Skip to content

2.4beta

Pre-release
Pre-release
Compare
Choose a tag to compare
@krajcsi krajcsi released this 21 Apr 14:56
· 185 commits to master since this release

2.4 beta (21 April 2023)

New features

  • Data handling
    • New data view to see the data together with the results (thanks to Belma Bumin)
    • Reload actual data file
    • Multivariate outlier filtering with Mahalanobis distance (Tamás Szűcs)
    • New demo data files https://learningstatisticswithcogstat.com/ (Róbert Fodor)
  • Ability to rerun the analyses in the Results pane
  • Multiple linear regression analysis (Tamás Szűcs)
    • Model fit metrics
    • Partial correlation
    • Partial regression plots
    • Scatterplot matrix of raw data
    • Analysis of multicollinearity
    • Hypothesis tests
  • Reliability analyses (Tamás Szűcs)
    • Internal consistency reliability analysis
      • Item-total scatter plots
      • Cronbach's alpha with and without items and their CIs
      • Item-rest correlation and their CIs
    • Interrater reliability analysis
      • Chart showing scores from different raters
      • ICC values and their CIs
      • Assumption checks for inferential statistics
      • Hypothesis tests whether ICC is 0
  • Displaying groups and factors
    • In comparing groups, display groups not only on x-axes but also with colors or in panels
    • In comparing repeated measures variables, display conditions not only on x-axes but also with colors
      • Rearrange the factors flexibly
    • For ordinal repeated measures variables, display the rank of the values
  • Comparing variables and groups in mixed design
    • Raw data
    • Descriptives and related charts
    • Parameter estimations and related charts
  • Behavioral data diffusion analysis
    • The time unit (sec or msec), error coding (1 or 0), and scaling parameter (0.1 or 1) can be set
    • Slow trials are filtered before the analysis is run
    • Display the number of filtered (missing and slow outlier) trials
    • Number of included trials per conditions are displayed
  • Output handling
    • Save results into html file instead of pdf file (Róbert Fodor)
    • Ability to use png or svg image formats for charts
  • Possibility to print detailed Python error messages to results pane
  • New localization
    • Chinese (Xiaomeng Zhu)
    • Malay (Nur Hidayati Miza binti Junaidi)
    • Arabic (Rahmeh Albursan)

Fixes

  • ⚠️ In outlier filtering, the cases with the limit value will be included and not excluded
  • ⚠️ With the update of the scipy module, the p values of the Wilcoxon tests are fixed
  • Extended calculation validations (thanks to Eszter Miklós)
  • Most settings in Preferences are applied without the need to restart
  • Various GUI, and output fixes