nomo_table() returns compact tibbles intended for manuscripts, audit
appendices, teaching materials, and nomo_report(). It never rounds away
the underlying object: full engine results remain stored in the original
nomologR object.
Value
A tibble. A table the object has no rows for, such as "partial"
for an invariance analysis without releases or "replication" for a
network without validation data, has no rows and the table's usual
columns.
Details
Supported objects and type values. The first value listed is the default.
nomo_screen:"items","distribution","cases","relationships","effort","decision_log"; seenomo_screen().nomo_factors:"evidence","criteria","adequacy","concordance","decision_log"; seenomo_factors().nomo_efa:"items","pattern","factor_correlations","residuals","decision_log". The pattern and factor-correlation matrices are returned as tables with the item or factor in the first column; seenomo_efa().nomo_cfa:"fit","loadings","factor_correlations","heywood","residuals","modification_indices","decision_log"; seenomo_cfa().nomo_reliability:"coefficients","alpha_status","ci_status","decision_log"; seenomo_reliability().nomo_validity:"convergent","discriminant","htmt_status","decision_log"."discriminant"has one row per construct pair, with the latent correlation and the HTMT-family values together; seenomo_validity().nomo_scores:"scores","diagnostics","unit_weighting","notes"; seenomo_scores().nomo_power:"power"fromnomo_power_rmsea();"summary"(default) and"parameters"fromnomo_power_simulate().nomo_hypotheses: the machine-readable hypothesis table (notype; anytypegiven is an error).nomo_network:"hypotheses"(default),"fit","measurement","relations","replication","single_indicators","sensitivity","decision_log"."single_indicators"and"sensitivity"describe the composites modeled as single indicators; seenomo_network().nomo_retest:"icc"(default),"reliable_change","decision_log"; seenomo_retest().nomo_esem:"loadings"(default),"factor_correlations","models","comparisons","decision_log"; seenomo_esem().nomo_method_variance:"comparisons"(default),"models","loadings","reliability","correlations","decision_log"; seenomo_method_variance().nomo_invarianceandnomo_invariance_longitudinal:"fit"(default),"categories","partial","local_strain","latent_means","decision_log". The local-strain table keeps lavaan's internalconstraintlabel and adds a human-readableconstraint_displaycolumn (for example,Intercept: ag3 (online vs. paper)).nomo_compare:"comparisons"(default),"models","loadings","evidence","decision_log"; seenomo_compare().nomo_missing:"strategies"(default),"estimates","fit","reliability","pattern","variables","decision_log"; seenomo_missing().nomo_hierarchical:"indices"(default),"subscales","factors","loadings","notes","decision_log"; seenomo_hierarchical().nomo_run:"stages"(default),"requests","decisions","component_log","scales","recipe","settings","lineage","methods"; seenomo_run()andnomo_revise()."lineage"returns one row per revision, or no rows for a workflow that was never revised."methods"returns the registry entries for the methods the run actually used, and is equivalent tonomo_methods(x); seenomo_methods().
Fit tables
Fit tables name their indices in one of two ways, a known difference that
is kept as it is (see Conventions in returned tables in ?nomologR).
The nomo_cfa and nomo_missing tables spell them chi_square,
p_value, CFI, and so on. The others use the names of
lavaan::fitMeasures(): chisq, pvalue, cfi, and so on.
nomo_cfa,"fit": long by design, one row per index, named inmetric(chi_square,df,p_value,CFI,TLI,RMSEA,RMSEA_CI_lower,RMSEA_CI_upper,SRMR), with itsvalue,reference, andattentionflag.nomo_missing,"fit": one row perstrategy, withchi_square,df,p_value,CFI,TLI,RMSEA, andSRMR.nomo_esem,"models": one row permodel, withchisq,df,pvalue,cfi,tli,rmsea,srmr,aic, andbic.nomo_method_variance,"models": one row permodel, withchisq,df,pvalue,cfi,tli,rmsea, andsrmr.nomo_compare,"models": one row permodel, withnpar,df,chisq,cfi,tli,rmsea,srmr,aic, andbic. The difference tests and their p values are in"comparisons".nomo_invariance,"fit": one row perlevel, withchisq,df,pvalue,cfi,rmsea, andsrmr, the changes from the level before (delta_cfi,delta_rmsea,delta_srmr), and the likelihood-ratio test against it (lrt_chisq,lrt_df,lrt_p).nomo_network,"fit": one row, withchisq,df,pvalue,cfi,tli,rmsea, andsrmr.
Examples
h <- nomo_hypotheses(
"Agency -> Persistence" = positive(min = .20),
"Agency <-> SocialDesirability" = negligible(within = c(-.15, .15))
)
nomo_table(h)
#> # A tibble: 2 × 16
#> id relation source target operator relation_type prediction lower upper
#> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <dbl> <dbl>
#> 1 H1 Agency -> … Agency Persi… -> directed positive 0.2 Inf
#> 2 H2 Agency <->… Agency Socia… <-> association negligible -0.15 0.15
#> # ℹ 7 more variables: lower_inclusive <lgl>, upper_inclusive <lgl>,
#> # region <chr>, scale <chr>, origin <chr>, magnitude_specified <lgl>,
#> # confirmable <lgl>
inv <- nomo_invariance(
"Agency =~ ag1 + ag2 + ag3 + ag4",
data = nomo_demo_network,
group = "group",
levels = c("configural", "metric", "scalar")
)
nomo_table(inv, "fit")
#> # A tibble: 3 × 19
#> level constraints partial_requested status converged chisq df pvalue
#> <chr> <chr> <chr> <chr> <lgl> <dbl> <dbl> <dbl>
#> 1 configur… none "" estim… TRUE 0.327 4 9.88e- 1
#> 2 metric loadings "" estim… TRUE 5.71 7 5.75e- 1
#> 3 scalar loadings, … "" estim… TRUE 68.9 10 7.12e-11
#> # ℹ 11 more variables: cfi <dbl>, rmsea <dbl>, srmr <dbl>, delta_cfi <dbl>,
#> # delta_rmsea <dbl>, delta_srmr <dbl>, lrt_chisq <dbl>, lrt_df <dbl>,
#> # lrt_p <dbl>, warnings <chr>, error <chr>
head(nomo_table(inv, "local_strain"))
#> # A tibble: 6 × 8
#> level constraint_index constraint score_x2 df p_value diagnostic_only
#> <chr> <int> <chr> <dbl> <dbl> <dbl> <lgl>
#> 1 metric 3 .p3. == .p17. 4.92 1 2.66e- 2 TRUE
#> 2 metric 2 .p2. == .p16. 1.03 1 3.10e- 1 TRUE
#> 3 metric 4 .p4. == .p18. 0.690 1 4.06e- 1 TRUE
#> 4 metric 1 .p1. == .p15. 0.0273 1 8.69e- 1 TRUE
#> 5 scalar 7 .p7. == .p21. 61.1 1 5.33e-15 TRUE
#> 6 scalar 5 .p5. == .p19. 11.3 1 7.68e- 4 TRUE
#> # ℹ 1 more variable: constraint_display <chr>