New 2.4 CogStat version

[Krajcsi Attila]

Dear All,

We are happy to announce that the latest CogStat release is available. There are tons of improvements. See the detailed list of new features below.

You can download the new version from our website: https://www.cogstat.org/

Enjoy the new release! Let us know if you have any feedback.

Attila

CogStat 2.4 (September 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)
  • Scatterplot matrix of raw data
  • Linear regression function
  • Scatterplot with regression line
  • Partial regression plots with regression lines
  • Model fit metrics
  • Partial correlations
  • Residual plot and histogram of residuals
  • Assumptions of inferential statistics
    • Multivariate normality
    • Homoscedasticity
    • Analysis of multicollinearity
  • Population parameter point and interval estimations (including standardized effect sizes)
  • 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 (experimental svg support)
• Possibility to print detailed Python error messages to results pane
• New localization
  • Chinese (Xiaomeng Zhu)
  • Malay (Nur Hidayati Miza binti Junaidi)
  • Arabic (Rahmeh Albursan)
• Python package
  • Pandas DataFrames with MultiIndex columns can be imported
  • Diffusion analysis results are returned as pandas Stylers

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