Thinking with Data I - Lecture #2

A set of vocabulary flashcards covering key terms and concepts from the lecture on research methods and design.

Sources of Research Ideas

Common places to find research ideas include informal observations, practical problems, and research literature.

• Sources of research ideas

Research literature - all published research in a given field include journal articles; scholarly books

Review the literature

• Explore new research

• Find if the topic of interest has been Studied

• Provide ideas for how to conduct your own study

• Evaluate how your study contributes to the field

Research question

Empirically testable research questions

• Questions expressed in terms of a single variables(behaviours) or relationship among variable (mutiple behaviours)

  • Descriptive- how frequently or intense is the behaviour

  • Explanatory- what are some possible causes of the behaviour

  • Predictive- what are some possible effects of the behaviour

What if your question has already been studied empirically?

You may need refine it.

Methodology- are there others ways to define and measure the behaviour.

Sample- are there groups of people for whom the association between two behaviors might be stronger or weaker.

Context- are there situations in which the association between two behaviors intensifies or diminishes.

Interestingness

The answer is in doubt

Fill a gap in the literature

Practical implications

Feasibility

Time and Money

Equipment and Materials

Technical knowledge and skill

Acess to research participants

Research Hypotheses

A specific prediction about a particular phenomenon in a specific context of interest.

Not all research questions are accompanied by hypotheses

Often but no always derived from theories

• Theory: A coherent explanation or interpretation of one or more phenomena

Evaluate Research Hypotheses

• Testable and Falsifiable- Emipirical test & Evidence to disconfirm or support.

• Logical- informed by theories, observations, and/or logical reasoning

• Positive- Statement about the existence of a relationship or an effect.

Research design

Overall plan for how to conduct an emipircal study to answer research questions or test research hypotheses

• Sampling process, data collection procedure, measurement

Types of Research Design

• Experimental & non-experimental

• Cross-sectional & longitudinal

• Quantitative & Qualitative

When to use experiments

Causal or explanatory research questions

• Whether there is a causal relationship between two thing (variables)

• Whether changes in one variable (Independent variable;IV) lead to changes in another variables (dependent variable; DV)

Fundamental features of experiments

• Manipulation of the independent variable

• Control of the extraneous variables

Manipulation of Independent Variables

Researchers manipulate IV by systematically changing its levels • Conditions: Different levels of IV

• Participants exposed to different levels or conditions

• Factors and levels (e.g., single-factor two-level design)

• Must involve active intervention of researchers

• Crucial for eliminating potential alternative explanations

Control of Extraneous Variables

Extraneous variables

• Variables other than IV and DV that vary in a study

• E.g., situational/task variables, participant characteristics

• Make it difficult to detect effects of IV

• As noise that add (random) variability to data

• As confounding variables that vary systematically with IV and provide alternative explanations for observed difference in DV

Types of Experiments

• Between-subject experiment

• Each participant is assessed in only one condition

• E.g., independent groups experiment

• Matched-pairs experiment

• Participants are matched on certain characteristics then assessed

• E.g., matched-groups experiment

• Within-subject experiment

• Each participant is assessed under all conditions

• E.g., repeated measures experiment

Non-Experimental Research

Lacks the manipulation of independent variable

• Measure variables as they naturally occur

• Difficult to make causal conclusions

When to use non-experiements? If research questions are

• Broad and exploratory

• About a single variable

• About non-causal relationship between variables

• About causal relationship, but Independent variable cannot be manipulated

• Participants cannot be randomly assigned

Types of Non-Experiments

Correlational studies

• Variables of interest are measured, and their statistical relationship is assessed

• E.g., simple survey studies

Observational studies

• Behaviors of interest are observed in natural or laboratory setting

• E.g., naturalistic observations, participant observations, structured observations, case studies, archival studies

Research Validity

Validity

• True and accurate representation of Information

Types of Validity

• Internal validity

• External validity

• Construct validity

• Statistical validity

Extraneous Variables

Variables other than the independent variable that may influence the outcome of the study but are not being measured.

Random Assignment

The process of randomly allocating participants to different experimental conditions to eliminate bias.

Causal Relationship

A relationship where one variable directly influences another variable.

Validity

The extent to which a study accurately represents the phenomenon it is intended to measure.

• Internal Validityj

The ability to draw causal conclusions

• Nell-conducted experiments are usually high in

Internal validity

External Validity

The extent to which research findings generalize to settings or populations outside the study conditions.

Construct Validity

The degree to which a test measures what it claims to measure.

Statistical Validity

The appropriateness of the statistical analyses used in the study to answer the research question.

Hypothesis

A specific prediction about the outcome of a study based on theory or previous research.