Research Designs Lecture 4

Podcast

Play an AI-generated podcast conversation about this lesson

Questions and Answers

What is a key characteristic of probability sampling?

Results can often be generalized to the population. (correct)

Elements have unequal chances of selection.

It is not scientifically sound.

It requires more complex methodologies.

Which sampling method involves dividing the population into subgroups before sampling?

Cluster Sampling

Random Sampling

Convenience Sampling

Stratified Sampling (correct)

What is a primary disadvantage of non-probability sampling?

Results cannot typically be generalized to the population. (correct)

It is more time-consuming than probability sampling.

It allows for greater control over the sample selection.

It is less operationally convenient.

Cluster sampling relies on what primary method?

Dividing the population into clusters and then randomly selecting entire clusters. Signup and view all the answers

Which statement best defines non-probability sampling?

Selection is based on subjective judgment and convenience. Signup and view all the answers

What is a common feature of stratified sampling?

Population is divided into mutually exclusive groups. Signup and view all the answers

When might non-probability sampling be preferred?

When quick data collection is necessary and randomness can be compromised. Signup and view all the answers

Which of the following best describes probability sampling processes?

They rely on equal chances for each individual to be selected. Signup and view all the answers

Which technique is most likely to be affected by geographical considerations in sample design?

Cluster sampling Signup and view all the answers

What is a key consideration when determining sample size using the confidence interval approach?

Desired level of precision Signup and view all the answers

What is a key characteristic of two-stage cluster sampling?

Select clusters first and then random members within those clusters. Signup and view all the answers

Which of the following represents a subjective method for determining sample size?

Rule of thumb approach Signup and view all the answers

Which of the following is NOT a problem associated with probability sampling?

Leads to high representativeness. Signup and view all the answers

In what situation is accuracy in sampling less critical?

Exploratory research Signup and view all the answers

What does convenience sampling primarily rely on?

Subjective criteria based on the researcher's convenience. Signup and view all the answers

Which of the following is a disadvantage of non-probability sampling related to statistical outcomes?

Statistical reliability is difficult to ascertain Signup and view all the answers

In what situation is judgmental sampling particularly useful?

When specific characteristics need to be met. Signup and view all the answers

What is a defining feature of quota sampling?

Certain quotas for types of people are established before sampling. Signup and view all the answers

What effect does limited resources have on sample design?

Influences the chosen sampling technique Signup and view all the answers

What characteristic makes a research sample more difficult to achieve when using systematic sampling?

Limited availability of information Signup and view all the answers

What is the primary purpose of snowball sampling?

To access rare or hard-to-find respondents through referral. Signup and view all the answers

Which of the following best describes the sampling technique of cluster sampling?

Dividing the population into clusters and then selecting some of those clusters for study. Signup and view all the answers

How does the process of judgmental sampling differ from convenience sampling?

Judgmental sampling selects based on researcher judgment while convenience sampling selects based on availability. Signup and view all the answers

Study Notes

Lecture 4 - Research Designs, Topics and Questions

Research designs can be qualitative, quantitative, or mixed (combining both qualitative and quantitative approaches).
Using mixed methods is called triangulation.

Differences in Research Designs

Difference 1: Data Nature - Quantitative research uses hard data (e.g., numerical data), while qualitative research utilizes soft data (e.g., words, observations, contexts).
Difference 2: Language of Research - Quantitative research employs a positivist language focusing on hypotheses, variables, and measures of central tendency (mean, standard deviation). Qualitative research uses a language of cases and contexts, useful for explaining details.
Difference 3: Purpose of Study - Quantitative studies aim to verify or falsify relationships (e.g., a negative relationship between social class and COVID-19 deaths). Qualitative studies explore other reasons for these relationships.
Difference 4: Logical Path of Research - Quantitative research follows a linear, sequential path (methodology → data collection → analysis). Qualitative research might be cyclical, iterative, or involve back-and-forth processes.

Choosing a Research Method

The method depends on research questions, underlying philosophy of research, preferences, and skills.
The approach may be influenced by colleagues, organizational approach, supervisor beliefs, and personal experience.
There's no right or wrong answer.

Choosing a Research Method - Considerations

Unit of analysis: Country, company, or individual.
Theory vs. Local Knowledge: Universal or local knowledge; will results be generalizable?
Theory or Data First?: Read literature first, then form theory; or gather data and form theory.
Research Design: Cross-sectional (one point in time) or longitudinal (changes over time).
Verification/Falsification of theory?

Quantitative Approaches - Key Points

Explain phenomena using numerical data analysis.
Identifies differences but not necessarily "why".
Data is always numerical (or convertible) and analyzed statistically.

