable individuals (experts) related to specific cultural aspects, like local customs or practices, ensuring they have relevant insights.

• The importance of selecting appropriate participants based on their expertise is stressed; for instance, researching dance requires interviewing dancers rather than uninformed individuals.

Determining Sample Size and Saturation

• In qualitative research, sample sizes are determined by saturation—when additional interviews yield repetitive information about categories under investigation.

• The concept of saturation is further explained through examples involving small populations where repeated responses indicate a complete understanding of the phenomenon being studied.

Practical Considerations in Data Collection

• Researchers face practical constraints such as time and resources when conducting interviews; these factors influence how many subjects can be surveyed within a given timeframe.