Unit-III:

Descriptive
Statistics and
Hypothesis
Testing with
SPSS

©shumank deep
 Definition

• Tabulation refers to the systematic arrangement of data into rows

and columns for easy understanding and analysis.

Purpose

• Simplifies data presentation.

• Highlights patterns and trends.
• Facilitates comparison.

Tabulation Types of Tabulation

• Simple Tabulation: Single variable data (Frequency Tables)

• Complex Tabulation: Two or more variables (cross-tabulation).

of Data Structure of a Table:

• Title: Descriptive heading for the table.

• Row Headings: Categories of data.
• Column Headings: Measurements or classifications.

Importance of Tabulation:

• Saves time in interpreting data.

• Essential for statistical computations.

Software for Tabulation:

• SPSS, Excel, Google Sheets, etc.

©shumank deep
 Import Data:
Open your dataset in SPSS.

Select Frequencies:

Steps for
Navigate to Analyze > Descriptive Statistics > Frequencies.

Choose Variables:
Select the variables you want to analyze (e.g., Age or Income).

Frequency Customize Output:

Tabulation Include statistics like mean, median, and mode.

in SPSS
Add cumulative percentages and charts (e.g., histograms).

Generate the Table:

Click OK to produce the frequency table.

Export/Save Results:
Exportdeep
©shumank the results to a document or presentation format.
 Understand Column Layout:

• Value Column: Represents data categories (e.g., age ranges).

• Frequency Column: Shows the count of occurrences for each category.

Analyze Percentages:

• Look at valid percentages to assess category proportions.

Interpreting
• Example: “25% of respondents fall within the 25–34 age group.”

Cumulative Percentages:

a
• Useful for understanding data distribution thresholds.
• Example: “50% of respondents are under 35 years old.”

Spot Trends:

Frequency • Observe where data clusters or significant gaps exist.

Highlight Outliers:

Table • Identify unusually high or low frequencies in certain categories.

Report Key Findings:

• Summarize insights in plain language for your audience.

• Example: “The majority of respondents (45%) are aged 35–44, indicating
a mid-career demographic.”
©shumank deep
 Import Data:

Steps for
Open your dataset in SPSS.

Go to Analyze > Descriptive Statistics >

Navigate to Crosstabs:

Creating a
Crosstabs.

Choose a row variable (e.g., Gender).

Select Variables:

Cross-
Choose a column variable (e.g.,
Purchase Decision).

Click Statistics to add Chi-square tests

or percentages.

Tabulation
Customize the Table: Under Cells, select counts and
percentages (row, column, or total).

in SPSS
Click OK to view the cross-tabulation in
Generate the Crosstab: the SPSS output window.

Save or export the table for reporting

Export Results: purposes.
©shumank deep
 • Rows represent one variable’s categories (e.g.,
Understand Gender: Male/Female).
Table Structure: • Columns represent the other variable’s categories
Interpreting a Cross- (e.g., Purchase Decision: Yes/No).

Analyze Counts • Look at the counts and percentages in each cell.

and • Example: “60% of males made a purchase,
Percentages: compared to 40% of females.”

Identify • Compare distributions across rows and columns.

Patterns or • Example: “Younger participants show a higher
Trends: frequency of positive purchase decisions.”
Tabulation

• Review row and column totals to understand

Consider Totals:
overall distribution.

Link to • Use the data to confirm or challenge expectations

Hypotheses: or research hypotheses.

• Check Chi-square results to determine if

Statistical
relationships between variables are statistically
Significance:
significant.
©shumank deep
Understanding Chi-Square Test Results in
SPSS
Purpose of the •Tests if there is a significant association between two categorical variables.

Chi-Square Test
•Example: Does gender influence purchase decisions?

Key Outputs in •Observed Counts: Actual data counts in each category combination.
•Expected Counts: The counts expected if no relationship exists.
•Pearson Chi-Square Value: Statistic indicating the difference between observed and expected counts.

SPSS •Degrees of Freedom (df): Calculated as rows − 1 × columns − 1 .

•P-value (Asymptotic Significance): Determines statistical significance (p<0.05p < 0.05p<0.05).

Steps to Analyze •Compare Observed and Expected Counts for notable differences.
•Look at the Chi-Square Value and associated P-value.

SPSS Output •Interpret whether the relationship between variables is significant.

