| T O P I C R E V I E W |
| TeoTiGrand |
Posted - 05/28/2026 : 3:20:44 PM
In the 2018 version, the Kruskal-Wallis ANOVA produces a result, whereas in the 2021 version it returns an error: kwanova:X-Function failed to execute!
Unknown error; function return error code: -11
Logically, why not just carry on working in the 2018 version, right? Firstly, I’m curious as to why it isn’t working and how to fix it. Secondly, I need the Dunn’s Test.
I am analysing data for the open field test. For the ‘number of peripheral squares’ category, the distances covered by the rat are:
K 54
K 93
K 59
K 88
K 92
KKp 125
KKp 71
KKp 34
KKp 85
KKp 40
KSD 26
KSD 56
KSD 129
KSD 66
KSD 115
KpSD 11
KpSD 61
KpSD 58
KpSD 83
KpSD 69
Please help!
:))) |
| 2 L A T E S T R E P L I E S (Newest First) |
| NadeeshaSupport |
Posted - 06/01/2026 : 09:44:19 AM
Hi TeoTiGrand,
I was able to run the analysis in Origin 2026. Please see attached. You may need to upgrade the software to the latest version.
Thanks,
Nadeesha
https://my.originlab.com/ftp/forum_and_kbase/Images/TeoTiGrand%20forum%20question.opju |
| MedAnnotate |
Posted - 06/01/2026 : 04:30:32 AM
1. Deconstructing the kwanova Error Code -11The error message kwanova:X-Function failed to execute! Unknown error; function return error code: -11 in Origin 2021 points directly to a data structure or variable indexing mismatch within the compiled Nag/C++ background libraries.While Origin 2018 used an older, less restrictive internal structure parsing system, Origin 2021 modernized its statistical dialog boxes (including the kwanova X-Function) to automatically check for strict indexing alignments to support features like post-hoc testing (Dunn’s) natively.Why it breaks in Origin 2021 (and how to fix it):The error code -11 typically occurs due to one of three formatting issues:Improper Data Layout Selection: You are trying to run the tool in Indexed mode while your data is structured in Raw/Sequential columns (or vice versa). Your dataset consists of two distinct components: a categorical grouping variable (the 4 rat groups: K, KKp, KSD, KpSD) and a continuous measurement variable (distance).Text in a Numeric Column: Origin 2021 strictly checks the column designations. If your data columns are combined as strings (e.g., cell content reads exactly "K 54"), the X-Function fails because it cannot compute ranks on non-numeric strings.Trailing Whitespace or Group Level Mismatch: Hidden spaces in group designations can confuse the new indexer.The Solution: Correct Data FormattingTo resolve this error and make your data readable by both Origin 2021 and any subsequent post-hoc tests, restructure your worksheet into exactly two columns:Column A (Set as Group/Factor): Contains the group names (K, KKp, KSD, KpSD).Column B (Set as Y / Data): Contains the raw distance values.Column A (Group)Column B (Distance)K54K93......KKp125KSD26To run the analysis correctly:Highlight your restructured data columns.Navigate to Statistics > Nonparametric Tests > Kruskal-Wallis ANOVA.In the dialog box, set Input Data Form to Indexed.Set Group Range to Column A, and Data Range to Column B.2. Why Not Just Stick to the 2018 Version?It is tempting to stay with an older build when an upgrade breaks a script, but doing so compromises your analysis for two main reasons:Native Dunn's Test Support: Origin 2018 did not natively incorporate Dunn's Test as an unbundled, automated post-hoc addition inside the standard Kruskal-Wallis user interface. To run Dunn's post-hoc in 2018, users frequently had to rely on separate LabTalk scripts, Origin C extensions, or manual calculation. Origin 2021 handles this directly within the GUI once data indexing is satisfied.Missing Values Handling Bug: Versions of Origin prior to modern revisions handled missing data point combinations across groupings inconsistently, occasionally throwing off the tie-correction factor ($\sum (t^3 - t)$) applied to the Chi-Square statistic ($H$).3. Step-by-Step Execution: Dunn's Test in Origin 2021Your experimental setup for the open field test evaluates 4 distinct treatment conditions (Groups = 4, $N = 20$, $n = 5$ per group). Because you have multiple comparison groups, the Kruskal-Wallis test will only provide a global $p$-value telling you if at least one group stochastically dominates another. You need Dunn's Test to pinpoint exactly which pairs differ. Once your data is formatted into Indexed Columns (as shown above), follow these precise steps:Open the Kruskal-Wallis ANOVA dialog (kwanova).Look at the bottom of the dialog panel for the Significance Level / Post Hoc tab (this may be nested under an advanced expander triangle depending on your patch version).Check the box labeled Dunn's Test (you can also select Dunn-Bonferroni if you require conservative family-wise error rate adjustments for multiple pairwise comparisons).Click OK.What to Look for in the Output Sheet:Origin will generate a new report sheet containing:The Ranks Table: Validating the mean rank of your metrics across groups (K, KKp, KSD, KpSD).Kruskal-Wallis ANOVA Table: Giving your Chi-Square ($H$ statistic) and overall asymptotic $p$-value.Dunn's Test Multi-Comparison Table: Look at the Sig column or the Adjusted P-Value column. If the value is $< 0.05$, the distance variation in peripheral square movement between those two specific rat groups is statistically significant.
Medannotate |
|
|
