Research methods

• The notes cover quantitative and qualitative research methods, experimental designs, variables, data analysis, hypothesis testing, and different research paradigms.

Systemic Review Process

• Starts with a well-formulated question.
• Requires a search strategy across databases.
• Uses an unbiased selection process for studies.
• Employs a critical appraisal of data.
• Involves synthesis and interpretation of results.
• Aims to reduce bias, ensure replicability, resolve conflicting data, identify gaps, and inform decision-making.

Answerable Questions and Systemic Reviews

• Characteristics of an answerable question are: Description of the population (P), identified intervention (I), explicit comparison (C), and relevant outcomes (O).
• Stages include confirming a clear research question using a PICO chart, defining inclusion and exclusion criteria, searching with truncations and boolean operators, screening papers, and conducting a quality assessment using a PEDRO scale.

Quantitative Research and Study Design

• Involves reviewing literature to find gaps, formulating a research question with aims and hypotheses, selecting an appropriate study design.
• Also understanding statistical tests, collecting and analyzing data, and interpreting the study.

Experimental Study Designs

• Include pre-experimental, true experimental, and quasi-experimental designs.

Within vs. Between Subjects

• Within-group/subject designs involve all participants taking part in every condition and are less affected by individual differences.
• Between-group/subject designs involve participants taking part in only one condition, and group differences are then compared.

Variables

• Dependent variables are measurable and can be referred to as criterion variables.
• What makes the dependent variable change the way it does? (if it does).
• Example: Investigating the effect of age on the amount of television watched; the DV is the amount of television watched.
• Confounding variables can influence the relationship between two other variables; daily calcium intake affecting the relationship between exercise and bone density.
• Control confounding variables by identifying all potential confounders from the outset and add controls to the study design.

Categorical Data

• Nominal data includes gender, team position, and job title.
• Ordinal data includes age categories, finishing positions in a race.

Scale Data

• Interval data includes temperature and time.
• Ratio data includes height, weight, and distance.

Descriptive Statistics: Data Distribution

• Relative frequency describes how often values of a variable occur.
• Histograms display the frequency of value ranges within a variable.

Central Tendency

• Mean is the average of a data set.
• Median is the middle number, best used with skewed data.
• Mode is the most frequently occurring number in a data set.

Variation

• Variation considers the spread of numbers in a data set.
• Range measures the difference between the highest and lowest values but is crude and vulnerable to outliers.
• Interquartile range indicates the difference between the highest and lowest values for the middle half of cases.
• Standard deviation indicates the typical distance of all values from the mean, which is affected by outliers.
• Smaller variation and range indicate more consistent data. Skewness describes the symmetry or asymmetry in data dispersion.
• Normal distributions have cases evenly dispersed from either side of the center (mean), use mean and SD to describe data, employ parametric tests.
• Skewed data are non-normally distributed and use median and IQR for description. use non-parametric tests.

Data Definitions

• Case = a bound unit of analysis.
• Variable = mass, a quality or characteristic of a case.
• Value = quantity of a case within a variable.

Testing Differences

• Sensitivity = detect a real difference between two or more groups on a variable of interest.
• Reliability = consistency of measurements.
• Validity = results reflect what is intended to be measured.
• (internal) - impacted by study design, methods of quality control
• (external) - extent to which findings can be generalised to larger groups.

Hypothesis Testing Framework

• Always involves two hypotheses:
• Alternative hypothesis and null hypothesis.
• Steps: define the research question, scrutinize the data, compute the statistical test, derive the P-value, and interpret the results.
• P-value refers to the probability value; p<0.05 means less than 5% of the time, the data would accept the null hypothesis.

Paired Data

• Data from the same group of people in different conditions or repeated measures within-subject.

Sphericity

• The variances of the differences between all combinations of related groups are equal.
• Within-subject differences occur when variance is usually small due to the same people being tested.

Mauchly's Test of Sphericity

• If P > 0.05, assume sphericity (variance is similar between groups).
• If P < 0.05, sphericity is violated, and the variances are significantly different between groups.
• If P < 0.05 use the Greenhouse Geisser correction.

Unpaired Data

• Data from different/independent groups.

Normality Testing Shorthand

• If p<0.05, the data is non-parametric.
• If p>0.05, the data is parametric.

Testing Associations

• Correlation describes the size and direction of a relationship between two or more variables, without proof of causation.
• Correlation unlike ANOVA provides an estimate i.e. how large or small the effect is or the direction of the effect
• Correlation between variables does not represent causation.
• Correlations analysis involves two outcome variables, a scatter plot, R-value (correlation coefficient), and P-value.
• Data are normally distributed use Pearsons correlation.
• Data are not normally distributed use spearman's correlation.

Correlation Coefficient

• Represents how close variables are; also known as the R-value.
• The closer R is to 1, the stronger the correlation.
• ‘R^2’ coefficient of determination, explained variation as a percentage.

Association Interpretation

• Correlations are influenced by outliers and linearity of data (line of best fit).
• Avoid saying ‘effects' and 'impacts
• Variability data is important in correlation.

Qualitative Research

Research Paradigms

• Paradigm: Research worldview/framework.
• Ontology: What is real?
• Epistemology: How do we know?

Research Paradigm Concepts

• Paradigm is the "lens" through which we understand things.
• Ontology is whether the world exists independently of our thoughts or is shaped by our perceptions.
• Epistemology is determined through experience (empiricism) or through reason and logic (rationalism)?
• Post-positivism: aims to get as close to "truth" employing scientific methods.
• Constructivism: centers around how individuals create their own meanings.
• Constructionism: focuses on how society shapes shared understandings.
• Post-structuralism/Critical Theory: challenges power structures and dominant ideas.

Paradigm in Research

• It determines what is studied, how it is studied, and how results are interpreted.
• A positivist paradigm follows scientific, objective methods.
• A constructivist paradigm focuses on understanding personal experiences.
• A critical paradigm challenges power structures and promotes social change.

Ontology in Research

• Researchers consider what they believe about what exists; this affects how they frame their study.
• Realism (Objective reality exists) is used in post-positivist, measure climate change with data.
• Relativism (Multiple realities exist) is used in constructivist research, study different cultures.
• Epistemology dictates how we justify what we know.
• Observation, experiments, and measurement help positivist epistemology.
• Interaction and interpretation help constructivist epistemology.
• Power, culture, and social context help critical epistemology.

Ethnography

• Study culture by immersing yourself in their environment.
• Researchers observe and interact with people in their natural setting.
• Used in anthropology and sociology.

Grounded Theory

• Develop theory based on data.
• Researchers collect data first and then build a theory from patterns they find.
• Often used in social sciences to explain behaviors.

Phenomenology

• Studies how people experience something.
• Focuses on personal experiences and feelings about a phenomenon.
• Used in psychology and healthcare research.

Community-Based Participatory Research (CBPR)

• Research done with a community, not just about them.
• The community is involved in designing, conducting, and applying research, solving local problems.

Case Studies

• Deep investigation of one person, group, or event over time.
• Gives detailed insights but can't always be allpied to everyone.
• Used in psychology, business, and education.

Narrative Research

• Studies people's stories and experiences.
• Focuses on individual life stories and how people make sense of events.
• Used in psychology, education, and social work.