Module stikpetP.other.thumb_cliff_delta

Functions

def th_cliff_delta(d, qual='romano')

Rules Of Thumb For Cliff Delta

This function will give a qualification (classification) for Cliff Delta

Parameters

d : float

the Cliff Delta value

qual : {"romano", "metsamuuronen"} optional

the rule of thumb to be used. Default is "romano"

Returns

pandas.DataFrame

A dataframe with the following columns:

• classification, the qualification of the effect size
• reference, a reference for the rule of thumb used

Notes

The following rules-of-thumb can be used:

"romano" => Uses Romano et al. (2006, p. 14):

|d|	Interpretation
0.00 < 0.15	negligible
0.15 < 0.33	small
0.33 < 0.47	medium
0.47 or more	large

"metsamuuronen" => Metsämuuronen (2023, p. 17):

|d|	Interpretation
0.00 < 0.11	negligible
0.11 < 0.28	small
0.28 < 0.43	medium
0.43 or more	large

Before, After and Alternatives

Cliff delta could be converted to Cohen d (Marfo & Okyere, 2019, p. 4) or Vargha-Delaney A ( * es_convert to convert the effect size measure

References

Marfo, P., & Okyere, G. A. (2019). The accuracy of effect-size estimates under normals and contaminated normals in meta-analysis. Heliyon, 5(6), e01838. doi:10.1016/j.heliyon.2019.e01838

Metsämuuronen, J. (2023). Somers’ delta as a basis for nonparametric effect sizes: Grissom-Kim PS, Cliff’s d, and Vargha-Delaney A as specific cases of Somers delta. doi:10.13140/RG.2.2.36002.09925

Romano, J., Kromrey, J. D., Coraggio, J., Skowronek, J., & Devine, L. (2006). Exploring methods for evaluating group differences on the NSSE and other surveys: Are the t-test and Cohen’s d indices the most appropriate choices? 1–51.

Author

Made by P. Stikker

Companion website: https://PeterStatistics.com
YouTube channel: https://www.youtube.com/stikpet
Donations: https://www.patreon.com/bePatron?u=19398076

Expand source code

def th_cliff_delta(d, qual="romano"):
    '''
    Rules of Thumb for Cliff Delta
    --------------------------
     
    This function will give a qualification (classification) for Cliff Delta
    
    Parameters
    ----------
    d : float
        the Cliff Delta value
    qual : {"romano", "metsamuuronen"} optional 
        the rule of thumb to be used. Default is "romano"
        
    Returns
    -------
    pandas.DataFrame
        A dataframe with the following columns:
    
        * *classification*, the qualification of the effect size
        * *reference*, a reference for the rule of thumb used
   
    Notes
    -----
    The following rules-of-thumb can be used:
    
    *"romano"* => Uses Romano et al. (2006, p. 14):
    
    |\\|d\\|| Interpretation|
    |---|----------|
    |0.00 < 0.15 | negligible |
    |0.15 < 0.33 | small |
    |0.33 < 0.47 | medium |
    |0.47 or more | large |
    
    *"metsamuuronen"* => Metsämuuronen (2023, p. 17):
    
    |\\|d\\|| Interpretation|
    |---|----------|
    |0.00 < 0.11 | negligible |
    |0.11 < 0.28 | small |
    |0.28 < 0.43 | medium |
    |0.43 or more | large |
        
    Before, After and Alternatives
    ------------------------------
    Cliff delta could be converted to Cohen d (Marfo & Okyere, 2019, p. 4) or Vargha-Delaney A (
    * [es_convert](../effect_sizes/convert_es.html#es_convert) to convert the effect size measure
        
    References
    ----------
    Marfo, P., & Okyere, G. A. (2019). The accuracy of effect-size estimates under normals and contaminated normals in meta-analysis. *Heliyon, 5*(6), e01838. doi:10.1016/j.heliyon.2019.e01838
    
    Metsämuuronen, J. (2023). Somers’ delta as a basis for nonparametric effect sizes: Grissom-Kim PS, Cliff’s d, and Vargha-Delaney A as specific cases of Somers delta. doi:10.13140/RG.2.2.36002.09925
    
    Romano, J., Kromrey, J. D., Coraggio, J., Skowronek, J., & Devine, L. (2006). Exploring methods for evaluating group differences on the NSSE and other surveys: Are the t-test and Cohen’s d indices the most appropriate choices? 1–51.
    
    Author
    ------
    Made by P. Stikker
    
    Companion website: https://PeterStatistics.com  
    YouTube channel: https://www.youtube.com/stikpet  
    Donations: https://www.patreon.com/bePatron?u=19398076
    
    '''
            
    #Romano et al. (2006, p. 14).
    if (qual=="romano"):
        ref = "Romano et al. (2006, p. 14)."
        if (abs(d) < 0.15):
            qual = "negligible"
        elif (abs(d) < 0.33):
            qual = "small"
        elif (abs(d) < 0.47):
            qual = "medium"
        else:
            qual = "large"
            
    #Metsämuuronen (2023, p. 17).
    elif (qual=="metsamuuronen"):
        ref = "Metsämuuronen (2023, p. 17)"
        if (abs(d) < 0.11):
            qual = "negligible"
        elif (abs(d) < 0.28):
            qual = "small"
        elif (abs(d) < 0.43):
            qual = "medium"
        else:
            qual = "large"
    
    results = pd.DataFrame([[qual, ref]], columns=["classification", "reference"])
    
    return results