©shumank deep
 Interpreting Chi-Square Results

Effect Size (Cramér’s

Significance How to Report Actionable Insights
V)
• Significant (P < • Small: 0.1 • Significant • Use significant
0.05): • Medium: 0.3 Example: findings to guide
• Relationship exists • Large: 0.5 "Chi-Square = 8.5, df strategies.
between variables. = 1, P = 0.004, • Example: Adjust
• Example: Gender Cramér’s V = 0.2 marketing
affects purchase (small effect)." campaigns if gender
decisions. • Non-Significant impacts purchases.
• Non-Significant (P > Example: • Focus resources on
0.05): "Chi-Square = 3.2, df categories showing
= 1, P = 0.07." strong associations.
• No significant
relationship.

©shumank deep
 Definition
• Measures of location describe the central point of a dataset.
• They summarize where most values in the data tend to cluster.

Key Measures
• Mean (Average):
• Sum of all data values divided by the number of values.
Measures of • Example: Average exam score.
• Median (Middle Value):
Location • The central value when data is sorted in ascending order.
• Example: Median income to reduce skew by outliers.
• Mode (Most Frequent):
• The value that appears most often.
• Example: Most common shoe size.

Why Measures of Location are Important

• Provide a quick summary of data tendencies.
• Useful in comparing different datasets.

©shumank deep
 Definition:

• The arithmetic average of a data set.

Best Used When:

• Data is evenly distributed (no extreme outliers).

• Examples: Average income, average test scores.

Where to Use Physical Representation:

Mean • Imagine dividing 10 apples equally among 5 people;

each gets 2 apples (mean).

Example:

• Exam Scores: [70, 75, 80, 85, 90]

• Mean: (70+75+80+85+90)/5=80
Mean:{(70+75+80+85+90)}/5 =
80(70+75+80+85+90)/5=80
©shumank deep
Where to Use Median

Definition: • The middle value of an ordered dataset.

• Data has outliers or is skewed.

Best Used When: • Examples: Median household income, property prices.

Physical • If 5 people line up by height, the median is the height of

Representation: the middle person.

• House Prices: [50K, 100K, 150K, 200K, 10M]

Example: • Median: 150K (middle value when sorted).

©shumank deep
 The value that appears
Definition: most frequently.

Data has repeating

categories.
Best Used When: Examples: Favorite colors,

Where to
common shoe sizes.

Use Mode Physical If 10 people pick candies,

the mode is the flavor
Representation: picked most often.

Shoe Sizes: [8, 8, 9, 10, 8,

9, 10, 8]
Example: Mode: 8 (most common
size).

©shumank deep
 Mean: Best for evenly distributed
data without extreme values.

Choosing the Median: Best for skewed data with

Right Measure significant outliers.

Mode: Best for categorical data or

identifying the most common
occurrence.

©shumank deep
Graphical Data Display
Purpose of Summarize large datasets visually.
Graphical Identify patterns, trends, and outliers.
Displays Enhance understanding and decision-making.

Pie Charts
Common Types
Boxplots
of Graphs
Histograms

Why Use Simplify complex data.

Visuals? Engage your audience effectively.

Role of SPSS in
Graphical Easy creation and customization of visuals.
Displays

©shumank deep
 Definition
• Circular chart showing proportions of
categories.
When to Use
• Best for displaying parts of a whole

Pie
(percentages).
• Example: Market share distribution among
companies.

Charts
SPSS Steps
• Go to Graphs > Chart Builder.
• Choose Pie Chart and drag a variable to the
sector.
• Click OK to generate the chart.
Interpretation
• Larger slices indicate higher proportions.
• Example: 50% of customers prefer Product A.
©shumank deep
Histograms

Definition • Bar chart showing the frequency of data intervals.

When to Use • Analyze data distribution and detect skewness.

• Example: Age distribution in a population survey.

SPSS Steps • Go to Graphs > Chart Builder.

• Choose Histogram and select the variable to plot.

• Bars represent the frequency of data in intervals.

Interpretation • Example: Majority of participants are aged 25–34.
• Skewness: Symmetrical vs. left- or right-skewed distribution.

©shumank deep
Key Terms in a Boxplot
Median (Middle Line in the The middle value of the data.
Box) Splits the data into two equal halves.

Contains the middle 50% of the data.

