The quick route: one call per dependent variable
analyze_and_report() runs the whole per-DV pipeline: it
checks the assumptions (and phrases the justification for the methods
section), builds the matching ggstatsplot figure with
automatic parametric/non-parametric selection, reports the omnibus test,
and – for three or more groups – the significant post-hoc
comparisons.
res <- analyze_and_report(
main_df,
dv = "tlx_mental", iv = "ConditionID",
design = "within",
ylab = "Mental Demand (TLX)"
)
#> Shapiro--Wilk tests indicated no significant deviation from normality in any group (all $p \geq 0.05$); therefore, parametric tests were used.
#> A ANOVA estimation for factorial designs using 'afex' found a significant effect of \ConditionID on tlx\_mental (\F{1.92190837641126}{44.2038926574591}{5.75}, \p{0.007}, r=0.11).
#> A Student's t post-hoc test found that Baseline was significantly higher (\m{54.92}, \sd{11.10}) in terms of tlx\_mental compared to HUD (\m{45.88}, \sd{9.76}); \padj{0.033}).
#> A Student's t post-hoc test found that Baseline was significantly higher (\m{54.92}, \sd{11.10}) in terms of tlx\_mental compared to LED (\m{45.21}, \sd{12.51}); \padj{0.015}).The result carries everything separately, so you can place the pieces where they belong:
res$methods # for the Methods section
#> [1] "Shapiro--Wilk tests indicated no significant deviation from normality in any group (all $p \\geq 0.05$); therefore, parametric tests were used."
res$text # the omnibus result
#> [1] "A ANOVA estimation for factorial designs using 'afex' found a significant effect of \\ConditionID on tlx\\_mental (\\F{1.92190837641126}{44.2038926574591}{5.75}, \\p{0.007}, r=0.11). "
res$posthoc # significant pairwise comparisons (NULL for 2 groups)
#> [1] "A Student's t post-hoc test found that Baseline was significantly higher (\\m{54.92}, \\sd{11.10}) in terms of tlx\\_mental compared to HUD (\\m{45.88}, \\sd{9.76}); \\padj{0.033}). "
#> [2] "A Student's t post-hoc test found that Baseline was significantly higher (\\m{54.92}, \\sd{11.10}) in terms of tlx\\_mental compared to LED (\\m{45.21}, \\sd{12.51}); \\padj{0.015}). "For a whole questionnaire battery, report_all() does
this for every scale and adds a Holm-corrected summary across the
dependent variables:
battery <- report_all(
main_df,
dvs = c("tlx_mental", "trust"),
iv = "ConditionID",
design = "within",
labels = c(tlx_mental = "Mental Demand", trust = "Trust")
)
#> Shapiro--Wilk tests indicated no significant deviation from normality in any group (all $p \geq 0.05$); therefore, parametric tests were used.
#> A ANOVA estimation for factorial designs using 'afex' found a significant effect of \ConditionID on tlx\_mental (\F{1.92190837641126}{44.2038926574591}{5.75}, \p{0.007}, r=0.11).
#> A Student's t post-hoc test found that Baseline was significantly higher (\m{54.92}, \sd{11.10}) in terms of tlx\_mental compared to HUD (\m{45.88}, \sd{9.76}); \padj{0.033}).
#> A Student's t post-hoc test found that Baseline was significantly higher (\m{54.92}, \sd{11.10}) in terms of tlx\_mental compared to LED (\m{45.21}, \sd{12.51}); \padj{0.015}).
#> Shapiro--Wilk tests indicated a significant deviation from normality for at least one group (minimum $W = 0.79$, $p < 0.001$); therefore, non-parametric tests were used.
#> A Friedman rank sum test found a significant effect of \ConditionID on trust (\chisq(2)=10.93, \p{0.004}, r=0.23).
#> A Durbin-Conover post-hoc test found that HUD was significantly higher (\m{4.12}, \sd{1.08}) in terms of \trust compared to Baseline (\m{3.42}, \sd{1.10}); \padj{0.024}).
#> A Durbin-Conover post-hoc test found that LED was significantly higher (\m{4.46}, \sd{0.78}) in terms of \trust compared to Baseline (\m{3.42}, \sd{1.10}); \padj{0.003}).
battery$summary
#> dv method statistic
#> 1 tlx_mental ANOVA estimation for factorial designs using 'afex' 5.754671
#> 2 trust Friedman rank sum test 10.929577
#> p.value p.holm
#> 1 0.006577594 0.008466471
#> 2 0.004233235 0.008466471