Clinical Psych Unit 3

1. Measurement in Research

  • Measurement: The process of assigning values (numbers or labels) to represent attributes of organisms, objects, or events so they can be compared. Accurate measurement is crucial in research to ensure that valid conclusions can be drawn from data analysis.

Constructs

  • Construct: A concept that cannot be directly observed (e.g., kindness, intelligence). Constructs serve as fundamental building blocks in psychological research; therefore, researchers must operationally define constructs for clarity and precision.

Operational Definition

  • Operational Definition: The specific way a variable is measured in a study. - Example: For kindness, it can be operationalized as the number of helping behaviors observed, showcasing a quantifiable framework for analyzing abstract concepts.

2. Measurement Accuracy

Classical Test Theory

  • Fundamental Statement: Observed Score = True Score + Measurement Error.

Measurement Error

  • Definition: Factors that inflate or deflate a person’s true score. Understanding measurement error is essential for researchers to improve data collection methodologies.

  • Types: 1. Random Error

    • Characteristics: Unpredictable variations that can affect the outcomes.

    • Example: Fatigue or distraction affecting test performance, which underscores the importance of creating a conducive testing environment.

    1. Systematic Error

    • Characteristics: Consistent bias in measurement that can lead to inaccurate conclusions.

    • Example: Faulty equipment leading to consistently inaccurate results; this highlights the necessity for regular maintenance and calibration of measurement tools.

Methods to Reduce Error

  • Implementation of Standardized Procedures to ensure consistency and reliability across different test administrations.

  • Taking Multiple Measurements allows researchers to obtain an average score, thus reducing the impact of random errors.

  • Utilizing Reliable Instruments that have been honed through previous research helps in acquiring accurate data.

3. Types of Variables

Variables

  • Definition: Factors that can change or vary, providing the foundation upon which data collection and analysis are based.

Qualitative Variables

  • Definition: Categorical variables that represent distinct groups and cannot be measured numerically.

  • Examples: - Gender (male, female)

    • Eye color (brown, blue, green)

    • Political affiliation (Democrat, Republican, Independent)

Quantitative Variables

  • Definition: Numerical variables that can be measured, providing essential data for statistical analysis.

  • Examples: - Height (measured in centimeters or inches)

    • Weight (measured in kilograms or pounds)

    • Test scores (numerical grades or percentages)

Quantitative Variable Types

  1. Discrete Variables

    • Characteristics: Countable values; no fractions or decimals involved.

    • Examples: - Number of siblings (0, 1, 2,…)

      • Number of pets (cat, dog, etc.)

  2. Continuous Variables

    • Characteristics: Can take any value within a range, making them suitable for more complex analyses.

    • Examples: - Height (from 150 cm to 200 cm)

      • Temperature (measured in degrees Celsius or Fahrenheit)

      • Time (measured in seconds, minutes)

4. Scales of Measurement

  1. Nominal Scale

    • Characteristics: Categories only; no order or ranking among them.

    • Examples: - Gender (male, female)

      • Eye color (blue, brown, green)

      • Major (psychology, biology, history)

  2. Ordinal Scale

    • Characteristics: Ranked order; differences between ranks are not equal, indicating a hierarchy without a specific distance between ranks.

    • Examples: - Race finishing positions (1st, 2nd, 3rd)

      • Class rankings (top 10, middle tier, bottom)

  3. Interval Scale

    • Characteristics: Equal intervals between values but lacks a true zero; used for data that can be measured at uniform intervals.

    • Examples: - Temperature (Celsius or Fahrenheit)

      • IQ scores (on a standardized test)

  4. Ratio Scale

    • Characteristics: Equal intervals and a true zero exists, allowing for a comprehensive range of mathematical operations.

    • Examples: - Height (0 cm indicates no height)

      • Weight (0 kg indicates no weight)

      • Age (years lived)

      • Time (duration measured in seconds, minutes)

5. Types of Measurement

  1. Self-Report Measurement

    • Definition: Participants report their own thoughts, feelings, or behaviors, providing subjective data.

    • Examples: Surveys, interviews, questionnaires, where participants articulate their perspectives.

    • Strengths: Economical, easy to administer, allows for quick data collection with minimal equipment.

    • Limitations: Prone to dishonesty, susceptible to social desirability bias affecting honesty (respondents may answer to appear favorable), inaccurate memory, and retrospective bias leading to flawed recall.

  2. Behavioral Measurement

    • Definition: Measures observable behavior, providing an objective perspective on participant actions.

    • Examples: Counting the number of times someone checks their phone during a session, observing stress behaviors in different contexts.

    • Strengths: More objective than self-reports, reducing the habitual biases found in participant accounts.

    • Limitations: Time-consuming to conduct, may require extensive training, exposure effects (participants may alter their behavior when under observation), and risk of observer bias impacting results.

Reducing Observer Bias

  • Use a masked design with blind observers to prevent bias in data interpretation.

  1. Physiological Measurement

    • Definition: Measures biological processes, giving insight into physiological reactions.

    • Examples: EEG (brain activity analysis), fMRI (brain imaging), heart rate monitoring, cortisol levels (stress indicators), skin conductance (measuring arousal).

    • Strengths: Provides highly objective data with high accuracy levels and diverse applications in psychology.

    • Limitations: Often expensive, necessitates specialized training for personnel, and may involve invasive procedures.

6. Reliability vs Validity

Reliability

  • Definition: The consistency of a measurement over time, reflecting the stability of the tool used. A measure is reliable if it produces similar results repeatedly under consistent conditions.

Validity

  • Definition: Refers to whether a test measures what it claims to measure, making it crucial to establish the credibility of research findings.

