UNIT 1: Significance of Statistics
Meaning of Statistics
Statistics is the science of:
● Collecting
● Organizing
● Analyzing
● Interpreting data
Descriptive vs Inferential Statistics
Descriptive Statistics
Describes data collected.
Examples:
● Mean
● Median
● Standard deviation
● Graphs
It summarizes the sample only.
Inferential Statistics
Draws conclusions about population based on sample.
It answers:
“Is this difference real or due to chance?”
Levels of Significance (α)
Level of significance = probability of rejecting a true null hypothesis.
Common levels:
● 0.05 (5%)
● 0.01 (1%)
If p < 0.05 → result significant.
Errors in Statistical Decision Making
Type I Error (α)
Rejecting true null hypothesis.
False positive.
Type II Error (β)
Failing to reject false null hypothesis.
False negative.
Sampling Distribution
Sampling distribution = distribution of sample means taken repeatedly from population.
Important concept:
Central Limit Theorem:
Sampling distribution becomes normal with large sample size.
Confidence Intervals
Confidence interval gives range within which population parameter lies.
Example:
95% CI = we are 95% confident true mean lies in this range.
It provides more information than p-value.
UNIT 2: Inferential Statistics
Difference from Descriptive Statistics
Descriptive Inferential
Summarizes data Generalizes
data
No probability Uses probability
No hypothesis Hypothesis
testing testing
Statistical Hypothesis Testing
Steps:
1. State Null Hypothesis (H₀)
2. State Alternative Hypothesis (H₁)
3. Choose significance level
4. Calculate test statistic
5. Compare with critical value
6. Decision
Effect Size
Effect size measures magnitude of difference.
It tells:
“How strong is the effect?”
Example:
● Cohen’s d
● Eta squared (η²)
Even small p-values may have small effect size.
Statistical Power
Power = probability of detecting real effect.
Power increases when:
● Sample size increases
● Effect size increases
● Alpha increases
Low power → Type II error risk.
UNIT 3: t-Test
t-test compares two means.
1️⃣ Independent Samples t-test
Used when:
Two different groups.
Example:
Male vs Female anxiety scores.
Formula concept:
Difference between means ÷ standard error
Assumptions:
● Normal distribution
● Equal variance
2️⃣ Paired Samples t-test (Correlated)
Used when:
Same group measured twice.
Example:
Before therapy vs After therapy.
Also called:
Dependent t-test.
Conceptual Difference
Independent t Paired t
Different Same group
groups
Between Within
subjects subjects
UNIT 4: One Way ANOVA
ANOVA compares more than 2 means.
Independent Samples ANOVA
Used when:
3 or more independent groups.
Example:
Stress levels in 3 professions.
It calculates:
F-ratio = Between group variance / Within group variance
If F significant → at least one group differs.
Post hoc test needed.
Repeated Measures ANOVA
Used when:
Same participants tested multiple times.
Example:
Pre-test, mid-test, post-test.
Controls individual differences.
UNIT 5: Non-Parametric Statistics
Parametric vs Non-Parametric
Parametric Non-Parametric
Normal distribution No normality required
required
Uses mean Uses median/ranks
More powerful Less powerful but
flexible
Chi-Square (χ²)
Used for:
Categorical data.
Example:
Gender × Preference.
Tests:
Association between variables.
Median Test
Compares medians between groups.
Used when data not normally distributed.
Mann-Whitney U Test
Non-parametric alternative to independent t-test.
Used for:
Two independent groups.
Wilcoxon Signed Rank Test
Alternative to paired t-test.
Used for:
Related samples.
Kruskal-Wallis Test
Alternative to one-way ANOVA.
Used for:
3 or more independent groups.
Friedman Test
Alternative to repeated measures ANOVA.
Used for:
Related groups.
UNIT 6: Correlations (Special Types)
Biserial Correlation
Used when:
One continuous variable
One artificially dichotomized variable
Example:
Pass/Fail derived from marks.
Point-Biserial Correlation
Used when:
One continuous
One naturally dichotomous variable
Example:
Gender & test score
Tetrachoric Correlation
Used when:
Both variables artificially dichotomized.
Phi Correlation
Used when:
Both variables naturally dichotomous.
Example:
Yes/No × Male/Female
Inferential statistics helps us:
1. Compare groups (t-test, ANOVA)
2. Study relationships (correlation)
3. Test associations (Chi-square)
4. Estimate population values (CI)
Always remember:
● Significance ≠ importance
● Effect size matters
● Power matters
● Assumptions matter