Quantitative Data: Sources and Analysis

Data sources include surveys (especially using Likert scales), observations, and secondary data (e.g., government data).
Analysis involves Hypothesis testing, correlations, means, and standard deviations.

Likert Scale

A rating scale used in surveys; ranges from "very dissatisfied" to "very satisfied".

What Quantitative Researchers Worry About

Sample size adequacy.
Correct statistical tests used.
Generalizability of results.
Reproducibility of results/methods.
Accuracy of measurements.

What's Wrong with Quantitative Research

Can't measure everything accurately.
May not fully capture human behavior.
Can be impersonal, missing depth of engagement.
Data may be static, a snapshot of a particular moment.
May reveal only one version of the truth

Qualitative Approaches

Research that doesn't use numerical data; instead uses words, pictures, photos, videos, audio recordings, field notes, observations.
Tends to start with a broad question instead of a specific hypothesis.
Often involves developing theory inductively .

Using Qualitative Data

Facilitate exploring how and why events happen.
Does not require large sample sizes.
Potential issues: inaccurate/false responses; researcher bias/influence; and researcher objectivity challenges.

Sources of Qualitative Data

Interviews (structured, semi-structured, unstructured).
Focus groups.
Questionnaires/Surveys.
Secondary data (e.g., diaries, written accounts, company reports).
Direct observations (often recorded).
Ethnography (studying customs, beliefs).
Qualitative data analysis methods:
- Content analysis (categorizing and summarizing data)
- Narrative analysis (re-formulating stories based on context.)
- Discourse analysis (examining talk and written text).

What Qualitative Researchers Worry About

Data coding accuracy.
Realistic situation capture.
Detailed context presentation.
"Seeing the world" through participants' eyes.

What's Wrong with Qualitative Research

Can be very subjective.
Challenging to repeat or replicate findings.
Not always generalizable to other contexts.
Defining and measuring concepts challenges.
Potential for easier implementation of "poor quality" Research.

Triangulation

Using multiple approaches (e.g., qualitative and quantitative methods) to study a research question, enhances confidence and credibility.
Increases credibility and internal validity of research.
Three forms of Triangulation include — data triangulation, researcher triangulation & methodological triangulation.

Effects of Triangulation

Combines multiple observers, theories, methods, and materials to overcome biases and weakness inherent in single-method studies.
Seeks confirmation through convergence (agreement or alignment) of perspectives from multiple sources
Increases research credibility and internal validity.

Types of Triangulation

Data Triangulation: gathering data from multiple sources (e.g., multiple interviews or different types of documentation).
Researcher Triangulation: Using multiple researchers to gather and analyze data.
Methodological Triangulation: Using multiple methods (e.g., both qualitative and quantitative).

Data Triangulation

Involves using different information sources (e.g., participants, other researchers, family, or other community members).
In-depth interviews gain stakeholders' perspectives and allow for diverse perspectives within a community.
Analysis involves comparing the different perspectives and looking for areas of agreement and divergence.

Researcher Triangulation

Involves multiple researchers examining the same group using the same method.
Comparing the findings of multiple researchers yields a broader, deeper understanding.
Agreement amongst multiple researchers increases the credibility of the findings.

Methodological Triangulation

Involves using multiple qualitative &/or quantitative methods to study a program.
Comparing results from multiple methods (e.g., surveys, focus groups and interviews) determines if similar results are observed.
Agreement between methods increases credibility.

Other Aspects of Research Design

Validity: Accuracy and factual soundness of the research.
Reliability: Consistency of research findings.
Trustworthiness: consistency & reproducibility of findings.
Dependability: Finding are consistent and replicable.
Confirmability: Extent findings are not influenced by researcher bias.

Summary

The research approach should be aligned with the research question and the researchers' abilities.
A mixed methods (triangulation) approach is often a suitable approach for thorough research.
Informed choices, justified and understood, are essential.
All approaches have limitations; recognizing and addressing these is crucial.

The Beginnings of a Research - Topic Sources

Personal experience and desire to investigate further.
Curiosity based on media reports.
State of knowledge in the field.
Solving a problem through research.
Social premiums (hot topics).
Personal values.
Everyday life and observations of behavior.

Arriving at Research Questions

Research questions are the starting point for research in the social sciences.
They guide and focus research efforts.
Crucial to base a research project on well-defined, clear and focussed questions.

Broad Topic

The broadest area of interest.
Guides literature review directions.
Identifies dominant questions in the topic.

Narrowing the Topic

Refines the broad topic based on specific interests.
Important factors to consider include: Relevant theories/models, time period, particular events, geographical area, biographical information, considerations of other aspects (economic, historical, etc.), available data.

