SPSS Analysis Test Notes

ANCOVA NOTES


Correlation between covariate and DV:
 SPSS: Analyze > Correlate > Bivariate. Move DV & Covariate.
Check Pearson.
 Write-up: "A Pearson correlation indicated a significant positive/negative
relationship between [covariate] and reaction time, r([df]) = [value], p =
[value]."
Linearity assumption check:
 Based on correlation p-value and scatterplot inspection.
 SPSS (scatterplot): Graphs > Chart Builder > Scatter/Dot. DV on Y,
Covariate on X, Group for colors.
 Write-up: "The assumption of linearity was met/not met, as the correlation
was significant/non-significant (p = [value]) and visual inspection of the
scatterplot showed a linear/curvilinear pattern."
Homogeneity of regression slopes:
 SPSS: Analyze > General Linear Model > Univariate. DV: RT, Fixed Factor:
Group, Covariate: Covariate. Model > Custom. Add: Group,
Covariate, Group*Covariate.
 Write-up: "The test for homogeneity of regression slopes was non-
significant, Group × Covariate: F([df1], [df2]) = [value], p = [value],
indicating the assumption was met."
Group means (raw/descriptive):
 SPSS: Analyze > Compare Means > Means. DV: RT, Independent: Group.
 Write-up: "Mean reaction times were: Control (M = [value], SD = [value]),
Meaning Distractor (M = [value], SD = [value]), Sound Distractor (M =
[value], SD = [value])."
One-Way ANOVA:
 SPSS: Analyze > Compare Means > One-Way ANOVA. DV: RT, Factor:
Group. Options: Check Descriptive, Homogeneity of variance.
 Write-up: "A one-way ANOVA revealed a significant/non-significant effect of
group, F(2, [df]) = [value], p = [value], partial η² = [value]."
ANCOVA with adjusted means:
 SPSS: Analyze > General Linear Model > Univariate. DV: RT, Fixed Factor:
Group, Covariate: Covariate. Options: Move Group to Display Means for,
Check Compare main effects, Estimates of effect size.
 Write-up: "After controlling for [covariate], the ANCOVA showed a
significant/non-significant effect of group, F(2, [df]) = [value], p = [value],

, partial η² = [value]. Adjusted means were: Control = [value], Meaning
Distractor = [value], Sound Distractor = [value]."
Covariate evaluation value:
 Found in ANCOVA output under "Parameter Estimates" or "Estimated
Marginal Means" note.
 Write-up: "The covariate was evaluated at its mean value of [value]."
Error terms comparison:
 ANOVA: From ANOVA table, SS Error or Within Groups SS.
 ANCOVA: From ANCOVA table, Error SS.
 Write-up: "The error sum of squares decreased from [ANOVA value] in the
ANOVA to [ANCOVA value] in the ANCOVA, reflecting variance accounted
for by the covariate."
Plain English summary:
 Write-up: "We first verified that our control variable was linearly related to
reaction time and that this relationship was consistent across groups. After
statistically adjusting for baseline differences in processing speed, we
found that distractor type [did/did not] significantly affect naming speed,
indicating that [the interference effect is independent of/can be explained
by] individual differences in baseline speed."




ANOVA NOTES


One-Way ANOVA for group differences:
 SPSS: Analyze > Compare Means > One-Way ANOVA. DV in "Dependent
List", IV in "Factor". Options: Check Descriptive, Homogeneity of variance
test, Means plot.
 Write-up: "A one-way ANOVA revealed a significant effect of [IV] on [DV],
F([df_between], [df_within]) = [F value], p = [p value], η² = [effect size]."
Post-hoc tests (if ANOVA significant):
 SPSS: In One-Way ANOVA dialog, click Post Hoc. Check Tukey (equal
variances assumed) or Games-Howell (equal variances not assumed).
 Write-up: "Tukey post-hoc tests indicated that [Group A] (M = [value], SD
= [value]) was significantly different from [Group B] (M = [value], SD =
[value]), p = [value], but not from [Group C], p = [value]."
Assumption checks:
 Normality: SPSS: Analyze > Descriptive Statistics > Explore. DV to
"Dependent List", Group to "Factor List". Check Normality plots with tests.

,  Write-up: "Shapiro-Wilk tests indicated no significant deviations from
normality for any group (all p > .05)."
 Homogeneity of variance: From One-Way ANOVA output: Levene's Test.
 Write-up: "Levene's test indicated homogeneity of variances, F([df1],
[df2]) = [value], p = [value]."
Effect size calculation:
 η² (eta squared) = SS_between / SS_total (available in ANOVA output with
"Estimates of effect size" checked).
 Write-up: "The effect size was [small/medium/large] according to Cohen's
(1988) conventions (η² = [value])."
Descriptive statistics table:
 SPSS: Analyze > Descriptive Statistics > Explore or Analyze > Compare
Means > Means.
 Write-up: "Table 1 presents means and standard deviations for [DV]
across [IV] conditions."
Interpreting non-significant results:
 Write-up: "The one-way ANOVA revealed no significant effect of [IV] on
[DV], F([df_between], [df_within]) = [F value], p = [p value] > .05,
suggesting that [practical interpretation of null finding]."
Assumption violations and corrections:
 If variances unequal: Use Welch ANOVA (SPSS: Check Welch in One-Way
ANOVA options).
 Write-up: "Due to violation of homogeneity of variance (Levene's F =
[value], p < .05), Welch's ANOVA was conducted, yielding F([df1], [df2]) =
[value], p = [value]."
Reporting degrees of freedom:
 Write-up format: "F([df_between = groups-1], [df_within = N-groups]) =
[value], p = [value]."
Graphical presentation:
 SPSS: Graphs > Chart Builder > Bar Chart. DV on Y-axis, IV on X-axis. Add
error bars (95% CI).
 Write-up: "Figure 1 displays mean [DV] scores with 95% confidence
intervals across [IV] conditions."
Power analysis (if needed):
 G*Power software or SPSS syntax: Calculates achieved power.
 Write-up: "With α = .05 and N = [sample size], power to detect a
medium effect (f = .25) was [value]%."