Box (Middle Section) This range goes from the 25th percentile (lower
quarter) to the 75th percentile (upper quarter).

Whiskers (Lines Outside the Show the smallest and largest values within a
normal range.
Box) They do not include extreme or unusual values.

Outliers (Dots Outside the Values that are far away from the rest of the data.
Whiskers) These are considered unusual.

Minimum and Maximum The smallest and largest values that are still
(Ends of Whiskers) within the normal range.

©shumank deep
Boxplots in Action
When to Use Boxplots

• Compare data distributions across groups.

• Identify outliers and assess data symmetry.

Example Use Case

• Comparing test scores for different student groups.

Steps to Create in SPSS

• Go to Graphs > Chart Builder.

• Select Boxplot and specify variables (e.g., Test Score by Group).
• Click OK to generate the plot.

Interpreting a Boxplot

• Median Line: Shows the center of the data.

• Wide Box: Indicates high variability.
• Long Whiskers or Outliers: Suggest data spread or anomalies.
©shumank deep
 Definition
• Measures of variability describe the spread or dispersion of data.
• They show how much data points differ from the center (mean or
median).

Key Measures
• Range:
• Difference between the largest and smallest values.

Measures of • Example: Scores range from 50 to 90; Range = 40.

• Variance:

Variability • The average squared difference from the mean.

• Example: High variance indicates widely spread data.
• Standard Deviation (SD):
• The square root of variance, showing the average distance from the
mean.
• Example: SD = 5 means most values are ±5 from the mean.

Why Measures of Variability Matter

• They help understand data consistency and predictability.
• Example: Comparing two products’ defect rates with similar
averages but different variability.
©shumank deep
 Range
• Use for quick comparisons but be cautious of outliers.
• Example: Income range in two cities (e.g., $20K–$150K vs.
$40K–$80K).

When to
Standard Deviation
• Use to understand the spread of data around the mean.
• Smaller SD: Data points are close to the mean.

Use
• Larger SD: Data points are widely spread.

Illustration
• Dataset 1: {50, 52, 54, 56, 58} → Low SD (values close to the
mean).
• Dataset 2: {30, 50, 70, 90, 110} → High SD (values spread out).

Key Insights
• Small variability: Data is consistent and predictable.
• Large variability: Data shows more diversity or uncertainty.
©shumank deep
How to Measure Variability in SPSS
Open your dataset in SPSS.
Navigate to Analyze > Descriptive Statistics > Descriptives.
Steps to Calculate Select the variable(s) you want to analyze.
Variability in SPSS
Check the option for Standard Deviation (default).
Click OK to view the output table.

Range: Maximum value - Minimum value.

Interpreting the Output Standard Deviation (SD): Indicates data spread around the mean.
•Low SD: Data points are close to the mean.
•High SD: Data points are widely spread.

Small SD (< 10% of the mean): Data is tightly clustered.

Threshold for SD Moderate SD (10–30% of the mean): Data shows moderate spread.
Large SD (> 30% of the mean): Data is widely spread.

Example in SPSS Mean = 50, SD = 5 → Tight clustering.

Output Mean = 50, SD = 20 → High variability.

©shumank deep
Introduction to Measures of Shape

Why Measures of
Definition Key Metrics When to Use
Shape Matter
• Measures of • Skewness: • They help • Use to check if
shape describe Indicates the identify: data meets
the distribution symmetry of the • Asymmetry in assumptions for
of data relative to distribution. data statistical tests
symmetry and • Kurtosis: distribution. (e.g., normality).
tail behavior. Describes the • Presence of • Compare
"tailedness" or extreme values distributions
concentration of or heavy tails. across datasets.
data in the tails
vs. the center.

©shumank deep
 Definition of Skewness

Skewness - Symmetry • Describes the degree of asymmetry in a dataset.

Symmetrical Distribution:
• Skewness = 0.
• Right-Skewed (Positive Skew): Tail extends to the right (e.g.,
income data).
• Left-Skewed (Negative Skew): Tail extends to the left (e.g.,
retirement age).

Thresholds for Skewness

• Between -0.5 and 0.5: Data is approximately symmetric.
of Data

• Between -1 and -0.5 or 0.5 and 1: Moderately skewed.

