3.1 Types of Experimental Designs

The choice of experimental design is critical and depends on the research question, available resources, and ethical considerations.

• Independent Groups Design (Between-Subjects): Different groups of participants are used for each experimental condition. Each participant experiences only one level of the IV.
  • Advantages: No order effects (practice or fatigue).
  • Disadvantages: Requires more participants; potential for participant variables (differences between groups) to confound results.
  • Dealing with problems: Random allocation of participants to conditions to minimize participant variables.
• Repeated Measures Design (Within-Subjects): The same participants are used in all experimental conditions.
  • Advantages: Fewer participants needed; eliminates participant variables.
  • Disadvantages: Order effects (e.g., practice effects, fatigue effects, boredom effects) can confound results.
  • Dealing with problems: Counterbalancing (e.g., ABBA design, where half participants do A then B, other half do B then A) to distribute order effects evenly across conditions.
• Matched Pairs Design: Different participants in each condition, but they are matched on characteristics relevant to the study (e.g., age, IQ, pre-test scores). One member of each pair is assigned to each condition.
  • Advantages: Reduces participant variables; no order effects.
  • Disadvantages: Difficult and time-consuming to find suitable matches; if one pair drops out, two participants' data are lost.

3.2 Experimental Settings

• Lab Experiment: Conducted in a highly controlled environment, allowing for precise manipulation of the IV and control over extraneous variables.
  • Advantages: High internal validity.
  • Disadvantages: Low ecological validity; artificial setting may lead to demand characteristics.
• Field Experiment: Conducted in a natural, real-world setting, where the IV is still manipulated by the researcher.
  • Advantages: Higher ecological validity.
  • Disadvantages: Less control over extraneous variables; ethical concerns (e.g., lack of informed consent if participants don't know they're in a study).
• Natural Experiment: The IV is naturally occurring and not manipulated by the researcher (e.g., comparing psychological effects before and after a natural disaster or policy change).
  • Advantages: High ecological validity; allows study of phenomena that would be unethical or impractical to manipulate.
  • Disadvantages: No control over IV; difficult to establish causality due to potential confounding variables.
• Quasi-Experiment: The IV is a pre-existing characteristic of the participants (e.g., gender, age group, clinical diagnosis) that cannot be randomly assigned.
  • Advantages: Can investigate real-world variables; higher ecological validity than lab experiments.
  • Disadvantages: Lack of random assignment means causal inferences are weaker; potential for participant variables to confound results.

3.3 Problems in Experimental Research and Solutions

• Manipulation of the IV: Ensuring the IV is effectively and consistently manipulated across conditions. Needs careful operationalization.
• Demand Characteristics: Cues in the experimental setting that inform participants about the hypothesis, leading them to alter their behavior to confirm or disconfirm expectations.
  • Dealing with problems: Single-blind design (participants unaware of condition), deception (minimally and ethically), use of filler tasks.
• Investigator Effects: Unintended cues or biases from the researcher that influence participant behavior.
  • Dealing with problems: Double-blind design (both participants and researchers interacting with them are unaware of conditions), standardized procedures, automated instructions.
• Participant Variables: Differences between individuals that can affect the DV.
  • Dealing with problems: Random allocation (independent groups), repeated measures design, matched pairs design.
• Situation Variables: Features of the environment or experimental context that can affect the DV (e.g., lighting, noise).
  • Dealing with problems: Standardized procedures, controlled environment.
• Experimental Realism: Forgetting the mundane realism part, the extent to which the experiment really 'grabs' participants, making them take the task seriously and behave naturally. High experimental realism can offset low mundane realism.

3.4 Observational Methods

Observational research involves watching and recording behavior in a systematic way. It can provide rich, detailed data about behavior in natural settings.

• Naturalistic Observation: Observing behavior in its natural environment without intervention.
  • Advantages: High ecological validity.
  • Disadvantages: Lack of control; difficult to establish causality; observer bias.
• Controlled Observation: Observing behavior in a structured environment (often a lab) where some variables are controlled.
  • Advantages: More control, easier to classify and compare behavior.
  • Disadvantages: Lower ecological validity.
• Overt Observation: Participants are aware they are being observed.
  • Advantages: Ethical (informed consent).
  • Disadvantages: Potential for reactivity (participants change behavior).
• Covert Observation: Participants are unaware they are being observed.
  • Advantages: Reduced reactivity.
  • Disadvantages: Ethical concerns (lack of informed consent, privacy violation).
• Participant Observation: The researcher becomes part of the group being observed.
  • Advantages: Rich qualitative data, deeper insight.
  • Disadvantages: Loss of objectivity, ethical dilemmas.
• Non-Participant Observation: The researcher observes from a distance without interacting with the group.
  • Advantages: Greater objectivity.
  • Disadvantages: Less in-depth understanding.
• Interobserver Reliability: The extent to which two or more observers agree on their observations, indicating consistency. This is crucial for objective observational data.

