qualitative + quantitative research methods

What is research?

systematic and objective inquiry process aimed at expanding knowledge and solving problems

qualitative research

- thoughts, behaviors, and meanings.

- exploratory

- subjective experiences

- studying text and context.

- Findings depend on understanding how people interpret their experiences.

- May lead to new hypotheses

- Flexible Design: Adapts to research needs, such as theoretical sampling

qualitative methods

Observation​

Interviews​

Focus groups​

Case studies

strengths + weaknesses of observation

- more reliable than interview

- Provides real-time data

- Captures behaviors in context

- Minimizes self-report bias

- time-consuming

- Observer bias

- Limited to observable behavior

strengths + weaknesses of interview

- In-depth insights

- Flexibility in questioning

- Establishes rapport for honest responses

- hear from participant in their own words

- Time-intensive

- Potential for interviewer bias

- Analysis can be complex

strengths + weaknesses of focus group

- diverse perspectives

- Encourages interaction and discussion

- Quick data collection

- Groupthink may occur

- Dominant voices can skew results

- Less control over the discussion

samples

Unlike quantitative research, qualitative research doesn't need a sample that represents the entire population. Instead, it focuses on gathering detailed insights from specific individuals

> Purposeful Sampling:

Selecting participants who have specific characteristics or experiences that are relevant to the research question, to gather rich and meaningful information.

> Convenience Sampling:

Choosing participants who are easy to reach or readily available.

Example: Using friends, family, or colleagues for the study.

Note: While convenient, it may not provide the most varied insights.

> Snowball Sampling:

Existing participants help recruit new participants.

Example: After an interview, a participant might refer someone else who fits the criteria.

Use: Useful for reaching hard-to-find populations.

> Maximum Variation Sampling:

Selecting a diverse group of participants to capture a wide range of perspectives.

Goal: To understand how different factors influence experiences or opinions.

types of interview

Fully structured interviews​

Unstructured interviews​

Semi-structured interviews

fully structured

- interviewer follows a predetermined set of questions exactly as they are written

- Consistent format for all participants

- Questions are closed-ended, often with specific response options

- Limited flexibility for follow-up questions

Unstructured

- open conversations, where interviewer has a general topic but no fixed questions

- can explore topics in depth

- Participants can steer the conversation, leading to unexpected insights

- May cover a wide range of topics beyond initial interests.

- exploring complex issues where depth of understanding is prioritized over uniformity

Semi-structured

- combine elements of both structured and unstructured interviews

- The interviewer has set of guiding questions but can adapt based on the participant's responses

- Allows follow-up questions and deeper exploration

- Balances consistency and flexibility

- obtaining detailed information while still allowing for comparability across interviews

analysing qualitative data

> Content Analysis

- Examines text or visual content to find patterns

- identifies themes or specific words/phrases

- Analyzes media, documents, or interviews

> Thematic Analysis

- looks for themes in qualitative data

- Codes data and groups codes into themes

- Understands broader patterns and meanings

> Framework Analysis

- Uses a structured approach to analyze data.

- Organizes data into key themes and concepts.

- Helpful in applied research with specific questions.

> Narrative Analysis

- Studies the stories people tell.

- Analyzes how narratives are structured.

- Understands personal experiences and communication.

> Conversation Analysis

- Examines how people interact in conversation.

- Looks at the structure of spoken interactions.

- Understands social communication patterns.

> Discourse Analysis

- Analyzes language use in context.

- Investigates meanings behind communication.

- Explores deeper social implications in texts.

> Interpretive Phenomenological Analysis (IPA)

- Focuses on how people understand their experiences.

- Explores the meanings of personal experiences.

- Understands individual perspectives.

challenges of qualitative data

- Subjectivity

- Complexity

- Time-Consuming

- Lack of Standardization:

- Organizing and storing data, like audio or notes, can be tricky

- Context Sensitivity

- Ethical Considerations

strategies to ensure trustworthiness in qualitative ressearch

triangulation

respondent validation

self- description

prolonged engagement

audit trail

thick description

peer debriefing

triangulation

Using multiple sources or methods to make research findings more reliable

Respondent Validation (Member Checking)

Sharing findings with participants to confirm they are accurate.

Ensures the conclusions match the participants' views.

Self-Description/Reflexivity

Researchers reflect on how their beliefs and experiences affect the study

Acknowledges personal biases that might influence the results.

prolonged engagement

pending an extended period of time in the field or with participants during the research process

- Builds Trust

- Deepens Understanding: Provides a better grasp of the context and dynamics.

- Recognizes Variability: Helps identify different perspectives and experiences.

- Refines Focus

audit trail

Keeping detailed records of the research process, including data collection and analysis

thick description

Providing rich, detailed descriptions of the context and participants

Helps readers understand the setting and nuances, making findings more relatable and credible

peer debriefing

discussing the research process and findings with colleagues or peers

Offers external perspectives that can identify biases and strengthen interpretations, enhancing credibility.