  • Important rule: Reliability does NOT guarantee validity; a tool can yield consistent results but still measure the wrong construct.

  • Example: A scale consistently measuring weight in kilograms when you believe it is measuring in pounds may show reliability but lack validity.

Reliability Types

  1. Internal Consistency

    • Definition: Measures whether items on a test assess the same construct cohesively, ensuring data collected aligns with what is intended to be measured.

    • Example: A 40-question shyness test where responses correlate positively if measuring the same underlying construct.

    • Split-Half Reliability: Divide the test into two halves; scores from both should be similar to confirm consistency.

    • Cronbach’s Alpha (α): A statistic measuring internal consistency, ranging from 0 to 1; higher values (typically above 0.7) indicate more reliable measures.

  2. Test-Retest Reliability

    • Definition: Measures consistency over time by evaluating the same participants using the same test at different points.

    • Example: A person taking the same personality test twice over a period should yield similar results if measuring stable traits, not transient states.

    • Important: More applicable for traits (stable attributes like extraversion) rather than states (temporary conditions like current mood).

Types of Validity

  1. Face Validity

    • Question: Does the test appear to measure the construct to the casual observer?

    • Example: A depression questionnaire directly asking about feelings of sadness.

    • Note: Although useful, it is not regarded as the strongest form of validity due to potential superficial assessments.

  2. Internal Validity

    • Question: Does the study demonstrate a real relationship between variables and accurately reflect cause-and-effect relationships?

    • Example: Drug A successfully reducing symptoms demonstrates high internal validity.

    • Low internal validity example: Height predicting intelligence is flawed, as it lacks logical reasoning.

    • Example Correlation: Observing ice cream sales rise concurrently with an increase in violent crime indicates a correlation but does not imply causation.

  3. External Validity

    • Question: Do results generalize to other people or situations?

    • Example: A study finding a specific dog treat preferred by most dogs would suggest applicability to dog owners at large.

    • Low external validity example: A study limited to white New Yorkers generalized to all Americans could lead to erroneous conclusions.

  4. Content Validity

    • Definition: Assesses whether the measure covers the full range of the construct being evaluated.

    • Example: A test covering chapters 1–3 that only poses questions about chapter 1 would demonstrate poor content validity.

  5. Construct Validity

    • Definition: Measures whether a test truly assesses the theoretical construct it is intended to measure.

    • Two Types:

      • Convergent Validity: Measures of the same construct correlate strongly.

        • Example: Two different depression scales resulting in similar scores indicate strong convergent validity.

      • Discriminant Validity: The measure does NOT correlate with different constructs to ensure a precise assessment.

        • Example: An empathy scale should not correlate with extroversion level indicators, demonstrating effective discrimination between constructs.

  6. Criterion Validity

    • Definition: The ability of a test to predict outcomes accurately based on specific criteria.

    • Predictive Validity: Predicts future outcomes based on current assessments.

      • Example: SAT scores forecasting college GPA success based on standardized testing results.

    • Concurrent Validity: Predicts current outcomes measured at the same time as other variables.

      • Example: An aggression survey compared with observed aggression behaviors directly spanning a similar timeframe.

7. Recruitment of Participants

Types of Sampling

  • Probability Sampling: Every individual within a population has a chance of being selected, enhancing the randomness of obtained samples and improving external validity.

Types

  1. Simple Random Sampling

    • Definition: Random selection from the entire population, ensuring each member has equal representation.

    • Example: Drawing student names from a hat to make the selection process unbiased.

  2. Stratified Random Sampling

    • Definition: The population is divided into subgroups based on specific characteristics, which are then randomly sampled to ensure representation across relevant categories.

    • Example: Sampling by ethnicity to match population proportions for comprehensive analysis.

  3. Cluster Sampling

    • Definition: The population is divided into clusters (often geographical), and entire clusters are randomly selected to reduce sampling complexity.

    • Example: Randomly selecting universities in Michigan where all students from chosen universities participate in the study.

  4. Multistage Sampling

    • Definition: Random sampling conducted in multiple stages to simplify data collection logistics.

    • Example:

      1. Select universities

      2. Select students within them to participate in a study ensuring robustness.

Law of Large Numbers

  • Point: Larger sample sizes tend to more accurately represent the population and its characteristics.

  • Important Notes:- Larger samples yield greater representativeness and reduced margin for error.

    • Benefits of larger samples decrease after a certain size is achieved, indicating optimal sampling strategies.

    • Typical research target: 25–30 participants per group, balancing practicality with statistical power.

Non-Probability Sampling

  • Definition: Not every individual has a chance of being selected, which may enhance ease of access but limits representativeness.

Types

  1. Convenience Sampling

    • Definition: Participants are selected based on their easy accessibility, often leading to a biased sample.

    • Example: Recruiting college students in psychology classes for studies, which may not reflect broader perspectives.

  2. Snowball Sampling

    • Definition: Existing participants recruit other participants, forming a network of study participants.

    • Useful for: Accessing rare populations and hard-to-reach groups, thereby enhancing diversity in study samples.

Recruitment Methods

  • Examples: - Subject pools from universities

    • Flyers posted in accessible locations

    • Online advertisement efforts

    • Social media outreach for diverse audiences

    • Craigslist listings targeting specific demographics

    • Collaborations with clinics and schools for direct access

    • Platforms like Amazon Mechanical Turk for vast participant pools with quick turnaround.

  • Recruitment methods should align carefully with the chose sampling strategy to ensure efficiency and effectiveness in data collection, reinforcing the integrity of research findings.