• Beyond ±1: Highly skewed.
Example
• Dataset A: Skewness = 0.2 → Symmetrical.
• Dataset B: Skewness = 1.5 → Right-skewed.
©shumank deep
 Definition

•Kurtosis measures how data is distributed in terms of its peakedness

(how sharp or flat the center is) and tail heaviness (how extreme values
occur at the ends).

Breaking It Down

•Peakedness: Describes whether the center of the distribution is sharp or

Kurtosis -
flat.
•Sharp peak: Data is concentrated around the center.
•Flat peak: Data is more evenly spread out.
•Tails: Refers to the extreme values at the edges of the distribution.
•Heavy tails: More extreme values than normal.

Tails of the
•Light tails: Fewer extreme values than normal.

Thresholds for Kurtosis

•Excess Kurtosis near 0: Data is close to a normal distribution.

Distribution
•Example: Heights of adults in a population.
•Positive Excess Kurtosis (>0): Data has a sharp peak with heavy tails
(Leptokurtic).
•Example: Extreme events like high stock market returns.
•Negative Excess Kurtosis (<0): Data has a flat peak with light tails
(Platykurtic).
•Example: Data evenly spread, like grades in an easy test.

Thresholds for Interpretation

•Between -2 and +2: Indicates normal kurtosis.

•Beyond ±2: Suggests significant deviation from normality.
•Positive (>2): More extreme values than expected.
•Negative (<-2): Fewer extreme values than expected.

©shumank deep
 Definition

• Hypothesis testing is a statistical method to determine if there

is enough evidence in a sample to support a specific claim
about a population.

Key Terms

Introduction to • Null Hypothesis (H₀): Assumes no effect or relationship exists.

• Example: "The new drug has no effect on blood pressure."
• Alternative Hypothesis (H₁): Assumes an effect or

Hypothesis relationship exists.

• Example: "The new drug reduces blood pressure."

Testing Steps in Hypothesis Testing

• Define H₀ and H₁.

• Choose a significance level (α\alphaα), typically 0.05.
• Perform the test using sample data.
• Make a decision: Reject or fail to reject H₀.

Importance

• Helps make data-driven decisions in research and business.

©shumank deep
 Null Assumes no relationship or effect exists between variables.
Hypothesis Example: "There is no difference in test scores between students

Types of Hypotheses
(H₀) taught online and in-person."

Alternative Proposes that there is a relationship or effect.

Hypothesis Two types:
(H₁) •Directional: Specifies the direction of the effect (e.g., "Online teaching results in higher scores").
•Non-Directional: Indicates a difference but not the direction (e.g., "There is a difference in scores").

Directional Directional: Tested with a one-tailed test (e.g., "Treatment A is better

vs. Non- than Treatment B").
Directional Non-Directional: Tested with a two-tailed test (e.g., "There is a
Hypotheses difference between Treatment A and B").

Importance Null and alternative hypotheses guide statistical tests.

in Research Clear definitions ensure reproducible and valid results.

©shumank deep
 Definition
• Parametric tests assume the data follows a specific distribution,
typically normal.

Common Parametric Tests

• T-Test: Compares means between groups.
• Example: Comparing male and female test scores.

Parametric
• ANOVA: Compares means across multiple groups.
• Example: Comparing sales performance of three regions.

Tests
• Pearson Correlation: Measures the relationship between two
variables.
• Example: Correlation between study hours and exam scores.

When to Use
• Data is continuous.
• Follows normal distribution.
• Homogeneity of variances.

Advantages
• More powerful when assumptions are met.
©shumank deep
 Definition

• Non-parametric tests do not assume any specific data distribution.

Common Non-Parametric Tests

• Mann-Whitney U Test: Compares medians of two independent

groups.
• Example: Comparing satisfaction scores of two customer groups.

Non-Parametric • Wilcoxon Signed-Rank Test: Compares medians of paired

samples.
• Example: Pre-test vs. post-test scores.

Tests • Chi-Square Test: Tests for association between categorical

variables.
• Example: Relationship between gender and product preference.

When to Use

• Data is ordinal or categorical.

• Data does not meet parametric assumptions.

Advantages

• Works with small sample sizes.

• Robust to non-normal distributions.

