Advanced Psychological Statistics

Two-way Anova F Test

August 26,2023
MALINIS, Kendies Thea T. Dr. Elmer G. De Jose
My Team

Kendies Thea Malinis, LPT Ryan Nitollano, RPm

Two-way Anova F Test ART

1
Overview:

History
Description
Assumptions and Limitations
Sample Problem
Computation using 3 methods: manual-
excel-jamovi/spss
References

2
HISTORY

Analysis of Variance (ANOVA) is a statistical technique used to

compare means of different groups. The two-way ANOVA extends
this concept to analyze the effects of two independent categorical
variables (factors) on a continuous dependent variable. The F-test
is a key component of ANOVA, used to assess whether the means
of the groups are significantly different from each other.

3
DESCRIPTIONS

In a two-way ANOVA, you have two factors, each with multiple

levels, and you're interested in how these factors, individually and
interactively, affect the dependent variable. The F-test is used to
determine whether there are significant differences in the means
between the groups formed by the combinations of factor levels.

4
ASSUMPTIONS:
The assumptions for a two-way ANOVA include:

• Independence: Observations are independent within and

between groups.
• Normality: The residuals (differences between observed and
predicted values) are normally distributed within each group.
• Homogeneity of Variance (Homoscedasticity): The variance of
the residuals is approximately constant across all groups.

5
Limitations of the assumptions for a two-way ANOVA:
1. Independence
• Limitation: Violation of independence assumptions can occur in cases where observations within
groups are correlated, such as in longitudinal or repeated measures designs.
• Impact: Correlated observations can lead to inflated or deflated F-statistics, potentially affecting the
accuracy of hypothesis testing.

2. Normality:
• Limitation: The normality assumption might be difficult to meet, especially with small sample sizes or
skewed data.
• Impact: Departure from normality can influence the accuracy of p-values and confidence intervals,
particularly for small sample sizes. However, ANOVA is known to be robust to mild deviations from
normality, especially when sample sizes are reasonably large.

6
Limitations of the assumptions for a two-way ANOVA
3. Homogeneity of Variance (Homoscedasticity):
⚬ Limitation: Violations of homoscedasticity occur when the variability of the dependent
variable is not consistent across groups.
⚬ Impact: Unequal variances can affect the significance of the F-test. If the assumption is
severely violated, the results might be unreliable. Robust techniques or transformations
might be considered in such cases.

4. Interaction Effect Interpretation:

⚬ Limitation: The interaction effect (if present) between two factors can be challenging to
interpret, especially with more than two factors or higher-order interactions.
⚬ Impact: Misinterpretation of interaction effects can lead to incorrect conclusions about the
relationships between variables.

7
Limitations of the assumptions for a two-way ANOVA
5. Multiple Comparisons:
⚬ Limitation: Conducting multiple pairwise comparisons without adjusting for the
family-wise error rate increases the risk of Type I errors (false positives).
⚬ Impact: Unadjusted multiple comparisons can lead to inflated significance levels and
incorrect identification of significant group differences.

6. Small Sample Sizes:

⚬ Limitation: With small sample sizes, the assumptions might be difficult to evaluate,
and the results can be sensitive to outliers.
⚬ Impact: Small samples can lead to less reliable estimates of population parameters
and can result in inaccurate inference.

8
Limitations of the assumptions for a two-way ANOVA
7. Non-Parametric Alternatives:
⚬ Consideration: When assumptions are severely violated or when data are
not amenable to normal distribution assumptions, non-parametric
alternatives (e.g., Kruskal-Wallis test) might be more appropriate.

8. Ecological Validity:
⚬ Consideration: The assumptions might not hold in real-world settings due to
the complexity and diversity of data.
⚬ Impact: The practical significance of findings might be limited if
assumptions do not hold under specific conditions.

9
Sample Problem:

Let's consider an example of a two-way ANOVA where

you're investigating the effects of two factors, A and B, on a
dependent variable Y. Factor A has three levels (A1, A2, A3),
and Factor B has two levels (B1, B2). You collect data from
each combination of factor levels and want to know if there
are significant differences in the means.

10
Computation Methods:

You can perform the computations for a two-way ANOVA F-

test manually, using Excel, or using statistical software like
Jamovi or SPSS. Here's a brief overview of the computation
steps for each method:

11
Manual Calculation:

• Calculate the overall mean (grand mean) of the dependent variable.

• Calculate the sum of squares (SS) for each source of variation (factors A,
B, interaction AB, and error).
• Calculate the degrees of freedom (df) for each source of variation.
• Calculate the mean squares (MS) for each source by dividing SS by df.
• Calculate the F-statistic for each source by dividing MS between groups by
MS within groups.

12
Sample problem using two-way anova
13

f test in manual computation:

Excel:

• Input your data into an Excel worksheet.

• Use the built-in ANOVA tool in Excel's Data Analysis
add-in.
• Specify the factors and dependent variable, and Excel
will compute the F-statistics, p-values, and other relevant
statistics.

14
Sample problem using two-way
15

anova f test in excel:

Jamovi/SPSS:

• Input your data into either Jamovi or SPSS.

• Navigate to the appropriate menu for performing a two-
way ANOVA.
• Specify the factors and dependent variable.
• The software will provide output including F-values, p-
values, and other relevant statistics.

16
Sample problem using two-way anova
17

f test in jamovi:
REFERENCES:

• Montgomery, D. C. (2017). Design and Analysis of Experiments. Wiley.

• Agresti, A., & Finlay, B. (2009). Statistical Methods for the Social Sciences.
Pearson.
• Jamovi: https://www.jamovi.org/
• SPSS: https://www.ibm.com/analytics/spss-statistics-software
• https://www.youtube.com/watch?v=0K-bfzLTRiY
• https://www.youtube.com/watch?v=Gsq3J0pQ2Rk&t=309s
• https://www.youtube.com/watch?v=_eDB7yV_SQc

18
 Advanced Psychological Statistics
Thank you
for listening!

August 26,2023
NEXT presenter: ART: Ryan Nitollano, RPm