Statistical Definitions and Categorizing Variables

Foundational Definitions * Statistics: Numbers that describe a sample. * Parameters: Numbers that describe a population. * Core Metrics: The values for mean, median, and mode (the three primary averages) can be either statistics or parameters depending on whether they are calculated for a sample or an entire population.
Two primary uses of statistics * Descriptive Statistics: Used purely to describe or summarize the characteristics of the specific sample being studied. It involves calculations like finding the mean or median to organize and present data clearly. * Inferential Statistics: Using sample statistics to make an "educated guess" or a "leap" of inference about a corresponding population parameter. * Correspondence: If the statistic is a sample mean, the inference is directed toward the population mean. If the statistic is a sample mode, the inference is directed toward the population mode.
Class Progression * The first portion of the course (typically through Chapter 6) focuses on descriptive statistics. * The remainder of the course (from Chapter 7 onwards) focuses on inferential statistics.

Definition: Sampling error is the natural difference, discrepancy, or error that exists between a sample statistic and its corresponding population parameter.
Core Role of the Field: If tasked with describing the field of statistics in one sentence, it could be defined as the study of how to deal with sampling error.
Characteristics of Sampling Error * It is unavoidable; even with a "perfect" random sample, the sample will not be identical to the population it was drawn from. * Hypothetical Example: * Assume a population has a mean ( $\mu$ ) of $50$ . * A random sample drawn from this population will likely not have a mean ( $\bar{x}$ ) of exactly $50$ . It might be $50.4$ . * Multiple random samples will produce different results, such as $50.1$ , $49.9$ , or occasionally more distant values like $55$ or $43$ . * The fact that these samples fluctuate rather than being exactly $50, 50, 50, 50, 50$ is the result of sampling error. * Research Implications: In actual research, the population parameter is usually unknown (a "big fat question mark"). Researchers calculate a sample statistic (e.g., $50.4$ ) and use inferential statistics to estimate the population value (e.g., "there is a certain percent chance the population mean is between this and this").

Correlational Method * Involves measuring two different variables to see if there is a relationship or trend between them. * Example: Height and Weight. Generally, taller people tend to be heavier and shorter people tend to be lighter. While individual exceptions exist (e.g., a friend who is taller but lighter), the general trend remains. * Limitations: One cannot draw cause-and-effect conclusions from correlational studies. In the height/weight example, height does not "cause" weight; rather, underlying biology/genetics/diet likely cause both traits to develop together.
Experimental Method * Involves two variables, but one is manipulated and one is measured. * Goal: To look for an effect of the manipulated variable on the measured variable. * Capability: This method allows for cause-and-effect conclusions. If a group receiving a high dose of a pill has less pain than those receiving a low dose, the researcher can conclude the pill caused the reduction in pain.

Independent Variable (IV) * The variable that is actively manipulated by the researcher. * Consists of different conditions or levels decide by the researcher. * Note: IVs exist only in experimental studies.
Dependent Variable (DV) * The variable that is simply measured by the researcher to see the effect of the IV. * Note: All studies (both experimental and correlational) involve measured variables, though they are primarily referred to as DVs in experimental contexts.
Experimental Groups vs. Control Groups * Control Group/Condition: Does not receive the treatment. They provide a baseline for comparison. * Experimental Group/Condition: Receives the treatment. * Treatment: The specific intervention believed to make a difference (e.g., a pill, a type of television content).
Examples of IV/DV Relationships * Pill Studies: IV is the dose (high, medium, low); DV is the outcome (e.g., pain levels, depression levels, memory capacity). * Aggression Studies: IV is the television content (Violent vs. Non-violent); DV is the measured aggression level in children. * Control condition: Watching a non-violent show (e.g., Daniel Tiger). * Experimental condition: Watching a violent show (e.g., Power Rangers or Paw Patrol). * Concentration Studies: IV is the presence of music (Music vs. Silence); DV is the ability to concentrate.

Random Selection (Random Sampling) * The process of picking a sample from a typically large population (e.g., picking $100$ people from millions). * Requires that everyone in the population has an equal chance of being selected.
Random Assignment * The process of dividing the selected sample into the experimental and control groups (e.g., splitting $100$ people into $50$ who get the pill and $50$ who do not). * Requires that every person in the sample has an equal chance of being assigned to either group (e.g., using a coin flip to decide: heads for the pill, tails for no pill). * Importance: This is the characteristic that defines a "true" experiment. It prevents self-selection bias. (e.g., if children were allowed to choose the violent TV show themselves, any subsequent aggression might be due to their personality rather than the TV show).

Operational Definition * The exact, specific way a researcher chooses to measure a variable for a particular study. * Utility: Different researchers may define the same concept differently. For example, one study might define "aggression" only as physical acts (punching/kicking), while another might include verbal aggression or social aggression (spreading rumors). This can explain why two studies on the same topic produce different results.
Quasi-Independent Variable * Variables that are "sort of" like independent variables but do not allow for random assignment. * Researchers can use these variables to create groups, but the participants already belong to those groups naturally. * Examples: * Gender: You cannot randomly assign a participant to be male or female. * Major: You cannot randomly assign a student to be a Physics major or an English major. * Quasi-Experiment: A study that uses these non-manipulated grouping variables. It cannot reach the same level of definitive cause-and-effect as a true experiment because random assignment is impossible.