©shumank deep
Parametric vs. Non- Non-
Parametric Tests Aspect Parametric
Parametric
Assumes No
Assumptions normal distribution
• Key Differences
• Choosing the Right Test
distribution assumption
• Use Parametric: When assumptions Ordinal,
are met (e.g., normality).
Continuous, categorical,
• Use Non-Parametric: When data is Data Type
skewed or assumptions are not met. interval/ratio or non-
• Example Scenarios normal
• Parametric: Testing average sales
between two regions. More efficient Works well
• Non-Parametric: Testing customer Efficiency with large with small
satisfaction rankings between two samples samples
stores.

©shumank deep
 Why Analyze Samples?

• To draw conclusions about a population based on a

subset of data.
• Identify differences or relationships between variables.

Overview of Key Types of Sample Tests

Hypothesis • One-Sample Test: Compare a sample mean to a known

population mean.
Testing for • Two Independent Samples Test: Compare means of two
independent groups.

Samples • Paired Samples Test: Compare means of the same

group at two different times.

Importance

• Helps validate research claims.

• Supports data-driven decision-making.

©shumank deep
 Test if the sample mean
Purpose differs significantly from
a known value.

One- Example
Testing whether the
average height of

Sample
students is different
from 5.5 feet.

Test - Threshold for

p < 0.05: Significant
difference.

Overview Significance p > 0.05: No significant

difference.

Useful for quality

4. Importance control or population-
level benchmarks.

©shumank deep
 1. Steps

• Go to Analyze > Compare Means > One-

One-
Sample T Test.
• Select the variable (e.g., test scores).

Sample
• Enter the test value (e.g., 50).
• Click OK.

Test in Interpret Results

SPSS • Check the p-value:

• If p < 0.05: Sample mean is significantly
different from the test value.
• If p > 0.05: No significant difference.

©shumank deep
Two Independent Samples Test - Overview

Purpose Compare means of two unrelated groups.

Example Comparing average scores of male and female students.

Threshold for p < 0.05: Significant difference between groups.

Significance p > 0.05: No significant difference.

Importance Common in experimental studies comparing control vs. treatment groups.

©shumank deep
Two Independent Samples Test in SPSS
Steps

• Go to Analyze > Compare Means > Independent-Samples T Test.

• Define grouping variable (e.g., gender).
• Select test variable (e.g., scores).
• Click OK.

Interpret Results

• Levene’s Test for Equality of Variances:

• If p > 0.05: Use 'Equal variances assumed' row.
• If p < 0.05: Use 'Equal variances not assumed' row.
• Check the p-value for significance.

©shumank deep
Paired Samples Test - Overview
Compare means of the same group at two different
Purpose times or under two conditions.

Comparing blood pressure before and after taking

Example medication.

p < 0.05: Significant difference between conditions.

Threshold for Significance p > 0.05: No significant difference.

Common in longitudinal studies or repeated

Importance measures experiments.

©shumank deep
Paired Samples Test in SPSS

Steps Interpret Results

Go to Analyze > Compare Means > Paired- Check the p-value:
Samples T Test. If p < 0.05: Significant difference between the
Select the two variables to compare (e.g., pre- two conditions.
test and post-test scores). If p > 0.05: No significant difference.
Click OK.

©shumank deep
 Threshold for Use p < 0.05 as the standard for determining significant results.
Significance

Effect Size Evaluate practical significance using effect size metrics (e.g., Cohen’s d).

Key Insights
for All Tests
Visualize Use charts (e.g., bar charts, boxplots) to illustrate findings.
Results

One-Sample: Benchmark testing.

Practical Independent Samples: Comparing distinct groups.
Applications Paired Samples: Pre- and post-intervention studies.

©shumank deep
Understanding Cohen’s d

What is Cohen’s d?
• A measure of effect size that shows the difference between two group averages in terms of standard deviation units.
• Helps to understand the practical importance of a difference.

Why Use Cohen’s d?

• Complements p-values by quantifying the magnitude of a difference.
• Useful in comparing results across studies.

Effect Size Thresholds:

• Small: 0.2
• Medium: 0.5
• Large: 0.8

Example Scenario:
• Comparing the effectiveness of two teaching methods on student test scores.

©shumank deep
 Definition

• Effect size is a quantitative measure of the magnitude of a

phenomenon or difference.
• Unlike p-values, it describes practical significance, showing how
much of an impact was observed.

Why is Effect Size Important?

• Complements statistical significance.

• Helps understand the practical relevance of findings.

