easystats/parameters

Is devel-`{parameters}` out of sync with devel-`{effectsize}`?

IndrajeetPatil opened this issue · 10 comments

IndrajeetPatil commented

A number of tests related to CIs for htest objects seem to be failing:

══ Failed tests ════════════════════════════════════════════════════════════════
── Failure ('test-model_parameters.aov_es_ci.R:15'): model_parameters.aov ──────
na.omit(mp$Omega2_CI_low) (`actual`) not equal to es$CI_low (`expected`).

  `actual`: 0.29
`expected`: 0.31
── Failure ('test-model_parameters.aov_es_ci.R:16'): model_parameters.aov ──────
mp$Omega2_CI_low (`actual`) not equal to c(0.3122, NA) (`expected`).

  `actual`: 0.29 NA
`expected`: 0.31 NA
── Failure ('test-model_parameters.aov_es_ci.R:17'): model_parameters.aov ──────
na.omit(mp$Omega2_CI_high) (`actual`) not equal to es$CI_high (`expected`).

  `actual`: 0.5
`expected`: 1.0
── Failure ('test-model_parameters.aov_es_ci.R:18'): model_parameters.aov ──────
na.omit(mp$Omega2_CI_high) (`actual`) not equal to 1 (`expected`).

  `actual`: 0.5
`expected`: 1.0
── Failure ('test-model_parameters.aov_es_ci.R:30'): model_parameters.aov ──────
na.omit(mp$Eta2_CI_low) (`actual`) not equal to es$CI_low (`expected`).

  `actual[1:4]`: 0.54 0.00 0.00 0.00
`expected[1:4]`: 0.56 0.00 0.00 0.00
── Failure ('test-model_parameters.aov_es_ci.R:31'): model_parameters.aov ──────
mp$Eta2_CI_low (`actual`) not equal to c(0.5572, 0, 0, 0, 0, 0, 0, NA) (`expected`).

  `actual[1:4]`: 0.54 0.00 0.00 0.00
`expected[1:4]`: 0.56 0.00 0.00 0.00
── Failure ('test-model_parameters.aov_es_ci.R:32'): model_parameters.aov ──────
na.omit(mp$Eta2_CI_high) (`actual`) not equal to es$CI_high (`expected`).

  `actual`: 0.7 0.0 0.0 0.0 0.0 0.0 0.0
`expected`: 1.0 1.0 1.0 1.0 1.0 1.0 1.0
── Failure ('test-model_parameters.aov_es_ci.R:33'): model_parameters.aov ──────
na.omit(mp$Eta2_CI_high) (`actual`) not equal to rep(1, 7) (`expected`).

  `actual`: 0.7 0.0 0.0 0.0 0.0 0.0 0.0
`expected`: 1.0 1.0 1.0 1.0 1.0 1.0 1.0
── Failure ('test-model_parameters.aov_es_ci.R:50'): model_parameters.anova ────
na.omit(mp$Omega2_CI_low) (`actual`) not equal to es$CI_low (`expected`).

  `actual[1:4]`: 0.52 0.00 0.00 0.00
`expected[1:4]`: 0.54 0.00 0.00 0.00
── Failure ('test-model_parameters.aov_es_ci.R:51'): model_parameters.anova ────
na.omit(mp$Omega2_CI_high) (`actual`) not equal to es$CI_high (`expected`).

  `actual`: 0.7 0.0 0.0 0.0 0.0 0.0 0.0
`expected`: 1.0 1.0 1.0 1.0 1.0 1.0 1.0
── Failure ('test-model_parameters.aov_es_ci.R:69'): model_parameters.anova ────
na.omit(mp$Omega2_CI_low) (`actual`) not equal to es$CI_low[2] (`expected`).

  `actual`: 0.25
`expected`: 0.30
── Failure ('test-model_parameters.aov_es_ci.R:70'): model_parameters.anova ────
na.omit(mp$Omega2_CI_high) (`actual`) not equal to es$CI_high (`expected`).

  `actual`: 0.631129676481392  
`expected`:                 1 1
── Failure ('test-model_parameters.aov_es_ci.R:96'): model_parameters.car-anova ──
na.omit(mp$Omega2_CI_low) (`actual`) not equal to es$CI_low (`expected`).

  `actual`: 0.000 0.028 0.000
`expected`: 0.000 0.051 0.007
── Failure ('test-model_parameters.aov_es_ci.R:97'): model_parameters.car-anova ──
mp$Omega2_CI_low (`actual`) not equal to c(0, 0.0511, 0.00666, NA) (`expected`).

  `actual`: 0.000 0.028 0.000 NA
`expected`: 0.000 0.051 0.007 NA
── Failure ('test-model_parameters.aov_es_ci.R:98'): model_parameters.car-anova ──
na.omit(mp$Omega2_CI_high) (`actual`) not equal to es$CI_high (`expected`).

  `actual`: 0.0 0.3 0.3
`expected`: 1.0 1.0 1.0
── Failure ('test-model_parameters.aov_es_ci.R:99'): model_parameters.car-anova ──
na.omit(mp$Omega2_CI_high) (`actual`) not equal to rep(1, 3) (`expected`).

