Quick Answer: Simple Effects Analysis Result

Simple Effects Analysis answers a more specific question than the ordinary main-effect table. A main effect asks whether a factor matters on average. A simple effect asks whether that factor matters within a specific level of another factor. This is especially important when an interaction is present because the effect of one factor may change depending on the level of the second factor.

In the worked example, the analysis is organized around two categorical factors and a continuous outcome. The interaction profile plot, simple effects charts, cell mean heatmap and pairwise comparison map are used together. The result should be reported as a cell-level story: which factor changes the outcome inside each subgroup, where the effect is strongest, and where the effect is weak or absent.

What Is Simple Effects Analysis?

Simple Effects Analysis is a follow-up method used in factorial designs. It examines the effect of one independent variable at a particular level of another independent variable. In a two-way ANOVA, the model may include Factor A, Factor B and the A × B interaction. If the interaction is meaningful, the average main effect may be incomplete or misleading.

For example, Factor A may appear to have a modest average effect. But inside one level of Factor B, Factor A may have a strong effect, while inside another level of Factor B, Factor A may have little or no effect. Simple Effects Analysis separates those conditional effects so the interaction can be explained properly.

The method is sometimes called simple main effects analysis. The phrase means that the analyst is still studying a main effect, but only inside a specific condition. The result is usually reported with cell means, standard errors, confidence intervals, F tests or pairwise comparisons.

Before reading this guide, review factorial ANOVA, one-way ANOVA, ANOVA assumptions, effect size, eta squared, p-values and confidence intervals.

When to Use Simple Effects Analysis

Use Simple Effects Analysis when a factorial ANOVA includes two or more factors and the interpretation requires conditional comparisons. It is most commonly used after a significant or practically meaningful interaction effect.

Simple Effects Analysis Formula

In a two-way ANOVA, the general model is:

A simple effect of Factor A within a specific level of Factor B compares A-level means while holding B fixed:

A simple effect of Factor B within a specific level of Factor A compares B-level means while holding A fixed:

For an F-style simple effect test, the mean square for the simple effect is compared with the appropriate error term:

Null and Alternative Hypotheses for Simple Effects Analysis

Simple Effects Analysis uses conditional hypotheses. The hypothesis is not only “Factor A has an effect.” It is “Factor A has an effect within this level of Factor B.”

Dataset and Variables Used

The worked example is structured as a two-factor ANOVA problem. The dependent variable is continuous. Factor A and Factor B are categorical predictors. The goal is to understand whether the effect of one factor changes across levels of the other factor.

Simple Effects Analysis is easiest to understand when paired with factorial ANOVA, ANOVA in Python, ANOVA in R, ANOVA in SPSS, descriptive statistics, box plot interpretation and effect size.

Two-Way ANOVA Context and Simple Effects Decision Table

Simple Effects Analysis is normally interpreted after the two-way ANOVA table has been reviewed. The ANOVA table shows the main effect of Factor A, the main effect of Factor B and the A × B interaction. When the interaction is meaningful, the simple effects table becomes the main interpretation tool.

Simple Effects Interpretation Summary

Python Chart-by-Chart Interpretation

The Python charts show the complete Simple Effects Analysis workflow. They move from the interaction profile to simple effects of each factor, cell means, pairwise comparisons, critical difference thresholds, marginal versus cell means and distribution context.

Python Chart 1: Simple Effects Interaction Profile

The interaction profile plot is the first chart to read. If the lines are parallel, the factor effects are similar across levels. If the lines separate, converge, cross or show different slopes, the chart suggests that one factor behaves differently depending on the level of the other factor.

This chart explains why Simple Effects Analysis is needed. The visible pattern should guide the follow-up tests. The final report should describe whether the interaction is ordinal, disordinal, crossing or mainly a difference in effect size across levels.

Python Chart 2: Simple Effects of Factor A Within Factor B

This chart isolates the effect of Factor A inside each level of Factor B. Instead of asking whether Factor A matters overall, it asks whether Factor A matters when B is held at a specific level.

The interpretation should name the B level where the Factor A difference is strongest. If Factor A is clear in one B level but weak in another, the interaction should be reported as a conditional Factor A effect.

Python Chart 3: Simple Effects of Factor B Within Factor A

This chart reverses the conditioning direction. It tests the effect of Factor B inside each level of Factor A. This is important because an interaction can be explained from either direction, depending on the research question.

The best report chooses the direction that answers the research question most clearly. If the study asks how Factor B works for each group of Factor A, this chart becomes the main simple effects result.

Python Chart 4: Simple Effects Cell Mean Heatmap

The cell mean heatmap gives the most direct view of the interaction. Each cell represents the mean outcome for one combination of Factor A and Factor B. Stronger or weaker cells show where the outcome is highest and lowest.