3.4.1 Observational Design: Sampling Procedures

To systematically record observations, researchers use sampling procedures:

• Event Sampling: Counting the number of times a specific behavior (event) occurs within a given time period. Useful for infrequent behaviors.
• Time Sampling: Recording behaviors at pre-determined intervals (e.g., every 30 seconds). Useful for continuous behaviors.

Structured vs. Unstructured Observations: Structured observations involve pre-defined coding schemes and categories, making data quantitative and easier to analyze. Unstructured observations record all relevant behavior, generating rich qualitative data, but can be susceptible to observer bias and harder to analyze.

3.5 Self-Report Techniques

Involve participants providing information about themselves, often through questionnaires or interviews.

• Questionnaires: A set of written questions used to gather information.
  • Construction: Careful wording, clear instructions, pilot testing.
  • Open Questions: Allow participants to answer in their own words, yielding qualitative data (e.g., "How do you feel about...").
  • Closed Questions: Provide fixed response options (e.g., "Yes/No," Likert scales), yielding quantitative data (e.g., "On a scale of 1-5, how much do you agree?").
  • Advantages: Can collect data from large samples efficiently; can be anonymous.
  • Disadvantages: Social desirability bias, misinterpretation of questions, response bias (e.g., acquiescence bias).
• Interviews: Direct, verbal questioning of participants.
  • Structured Interview: Pre-determined questions asked in a fixed order.
  • Unstructured Interview: General aims but no fixed questions, allowing for flexibility and follow-up questions.
  • Semi-structured Interview: A mix, with a core set of questions but flexibility to explore emergent themes.
  • Advantages: Rich, detailed data (especially unstructured); clarification possible.
  • Disadvantages: Time-consuming, interviewer effects (bias), social desirability.

3.6 Other Research Methods

• Case Study: An in-depth investigation of a single individual, group, event, or organization. Uses multiple sources of data (interviews, observations, documents).
  • Advantages: Rich, detailed insights; useful for rare phenomena; can generate hypotheses for future research.
  • Disadvantages: Findings may not be generalizable; prone to researcher bias; ethical challenges (confidentiality).
• Content Analysis: A systematic technique for analyzing communication (e.g., texts, images, videos) by identifying and counting themes, words, or concepts. Can involve converting qualitative data (e.g., interview transcripts) into quantitative data. [Simply Psychology].
• Meta-Analysis: A statistical technique that combines the results of multiple independent studies on a similar topic to derive an overall conclusion. It provides a more precise estimate of an effect size than any single study. Often the basis for systematic reviews.
• Systematic Review: A comprehensive review of all relevant research on a specific question, using rigorous, predefined methods to identify, select, assess, and synthesize findings. A meta-analysis may or may not be part of a systematic review.
• Effect Size: A quantitative measure of the strength of a phenomenon (e.g., the difference between two means, or the strength of a correlation). Unlike p-values, effect sizes indicate the practical significance of a result, not just statistical significance.
• Cross-Sectional Studies: Data are collected at one point in time from different groups (e.g., different age groups) to compare them.
  • Advantages: Quick, efficient.
  • Disadvantages: Cannot establish cause-and-effect; cannot track changes over time; prone to cohort effects (differences due to different life experiences of different age groups).
• Longitudinal Studies: Data are collected from the same group of participants over an extended period.
  • Advantages: Can track developmental changes, identify cause-and-effect relationships more plausibly.
  • Disadvantages: Time-consuming, expensive; participant attrition (drop-out); cohort effects remain a concern if findings are to be generalized to later generations.

3.7 Data Analysis and Evaluation in a Psychological Context

3.7.1 Types of Data

• Quantitative Data: Numerical data that can be measured and expressed in numbers (e.g., test scores, reaction times). Amenable to statistical analysis.
• Qualitative Data: Non-numerical, descriptive data, often in the form of words, images, or observations (e.g., interview transcripts, open-ended survey responses). Provides rich, in-depth understanding.
• Primary Data: Data collected first-hand by the researcher for the specific purpose of their study.
  • Advantages: Tailored to specific research questions.
  • Disadvantages: Time-consuming, expensive.
• Secondary Data: Data that already exists and has been collected by someone else for another purpose (e.g., government statistics, existing research papers, medical records).
  • Advantages: Cost-effective, accessible.
  • Disadvantages: May not perfectly fit research questions, quality varies, potential for bias.

3.7.2 Correlations

Correlational research examines the relationship between two or more variables as they naturally occur. It does not manipulate variables and therefore cannot establish cause and effect. [Simply Psychology].