  `actual`: 0.0 0.3 0.3
`expected`: 1.0 1.0 1.0
── Failure ('test-model_parameters.aov_es_ci.R:121'): model_parameters.maov ────
na.omit(mp$Omega2_CI_low) (`actual`) not equal to es$CI_low (`expected`).

  `actual`: 0.54 0.71 0.51
`expected`: 0.58 0.74 0.55
── Failure ('test-model_parameters.aov_es_ci.R:122'): model_parameters.maov ────
mp$Omega2_CI_low (`actual`) not equal to c(0.58067, NA, 0.74092, NA, 0.55331, NA) (`expected`).

  `actual`: 0.54 NA 0.71 NA 0.51 NA
`expected`: 0.58 NA 0.74 NA 0.55 NA
── Failure ('test-model_parameters.aov_es_ci.R:123'): model_parameters.maov ────
na.omit(mp$Omega2_CI_high) (`actual`) not equal to es$CI_high (`expected`).

  `actual`: 0.8 0.9 0.8
`expected`: 1.0 1.0 1.0
── Failure ('test-model_parameters.aov_es_ci.R:124'): model_parameters.maov ────
na.omit(mp$Omega2_CI_high) (`actual`) not equal to rep(1, 3) (`expected`).

  `actual`: 0.8 0.9 0.8
`expected`: 1.0 1.0 1.0
── Failure ('test-model_parameters.aov_es_ci.R:154'): works with aov ───────────
`df_aov` (`actual`) not equal to structure(...) (`expected`).

actual vs expected
                Eta2_CI_high Omega2_CI_high
- actual[1, ]      0.6090780      0.4877583
+ expected[1, ]    1.0000000      1.0000000
- actual[2, ]      0.5274280      0.5036954
+ expected[2, ]    1.0000000      1.0000000
- actual[3, ]      0.2562071      0.0000000
+ expected[3, ]    1.0000000      1.0000000
- actual[4, ]      0.3038357      0.10[332](https://github.com/easystats/parameters/actions/runs/3977528585/jobs/6818741637#step:6:334)57
+ expected[4, ]    1.0000000      1.0000000
  actual[5, ]             NA             NA

  `actual[[9]]`: 0.6 0.5 0.3 0.3 NA
`expected[[9]]`: 1.0 1.0 1.0 1.0 NA

  `actual[[12]]`: 0.5 0.5 0.0 0.1 NA
`expected[[12]]`: 1.0 1.0 1.0 1.0 NA
── Failure ('test-model_parameters.aov_es_ci.R:321'): works with Gam ───────────
`df_Gam` (`actual`) not equal to structure(...) (`expected`).

actual vs expected
                Omega2_CI_low Omega2_CI_high
- actual[1, ]       0.6498208      0.7521348
+ expected[1, ]     0.706[340](https://github.com/easystats/parameters/actions/runs/3977528585/jobs/6818741637#step:6:342)0      1.0000000
- actual[2, ]       0.3205417      0.4915730
+ expected[2, ]     0.4106700      1.0000000
- actual[3, ]       0.0000000      0.0000000
+ expected[3, ]     0.0000000      1.0000000
  actual[4, ]              NA             NA

  `actual[[8]]`: 0.65 0.32 0.00 NA
`expected[[8]]`: 0.71 0.41 0.00 NA

  `actual[[9]]`: 0.8 0.5 0.0 NA
`expected[[9]]`: 1.0 1.0 1.0 NA
── Failure ('test-model_parameters.aov_es_ci.R:371'): works with anova ─────────
`df_car` (`actual`) not equal to structure(...) (`expected`).

actual vs expected
                Eta2_CI_high Epsilon2_CI_high
- actual[1, ]     0.06015808        0.0000000
+ expected[1, ]   1.00000000        1.0000000
- actual[2, ]     0.[345](https://github.com/easystats/parameters/actions/runs/3977528585/jobs/6818741637#step:6:347)00743        0.3267[400](https://github.com/easystats/parameters/actions/runs/3977528585/jobs/6818741637#step:6:402)
+ expected[2, ]   1.00000000        1.0000000
- actual[3, ]     0.27295622        0.25[450](https://github.com/easystats/parameters/actions/runs/3977528585/jobs/6818741637#step:6:452)05
+ expected[3, ]   1.00000000        1.0000000
  actual[4, ]             NA               NA

  `actual[[9]]`: 0.1 0.3 0.3 NA
`expected[[9]]`: 1.0 1.0 1.0 NA

  `actual[[12]]`: 0.0 0.3 0.3 NA
`expected[[12]]`: 1.0 1.0 1.0 NA

[ FAIL 23 | WARN 0 | SKIP 7 | PASS 1269 ]
mattansb commented

Has the default CI type in parameters for ANOVA effect sizes changed? It seems to be two-sided, but the tests suggest that it use to be one sided...

strengejacke commented

It's NULL by default.

mattansb commented

Ah... Should be fixed in easystats/effectsize#555

strengejacke commented

Ok. I also set alternative = "greater" explicitly to the tests.

  • 👍1
strengejacke commented

Do we no longer have partial effect size types?

mattansb commented

We definitely do 😅

strengejacke commented

How do I calculate partial effect sizes then?

mattansb commented

Default is partial. If you want non-partial, set partial = FALSE.

strengejacke commented

Ah, just saw it's also documented :-)