This chart prevents a common mistake: interpreting only marginal means. When an interaction exists, the cell means are often more important than the average main effects. The final explanation should match the cell mean pattern visible in this heatmap.

Python Chart 5: Simple Effect Pairwise Comparison Map

The pairwise comparison map shows which specific conditional comparisons are strongest. This is the chart that turns the interaction pattern into reportable differences. It helps identify which levels differ inside each simple effect test.

The final article should report these comparisons with direction. Instead of saying “there was a significant simple effect,” state which group had the higher mean and at which level of the other factor the difference occurred.

Python Chart 6: Simple Effect Critical Difference Thresholds

The critical difference chart shows whether observed conditional differences are large enough to pass the selected decision threshold. A difference may look visible in a mean chart, but the critical threshold helps determine whether it is strong enough statistically.

This chart is useful for avoiding overclaiming. If a difference is below the threshold, it can be described as small or descriptive, but it should not be reported as a statistically supported simple effect.

Python Chart 7: Marginal Means vs Cell Means

This chart explains why marginal means can be misleading when an interaction exists. Marginal means average across the second factor, while cell means show each subgroup combination. If the cell means move in different directions, the marginal mean can hide the real pattern.

The chart supports the main teaching point of Simple Effects Analysis: when the interaction is meaningful, cell-level interpretation is more informative than broad average interpretation.

Python Chart 8: Simple Effects Distribution Context

The distribution context chart shows the spread of scores inside the A × B cells. It helps readers understand whether cell mean differences are supported by compact distributions or whether there is large overlap.

This chart should be used before final reporting because simple effects are not only about mean height. They also depend on variability, sample size and overlap inside the cells.

R Validation Chart

The validation chart repeats the interaction profile in a second output set. It is useful for checking that the interaction pattern is not only a Python artifact and that the same visual structure appears in the validation workflow.

Validation Chart 1: Simple Effects Interaction Profile

The validation interaction profile should be compared with the Python profile plot. If the same nonparallel or conditional pattern appears, it strengthens the conclusion that simple effects are needed.

This chart also helps confirm the direction of reporting. The final interpretation should explain the interaction through the factor direction that makes the profile easiest to understand.

SPSS, R, Python and Excel Workflows for Simple Effects Analysis

Simple Effects Analysis can be completed in SPSS, R, Python and Excel. SPSS is useful for estimated marginal means and pairwise comparisons. R and Python are best for reproducible model fitting, interaction plots, cell means and custom simple effects tables. Excel can calculate cell means and teach the logic, but full inferential testing is easier in statistical software.

SPSS Workflow

R Workflow

Python Workflow

Excel Workflow

Excel can support Simple Effects Analysis by calculating cell means, marginal means and simple mean differences. It is best for teaching the logic. For formal F tests, adjusted pairwise comparisons and interaction output, SPSS, R or Python is preferred.

Code Blocks for Simple Effects Analysis

SPSS Syntax

UNIANOVA outcome BY factor_a factor_b
  /METHOD = SSTYPE(3)
  /INTERCEPT = INCLUDE
  /PRINT = DESCRIPTIVE ETASQ HOMOGENEITY
  /EMMEANS = TABLES(factor_a*factor_b)
  /EMMEANS = TABLES(factor_a*factor_b) COMPARE(factor_a) ADJ(SIDAK)
  /EMMEANS = TABLES(factor_a*factor_b) COMPARE(factor_b) ADJ(SIDAK)
  /DESIGN = factor_a factor_b factor_a*factor_b.

R Code

data <- read.csv("dataset.csv")
data$factor_a <- factor(data$factor_a)
data$factor_b <- factor(data$factor_b)

# Two-way ANOVA with interaction

model <- aov(outcome ~ factor_a * factor_b, data = data)
summary(model)

# Cell means

aggregate(outcome ~ factor_a + factor_b, data = data, mean)

# Simple effects with emmeans

# install.packages("emmeans")

library(emmeans)

emm <- emmeans(model, ~ factor_a * factor_b)

# Effect of Factor A within each level of Factor B

pairs(emmeans(model, ~ factor_a | factor_b), adjust = "sidak")

# Effect of Factor B within each level of Factor A

pairs(emmeans(model, ~ factor_b | factor_a), adjust = "sidak")

Python Code

import pandas as pd
import itertools
from scipy import stats
import statsmodels.api as sm
import statsmodels.formula.api as smf
from statsmodels.stats.multitest import multipletests

df = pd.read_csv("dataset.csv")

df["factor_a"] = df["factor_a"].astype("category")
df["factor_b"] = df["factor_b"].astype("category")