Understanding • Allows comparison of results across studies.

Effect Size
Types of Effect Sizes

• Cohen’s d: Measures the difference between two group means.

• Pearson’s r: Measures the strength of the relationship between
two variables.
• Eta-Squared (η²): Indicates the proportion of variance explained
in ANOVA.

4. Key Thresholds

• Small Effect: 0.2 (Cohen’s d) or 0.1 (Pearson’s r).

• Medium Effect: 0.5 (Cohen’s d) or 0.3 (Pearson’s r).
• Large Effect: 0.8 (Cohen’s d) or 0.5 (Pearson’s r).
©shumank deep
 • To evaluate magnitude: How much of an effect or
relationship exists.
When to Use • In addition to p-values, to assess practical
implications.
• For comparing studies or conducting meta-analyses.

• Cohen’s d: Difference between means divided by

How to pooled standard deviation.

When and
• Pearson’s r: Correlation coefficient for relationships.
Calculate • Eta-Squared: Proportion of variance explained in
ANOVA.

How to Use
Effect Size
• Cohen’s d = 0.6: Moderate difference in test scores
Example between two teaching methods.
Interpretations • Pearson’s r = 0.4: Moderate positive correlation
between hours studied and exam scores.

• Effect size adds context and clarity to research

findings by quantifying their practical significance.
Key Takeaway • Always report alongside p-values for a complete
understanding of results.

©shumank deep
 Perform a T-Test:

• Go to Analyze > Compare Means > Independent-Samples T

Test.
• Select the two groups and the test variable.
• Obtain the group means and standard deviations from the
output.

Steps to Calculate Cohen’s d Manually:

Perform • Subtract the smaller mean from the larger mean.

• Divide by the average standard deviation of the two groups.

Cohen’s d in Example Calculation:

SPSS
• Group 1: Mean = 75, SD = 10
• Group 2: Mean = 70, SD = 12
• Approximation: Difference = 5; Average SD = 11
• Cohen’s d ≈ 5 ÷ 11 = 0.45 (Medium Effect).

Alternative Tools:

• Online calculators or software like R or Python can

automate this calculation.
©shumank deep
Interpreting Cohen’s d
What Does Cohen’s When to Use Example
Key Takeaway:
d Tell Us? Cohen’s d? Interpretation:
• Small Effect (d = • When comparing • If Cohen’s d = 0.6, • Cohen’s d is
0.2): Groups two independent the difference crucial for
slightly differ. groups (e.g., between groups is understanding the
• Medium Effect (d = treatment vs. moderate. real-world
0.5): Moderate, control). • This suggests that relevance of
meaningful • Works well with the intervention statistical results.
difference. normally (e.g., new teaching • Even significant p-
• Large Effect (d = distributed data. method) has a values can
0.8): Significant practical impact correspond to
and substantial on student negligible effect
difference. performance. sizes.

©shumank deep
 A statistical test used to
What is One-Way determine if there are
significant differences
ANOVA? between the means of
three or more groups.

Introduction Comparing one

independent variable with
When to Use
to One-Way
multiple levels (e.g., test
scores across teaching
methods).

Analysis of
Variance
Data is normally
distributed.

Assumptions Variances across groups

are approximately equal.

(ANOVA)
Observations are
independent.

Extends t-tests to more

Why is ANOVA than two groups.
Prevents inflation of Type I
Important? error from multiple
comparisons.
©shumank deep
 Groups (Levels): Categories of the independent
variable (e.g., Group A, Group B, Group C).

Between-Group Variance: Variability due to

differences between group means.

Key Terms in
One-Way Within-Group Variance: Variability within each group.

ANOVA F-Ratio: Ratio of between-group variance to within-

group variance. Larger F indicates greater
differences.

P-Value: Determines statistical significance of the F-

ratio.
©shumank deep
 Input Data

• Enter the dependent variable (e.g., test scores) and

independent variable (e.g., group).

Navigate in SPSS

Steps to • Go to Analyze > Compare Means > One-Way ANOVA.

Perform One-
Define Variables

• Set the dependent variable (e.g., scores) and factor

Way ANOVA (e.g., group).

in SPSS
Additional Options

• Click on Post Hoc to select tests (e.g., Tukey) for

pairwise comparisons.
• Click Options to include descriptive statistics.