In qualitative research, which of the following refers to the use of multiple methods or data sources to develop a comprehensive understanding of phenomena?​

Member checking.​

Snowball sampling.​

Quota sampling.​

Convenience sampling​

Triangulation.​

triangulation

A qualitative study is carried out to investigate the attitudes of pharmacy students on the feedback they receive from their tutors on exam performance. The researchers openly acknowledge that the relationship among the researchers, the research topic and subjects may have influenced the study results. This concept is known as:​

Triangulation.​

Iteration.​

Grounding.​

Reflexivity.​

Transferability

reflexivity

A. Snowball sampling.​

B. Maximum variation sampling.​

C. Reliability.​

D. Participant observation.​

E. Negative sampling.​

F. Focus group.​

G. In-depth interview.​

H. Reflexivity.​

I. Transferability.​

​

For each of the following definitions/statements, choose the most likely option

1. Collecting data on naturally occurring behaviours of participants in their usual setting. ​

​

2. Using study participants as informants to identify other people who could potentially participate in the study. ​

​

3. The researcher reflects on whether his or her values and attributes may have influenced (or biased) any stages of the study

1. Collecting data on naturally occurring behaviours of participants in their usual setting. Participant observation.​

​

2. Using study participants as informants to identify other people who could potentially participate in the study. Snowball sampling.​

​

3The researcher reflects on whether his or her values and attributes may have influenced (or biased) any stages of the study. Reflexivity.​

quantitative methods

- experimental

- quasi- experimental

- true experimental (RCT's)

- cross sectional

- case control

- cohort

observational vs experimental studies

Observational study- we observe and measure specific characteristics but we don't attempt to modify the subjects being study​

​

In an experiment, we apply some treatments (or interventions) and then proceed to observe the effects on the subjects or experimental units​

experimental

definition:

Researchers manipulate independent variables to observe effects

characteristics:

- Random assignment to groups

- Establishes cause-and-effect relationships

Quasi-Experimental

definition:

Similar to experimental but lacks random assignment. Instead, subjects are assigned to groups based on non-random criteria

characteristics:

- Real-world settings

- Manipulation of variables; potential selection bias

True Experimental (RCTs)

definition:

Participants are randomly assigned to intervention or control groups

characteristics:

- High internal validity

- blinding may occur

- Gold standard for causal relationships

- Strong evidence for effectiveness and safety

- Measures disease risk prospectively

- Minimizes bias with blinding

- Controls confounders through randomization

Strict eligibility criteria can limit external validity.

Needs large samples for small effects or rare outcomes.

Generally expensive.

High costs for long follow-up studies.

Intervention effects may differ in real-world practice.

Cross-Sectional

definition:

Collects data at a single point (snapshot)

characteristics:

- Assesses prevalence or relationships

- Does not infer causality

advantages:

- short duration

- find out prevalence of multiple predictors and outcomes​

disadvantages:

- cant find incidence​

- Not suitable for rare predictors or outcomes

cross sectional surveys

collect data from a population or a representative subset at a single point in time

Case-Control

definition:

Compares individuals with a condition (cases) to those without (controls)

- what are the factors that caused this event?​

- 1st group has condition of interest. These are the cases (e.g. children with asthma)​

- 2nd group will be the control. These don't have condition of interest (e.g. children without asthma- matched control)​

characteristics:

- Retrospective design- look back to find risk factors that may explain why the cases got the disease and controls did not​

- Useful for rare diseases or outcomes

- Sequence of events may be unclear​

- Information recall bias​

- If outcome is fatal, information may have to be obtained indirectly via relatives etc.​

- Do not yield prevalence and incidence

cohort

definition:

Follows a group with common characteristics over time to observe outcomes based on exposure

what are the effects of this particular exposure?​

used to identify risk factors (exposures) that lead to a particular outcome (e.g. smoking and lung cancer)​

characteristics:

- prospective or retrospective

- Establishes associations between exposures and outcomes

- Studies multiple outcomes for the same risk factor

- Good for rare exposures

- Provides incidence and risk data

- More control in prospective cohorts

- Less bias

Requires large sample sizes.

Harder for rare outcomes.

Lengthy and costly follow-up.

Difficult to maintain follow-up

good to group/ visualise data initially

Purpose: Identify patterns and outliers early

Outliers

Choosing an Average

Mean: Sensitive to outliers.

Median: Better for skewed data.

Mode: Most common value.

Spread of Data

Skewness:Positive Skew: More low values.Negative Skew: More high values.

Measures of Spread

- Range: Difference between highest and lowest.

-Variance: How values spread from the mean.

- Standard Deviation: Average distance from the mean.

quantitive analysis: Distinguishing meaningful patterns (signal) from random variations (noise) in data

- Parametric Tests: Used when data fits a specific pattern (e.g., t-tests).

- Non-Parametric Tests: Used when data doesn't fit a specific pattern (e.g., Mann-Whitney U test)

- Significance Level: A cutoff (usually 0.05) to decide if results are significant.

- p-Value: Tells you the probability that your results happened by chance; lower values suggest stronger evidence

Correlation vs. Causation

- Just because two things are related doesn't mean one causes the other. Always look for other explanations.

see slide 48