Focused Topic

Similar to the narrowing process.
Familiarity with major debates.
Awareness of available information & resources

Research Question

Allows for understanding the complexity of the topic.
May contain several subcomponents, leading to a coherent argument in the research paper, Avoid including numerous related issues and questions.
The rationale behind the research questions should be able to be explained.

Research Questions (Example)

How did Indigenous NGOs utilize liberal economic policies in the Anglophone Caribbean between 1999 and 2020 to promote democracy?
- Specific/Sub questions
  - What liberal economic policies were implemented ?
  - To what extent were these policies used ?
  - How successful was the use of these policies?
  - How did local contexts impact the use of these policies?

Pay attention to and define the key concepts.

Key concepts need to be clearly understood and defined within the introduction
Examples include "NGOs", "democracy", "liberal economic policies", etc.

Conceptualisation, Operationalisation and Levels of Measurement

Conceptualization: Defining the key concepts within a research question.
Operationalization: Defining how to measure or gather data to capture the concepts.
Levels of Measurement: Nominal, Ordinal, Interval, Ratio.

Research Design Questions: Example (Youth & Political Participation in the Public Sector)

What is the level of political participation among young people employed in the public sector?
To determine the level of political participation among young people employed in the public sector?

Concepts - Essential Principles

Concepts are labels for phenomena, aiding communication.
Some concepts are concrete (observable and measurable); others are abstract and intangible.
Conceptualization involves specifying the meaning of concepts within the research.

Levels of Measurement

Nominal level measurement has no mathematical interpretation, it describes differences in categorical variables.
Ordinal level measurement indicates an order, but not the distance between values (differences between ranks).
Interval level measurement defines and ranks values with consistent intervals (e.g., IQ scores or temperature).
Ratio level measurement has a true zero point and indicates a fixed measure of the variable (e.g., weight or height).

Measurement Quality – Reliability

A measurement approach's reliability can suffer if respondents must interpret questions. Fixed-response formats (e.g., multiple choice questions) help to minimize interpretation and maximize reliability.

Measurement Quality – Validity

The extent to which a measurement represents the intended concept.
Examples: Measuring 'university education' with either “number of years studied” or “qualifications received” have differing validities.

Validity is divided into types

Content validity:* How completely a measure reflects all components of the concept
Face validity:* How well the measure logically matches common agreement about the concept

Sampling: Techniques, Concepts and Terminology

Sampling is about efficiently selecting a subset from a bigger population to learn things about the larger group.
Universe: All possible elements (e.g., all past and present students).
Population: Specified elements within the "universe" (e.g., students who attended University between two given time periods).
Sample Frame: A list of eligible elements that can be drawn upon to select a sample (e.g. list from university registrar/databases).
Sampling: The act of selecting individuals for the sample.
Sample Element: Individual observation selected from the sample.

Types of Sampling

Probability Sampling: Every member of the population has a known chance of being selected; results can be generalized.
- Simple random sampling: Randomly selects individuals.
- Systematic sampling: Selects every nth individual.
- Stratified sampling: Divides the population into subgroups, then randomly samples from these.
- Cluster sampling: Randomly selects groups (clusters) and may include all individuals within those groups or random individual selections from each group.
Non-Probability Sampling: Probability is not considered in the selection process.
- Convenience sampling: Subjectively selects convenient units (individuals readily accessible).
- Judgmental sampling: Uses researcher judgment to select informative cases.
- Quota sampling: Sets quotas for specific characteristics in a sample
- Snowball sampling: Participants in the study refer new participants.

Problems with Probability Sampling

Overall cost of study.
Time required for the study; call-backs and follow-ups.
Sample frames become outdated quickly.

Non-Probability Sampling

Non-probability sampling approaches are often simpler, quicker and cheaper than Probability methods Advantages: speed, lower costs, simplicity Disadvantages: Unrepresentative samples, potential for bias

Issues in Sample Design and Selection

Accuracy: Samples should represent the target population.
Resources: Time, money, and resources are an important consideration in all types of research designs

Determination of Sample Size

Subjective Methods: Rules of thumb, averages from similar studies.
Cost/Time Basis: Factors relating to the available resources of the study (funding, personnel, time).
Statistical Formulae: More sophisticated methods that use confidence intervals, etc.

Studying That Suits You

Use AI to generate personalized quizzes and flashcards to suit your learning preferences.

Description

Explore the distinctions between qualitative, quantitative, and mixed research designs in this quiz. Understand the implications of using different data types and the language of research methodologies. Test your knowledge on triangulation and its importance in research.