RDA Week 2: Research Ideas & Hypotheses, Defining & Measuring Variables

Basic Research

Studies designed to address theoretical issues or expand general knowledge

Applied Research

Studies directed towards solving practical problems

Academic Literature

The body of information published worldwide

-Aim to identify studies directly related to research ideas

-Goal is to have a strong general understanding of the research area

-Keep an open mind…start broad and narrow

Primary Sources

Firsthand reports in which the authors describe their observations

-Empirical journals and theses

Secondary Sources

Reports in which the authors discuss someone else’s observations

-Textbooks, research report introductions, meta-analyses

Research Hypothesis

-A normal hypothesis

-HR or HA

Null Hypothesis

A statement that states there is not effect, no change, or no relationship among variables

-H0

A Good Hypothesis is…

-Logical

-Testable

-Refutable or Falsifiable

-Positive

Variables

Characteristics that change or differ between individuals

Independent Variable

Manipulated or controlled by experimenter

Dependent Variable

Measured by experimenter (hypothesized to change)

Operation Definitions

Allow us to discuss constructs (that cannot be observed or measured directly) by defining what they encompass and how we measure them

Validity

Does the procedure accurately capture the variable it is designed to measure?

Face Validity

Concurrent Validity

Reliability

Measurement Error

Due to limitations in measurement device

Observer Error

Due to human error in measurement

Environmental Changes

Make it difficult to recreate identical circumstances

Participant Changes

Participant changes between measurements can change measurements

Test-Retest Reliability

Compares scores of two successive measurements of the same individual

Inter-Rater Reliability

Compares the agreement between two observers or instruments simultaneously recording measurements

Validity = ?

Accuracy

Reliability = ?

Consistency of measurement

Nominal Scale

Represents qualitative differences in the measured variable

Ordinal Scale

Represents differences in terms of rank

Interval Scale

Measurement with an equal interval between values but without a true zero point (zero point is arbitrary)

Ratio Scale

Measurement with an equal interval between values and includes a true zero point (zero represents the complete absence)

Self-Report Measures

Directly asks participants questions in order to obtain information on behaviour

Pros and Cons of Self-Report Measures

Pro: Provides a direct measure from the participants perspective

Con: Reporting accuracy can result in low validity

Physiological Measures

Based on assumed manifestations of the underlying construct

Pros and Cons of Physiological Measures

Pro: Provides an object measurement

Cons: Require expensive equipment; often not feasible, may not measure constructs validly, assumes a physiological change will result in a behavioural difference

Behavioural Measures

Based on activities that can be observed and measured

Pros and Cons of Behavioural Measures

Pros: Can classify or analyze many aspects of a construct; often choose the best one

Cons: Behaviours can be highly situational, identifies change in behaviour, but not

physiological causes

Multiple Measure Approach

Using multiple methods to evaluate a variable or construct

Pros and Cons of Multiple Measure Approach

Pro: Increases confidence in validity

Cons: Complex statistical analyses, lack of agreement between measures, combining measures into a single score

Range Effect

A procedure is not sensitive enough to detect a measurement difference

Ceiling Effect

Scores clustered at high end of the scale

• Does not allow higher scores to be detected

Floor Effect

Scores clustered at low end of the scale

• Does not allow lower scores to be detected

An Artifact

A feature accidentally introduced into measurement experimental process

Experimenter Bias

Experimenter’s expectations influence results

Single-Blind Study

Research conducted by a naïve experimenter

Double-Blind Study

Both experimenter and participant are naïve

• Gold standard