# Two-way ANOVA with interaction

model = smf.ols("outcome ~ C(factor_a) * C(factor_b)", data=df).fit()
anova = sm.stats.anova_lm(model, typ=2)
print(anova)

# Cell means

cell_means = (
df.groupby(["factor_a", "factor_b"])["outcome"]
.agg(["count", "mean", "std"])
.reset_index()
)
print(cell_means)

# Simple effect of Factor A within each level of Factor B

rows_a_within_b = []

for b_level in df["factor_b"].cat.categories:
subset = df[df["factor_b"] == b_level]
a_levels = list(subset["factor_a"].cat.categories)

```
raw_p_values = []
comparisons = []

for a1, a2 in itertools.combinations(a_levels, 2):
    x1 = subset.loc[subset["factor_a"] == a1, "outcome"]
    x2 = subset.loc[subset["factor_a"] == a2, "outcome"]

    t_stat, raw_p = stats.ttest_ind(x1, x2, equal_var=True)
    raw_p_values.append(raw_p)
    comparisons.append((b_level, a1, a2, x1.mean() - x2.mean(), t_stat, raw_p))

if raw_p_values:
    reject, p_adj, _, _ = multipletests(raw_p_values, method="sidak")
    for item, adj_p, sig in zip(comparisons, p_adj, reject):
        rows_a_within_b.append(list(item) + [adj_p, sig])
```

a_within_b = pd.DataFrame(rows_a_within_b, columns=[
"factor_b_level", "factor_a_1", "factor_a_2",
"mean_difference", "t_statistic", "raw_p",
"sidak_adjusted_p", "significant"
])

print(a_within_b)

# Simple effect of Factor B within each level of Factor A

rows_b_within_a = []

for a_level in df["factor_a"].cat.categories:
subset = df[df["factor_a"] == a_level]
b_levels = list(subset["factor_b"].cat.categories)

```
raw_p_values = []
comparisons = []

for b1, b2 in itertools.combinations(b_levels, 2):
    x1 = subset.loc[subset["factor_b"] == b1, "outcome"]
    x2 = subset.loc[subset["factor_b"] == b2, "outcome"]

    t_stat, raw_p = stats.ttest_ind(x1, x2, equal_var=True)
    raw_p_values.append(raw_p)
    comparisons.append((a_level, b1, b2, x1.mean() - x2.mean(), t_stat, raw_p))

if raw_p_values:
    reject, p_adj, _, _ = multipletests(raw_p_values, method="sidak")
    for item, adj_p, sig in zip(comparisons, p_adj, reject):
        rows_b_within_a.append(list(item) + [adj_p, sig])
```

b_within_a = pd.DataFrame(rows_b_within_a, columns=[
"factor_a_level", "factor_b_1", "factor_b_2",
"mean_difference", "t_statistic", "raw_p",
"sidak_adjusted_p", "significant"
])

print(b_within_a)

Excel Formula Pattern

Cell mean:
=AVERAGEIFS(outcome_range,A_range,A_level,B_range,B_level)

Cell count:
=COUNTIFS(A_range,A_level,B_range,B_level)

Simple effect of A within B level 1:
=Mean_A1_B1-Mean_A2_B1

Simple effect of B within A level 1:
=Mean_A1_B1-Mean_A1_B2

Cell standard deviation:
=STDEV.S(FILTER(outcome_range,(A_range=A_level)*(B_range=B_level)))

Sidak adjusted p-value:
=1-(1-raw_p)^m

Decision:
=IF(adjusted_p<0.05,"Significant","Not significant")

APA Reporting Wording for Simple Effects Analysis

A two-way ANOVA was conducted to examine the effects of Factor A and Factor B on the outcome. Because the A × B interaction was meaningful, Simple Effects Analysis was used to interpret the conditional pattern. The simple effect of Factor A was examined separately within each level of Factor B, and the simple effect of Factor B was examined separately within each level of Factor A.

The simple effects results showed that the group differences were not uniform across all cells. The interaction profile and cell mean heatmap indicated that the size and direction of the effect changed across factor levels. Follow-up pairwise comparisons were used to identify which conditional differences were statistically meaningful. Therefore, the interaction should be reported using cell means and conditional comparisons rather than only marginal main effects.

Common Mistakes in Simple Effects Analysis

Downloads and Resources

Use the downloadable Python report to verify the Simple Effects Analysis charts, interaction profile, cell means, pairwise comparison map, critical difference thresholds and distribution context.

Related Guides

Continue with these related Salar Cafe guides to understand factorial ANOVA, interaction effects, assumptions, p-values and effect sizes.