Run the Test

• Click OK to generate results.

©shumank deep
 Descriptive Statistics

• Provides means and standard deviations

for each group.

ANOVA Table

ANOVA • Between-Group Sum of Squares (SS):

Variability due to differences between
Output in groups.
• Within-Group Sum of Squares (SS):
SPSS Variability within groups.
• F-Ratio and P-Value: Indicate whether the
group means differ significantly.

Post Hoc Tests

• Identify which specific groups differ.

©shumank deep
 Key Components

• F-Ratio: Larger values suggest greater differences

between groups.
• P-Value:
• p < 0.05: Reject the null hypothesis (significant
difference).
• p > 0.05: Fail to reject the null hypothesis (no
Interpreting significant difference).

ANOVA 2. Post Hoc Tests

Results
• Provide pairwise comparisons to identify specific
group differences.
• Example: Tukey test indicates Group A > Group B.

3. Effect Size (Eta-Squared, η²)

• Small: 0.01
• Medium: 0.06
• Large: 0.14
©shumank deep
 Comparing test scores across three teaching methods:
Scenario Traditional, Online, and Hybrid.

ANOVA Output: F(2, 27) = 5.43, p = 0.01.

Example
Results
Post Hoc Test: Online scores > Traditional and Hybrid.

Interpretation
Teaching method significantly affects test scores.
Conclusion
Online method shows significantly higher scores.

Implementing online teaching strategies may enhance

Insights learning outcomes.

©shumank deep
Limitations and Best Practices
Limitations Best Practices SPSS Advantages

• Sensitive to outliers. • Verify assumptions • Automates

• Assumes equal (normality and calculations.
variances (use homogeneity of • Provides detailed
Welch’s ANOVA if variances). statistical
violated). • Include Post Hoc summaries.
tests for detailed
insights.
• Report effect size for
practical
significance.

©shumank deep
 What is ANCOVA?
• Combines ANOVA and regression to evaluate mean

Introduction differences while controlling for covariates.

Purpose
to Analysis • Adjusts the dependent variable by removing the

of influence of a covariate.
• Increases precision by reducing variability caused

Covariance by the covariate.

Key Applications
(ANCOVA) • Evaluating the effect of an intervention while
accounting for preexisting differences (e.g.,
comparing teaching methods while controlling for
prior knowledge).

©shumank deep
 Dependent Variable
• The outcome being measured (e.g., test scores).

Independent Variable
• The categorical factor being tested (e.g., teaching
Key method).

Components of Covariate
ANCOVA • A continuous variable that influences the dependent
variable (e.g., prior knowledge).

Assumptions
• Linearity between covariate and dependent variable.
• Homogeneity of regression slopes.
• Normal distribution and homogeneity of variance.

©shumank deep
 Prepare • Ensure the covariate and dependent variable are

Data continuous.

Navigate in • Go to Analyze > General Linear Model > Univariate.

Steps to SPSS
Perform Set • Dependent Variable: Outcome (e.g., test scores).
• Fixed Factor: Group (e.g., teaching method).

ANCOVA in Variables • Covariate: Adjusting variable (e.g., prior knowledge).

SPSS Additional • Click Options to display descriptive statistics.

Options • Click Plots to visualize group means.

Run the • Click OK to view results.

Analysis
©shumank deep
 Key Outputs
• Descriptive Statistics: Adjusted and unadjusted means for
each group.
• Tests of BetweenSubjects Effects: Check the significance
of the independent variable after controlling for the
covariate.

Interpreting • Estimated Marginal Means: Means adjusted for the

covariate.
ANCOVA PValue Interpretation
Results • Significant (p < 0.05): Independent variable has a
significant effect.
• NonSignificant (p > 0.05): No significant effect.

Effect Size (Partial EtaSquared)

• Small: 0.01
• Medium: 0.06
• Large: 0.14

©shumank deep
 Controls for extraneous variables, improving accuracy.
Benefits
Reduces error variance, increasing statistical power.

Benefits and Limitations

Sensitive to assumption violations (e.g., linearity, homogeneity of
regression slopes).

Limitations
Misinterpretation risk if covariates are not independent of the
independent variable.

Verify assumptions before analysis.

Best Use graphical methods to check linearity and regression slopes.
Practices
Report both adjusted means and effect sizes for clarity.

©shumank deep