• Correlational Hypothesis: Predicts a relationship between two co-variables (e.g., "There will be a relationship between hours spent studying and exam scores").
• Co-variables: The two variables being measured in a correlational study.
• Positive Correlation: As one variable increases, the other variable also increases (e.g., height and weight).
• Negative Correlation: As one variable increases, the other variable decreases (e.g., hours of sleep deprivation and cognitive performance).
• Correlation Coefficient: A numerical value (ranging from -1.0 to +1.0) that indicates the strength and direction of a linear relationship between two variables. +1.0 indicates a perfect positive correlation, -1.0 a perfect negative correlation, and 0 indicates no linear relationship.
• Significance: Determined by statistical tests, indicating the probability that the observed correlation occurred by chance.
• Scattergram (Scatter Plot): A graph that displays the relationship between two quantitative variables. Each point represents an individual's score on both variables.
• Linear Correlation: Data points cluster around a straight line on a scattergram.
• Curvilinear Correlation: The relationship between variables is non-linear, forming a curve (e.g., arousal and performance, often an inverted U-shape).

3.7.3 Measures of Central Tendency

These statistics locate the central point of a data set. [Simply Psychology].

• Mean: The arithmetic average of all scores. Most sensitive measure, uses all data points. Affected by outliers.
• Median: The middle score when data are ordered from lowest to highest. Unaffected by extreme scores.
• Mode: The most frequently occurring score. Useful for categorical data or when data contains multiple peaks.

3.7.4 Measures of Dispersion

These statistics describe the spread or variability of data around the central tendency.

• Range: The difference between the highest and lowest scores. Simple to calculate but sensitive to outliers.
• Standard Deviation (SD): A measure of the average distance of each score from the mean. A small SD indicates data points are close to the mean, a large SD indicates data points are widely spread. More informative than range.

3.7.5 Levels of Measurement (Scales of Measurement)

The type of data determines which statistical tests are appropriate. [Simply Psychology].

• Nominal Data: Categorical data that cannot be ordered (e.g., gender, eye color, type of therapy). Only mode is a meaningful measure of central tendency.
• Ordinal Data: Data that can be ranked or ordered, but the intervals between ranks are not equal or meaningful (e.g., Likert scale responses "Agree," "Neutral," "Disagree"; finishing positions in a race). Median is appropriate.
• Interval Data: Data with equal intervals between values, but no true zero point (e.g., temperature in Celsius or Fahrenheit, IQ scores). Mean, median, and mode are appropriate.
• Ratio Data: Data with equal intervals and a true zero point, meaning the absence of the quantity (e.g., height, weight, reaction time, number of errors). All measures of central tendency and dispersion are appropriate.

3.7.6 Display of Quantitative Data

Visual representation of data is crucial for understanding and communicating findings.

• Tables: Organized presentation of raw numbers or summary statistics.
• Bar Charts: Used for categorical (nominal, ordinal) data, showing frequencies or means of discrete categories. Gaps between bars.
• Histograms: Used for continuous (interval, ratio) data, showing the distribution of scores within intervals. No gaps between bars.
• Line Graphs: Used to show changes over time or relationships between continuous variables.
• Scattergrams (Scatter Plots): As discussed, for displaying correlations between two continuous variables.

3.7.7 Data Distributions

• Normal Distribution: A symmetrical, bell-shaped curve where the mean, median, and mode are all at the center. Many natural phenomena follow a normal distribution. [Investopedia].
• Skewed Distribution: Asymmetrical distribution.
  • Positive Skew (Right Skew): Tail points to the right, most scores are low, few high scores. Mode < Median < Mean.
  • Negative Skew (Left Skew): Tail points to the left, most scores are high, few low scores. Mean < Median < Mode.

3.7.8 Statistical Testing: Inferential Statistics

Inferential statistics allow us to make inferences about a population based on sample data. They test hypotheses and determine the probability that observed effects are due to chance.

• Significance: The probability that an observed result actually happened by chance. In psychology, the conventional level of significance (alpha level) is p < 0.05, meaning there's less than a 5% chance the results occurred randomly.
• Probability: A measure of the likelihood of an event occurring.
• Calculated Value: The value obtained from performing a statistical test.
• Critical Value: A threshold value found in tables. If the calculated value exceeds (or falls below, depending on the test) the critical value at a chosen significance level and degrees of freedom, the result is considered statistically significant.
• Table of Critical Values: Used to interpret the calculated value of a statistical test.
• One-tailed Test (Directional): Used when the hypothesis predicts a specific direction of difference or relationship.
• Two-tailed Test (Non-directional): Used when the hypothesis predicts a difference or relationship exists, but not its direction.
• Degrees of Freedom (df): A parameter used in many statistical tests, related to the number of independent pieces of information in a sample.
• Justifying the use of test: The choice of statistical test depends on:
  1. The research design (e.g., independent groups, repeated measures).
  2. The level of measurement of the data (nominal, ordinal, interval/ratio).
  3. Whether testing for a difference or a correlation.

Example - The Sign Test: A non-parametric test used for a Repeated Measures design with nominal data. It compares the number of positive and negative differences between two conditions. The calculated value (S) is the less frequent sign. This value is then compared to a critical value from a table, considering N (number of valid data points) and the chosen significance level (e.g., 0.05, one-tailed or two-tailed). If S is less than or equal to the critical value, the result is significant [Statistics How To].