Summary of Introduction to Statistical Fundamentals
Introduction to Statistical Fundamentals: A Student Guide
Introduction
Statistics provides tools to collect, summarize and draw conclusions from data. This study guide covers practical methods you will use in applied research: descriptive summaries, inferential techniques, types of random variables, sampling rationales and common sampling methods.
Definition: Statistics is the science of collecting, organizing and interpreting data to make informed decisions.
1. Descriptive Statistics
Descriptive statistics summarize and describe features of a dataset so you can quickly understand its center, spread and shape.
Key measures
- Mean (Average): Sum of values divided by number of observations. Useful for symmetric data.
- Median (Middle value): The middle observation after sorting; robust to outliers.
- Mode (Most frequent value): The value(s) that occur most often.
- Charts and graphs: Histograms, boxplots, bar charts and scatterplots visualize distributions and relationships.
Definition: Descriptive statistics are numerical and graphical methods that summarize observed data.
Example: Scores of 4 students: $95$, $90$, $85$, $80$. The mean is $\frac{95+90+85+80}{4}=87.5$, the median is $\frac{90+85}{2}=87.5$ and there is no repeated mode.
2. Inferential Statistics
Inferential statistics let you draw conclusions about a larger population based on a sample.
Goals
- Estimate population parameters (e.g., population mean) using sample statistics.
- Test hypotheses about relationships or differences.
- Quantify uncertainty using confidence intervals and p-values.
Definition: Inferential statistics are methods that use sample data to make probabilistic statements about a population.
Common tools and when to use them
- Hypothesis testing: Decide if an observed effect is likely genuine or due to chance (e.g., $t$-test, chi-square test). Example: test if a new teaching method changes average test scores.
- Confidence intervals: Provide a range where the true parameter likely lies, e.g., a 95% CI for a mean.
- Regression analysis: Model relationships between variables (predictor(s) and outcome). Example: test whether study hours predict test score using linear regression $y = \beta_0 + \beta_1 x + \varepsilon$.
- ANOVA (Analysis of Variance): Compare means across more than two groups.
Example: You survey $50$ students and $40$ say they like pizza; you can estimate the proportion who like pizza in the whole school and calculate a confidence interval to express uncertainty.
3. Types of Random Variables
Random variables describe outcomes that vary by chance.
| Type | Description | Examples |
|---|---|---|
| Discrete | Takes countable values (typically integers) | Number of cars in a parking lot; number of defective items in a box |
| Continuous | Takes any value within a range (measurable) | Height of a student; time to run $100,$m |
Definition: A random variable is a variable whose value depends on chance.
Example: Number of defective items in a box can be $0$, $1$, $2$, ... (discrete). Height can be $170.3,$cm, $171,$cm, etc. (continuous).
4. Why Use a Sample instead of a Census?
Collecting data from a sample is often preferable to measuring the entire population (a census).
- Cheaper — uses less money
- Faster — requires less time
- More practical and controllable — easier to ensure data quality
- Sometimes necessary — e.g., destructive testing (testing blood or breaking items)
Definition: A sample is a subset of the population used to estimate population characteristics.
5. Sampling Methods
Sampling methods determine how units are selected from the population. Probability-based methods help avoid bias.
| Category | Meth
Sign up for the full summaryAlready have an account? Sign in
Statistical Methods Overview
Klíčové pojmy: Descriptive statistics summarize data using mean, median, mode and graphs, Inferential statistics generalize from a sample to a population using tests and confidence intervals, Use hypothesis tests (e.g., t-test, chi-square) to decide if an effect is likely real, Confidence intervals give a probabilistic range for population parameters, Regression models relationships: $y=\beta_0+\beta_1 x + \varepsilon$, ANOVA compares means across more than two groups, Discrete variables take countable values; continuous variables take values in a range, Probability sampling (simple, systematic, stratified) reduces selection bias, Samples are cheaper, faster and sometimes the only feasible option, Sampling error is the difference between sample estimates and true population values, Population ($N$) vs sample ($n$): know which you reference, Stratified sampling improves precision for subgroup estimates