Closed MindCartographer closed 3 years ago
Hi José, thanks for the suggestions. I have a question with regard to the calculation, however. Polarization is the proportion of "extreme" ratings for the whole grid, per construct or per element. So when we calculate the average of the polarization values for constructs or elements, shouldn't we just get the same results as the total polarization?
Let's take the boeker
grid as an example:
> l <- indexPolarization(boeker)
l
POLARIZATION OVERALL
Ratings Extreme Polarization Constructs_Avg Elements_Avg
1 210 26 12.4% 12.4% 12.4%
POLARIZATION BY CONSTRUCT
Construct Ratings Extremes Polarization
1 balanced - get along with conflicts 15 2 13.3%
2 isolated - sociable 15 1 6.7%
3 closely integrated - excluded 15 0 0.0%
4 discursive - passive 15 2 13.3%
5 open minded - indifferent 15 3 20.0%
6 dreamy - dispassionate 15 1 6.7%
7 practically oriented - depressed 15 1 6.7%
8 playful - serious 15 0 0.0%
9 socially minded - selfish 15 3 20.0%
10 quarrelsome - peaceful 15 2 13.3%
11 artistic - technical 15 1 6.7%
12 scientific - emotional 15 3 20.0%
13 introvert - extrovert 15 1 6.7%
14 wanderlust - home oriented 15 6 40.0%
POLARIZATION BY ELEMENT
Element Ratings Extremes Polarization
1 self 14 2 14.3%
2 ideal self 14 8 57.1%
3 mother 14 0 0.0%
4 father 14 3 21.4%
5 kurt 14 0 0.0%
6 karl 14 0 0.0%
7 george 14 0 0.0%
8 martin 14 1 7.1%
9 elizabeth 14 0 0.0%
10 therapist 14 1 7.1%
11 irene 14 7 50.0%
12 childhood self 14 2 14.3%
13 self before illness 14 1 7.1%
14 self with delusion 14 0 0.0%
15 self as dreamer 14 1 7.1%
The total and means of the constructs and elements are then ...
> l$polarization_total$Polarization
[1] 0.1238095
> mean(l$polarization_constructs$Polarization)
[1] 0.1238095
> mean(l$polarization_elements$Polarization)
[1] 0.1238095
Above, we get 12,4% in all cases. Could you please elaborate on how exactly you calculate the total polarization for constructs and elements in GRIDCOR, and outline what I am missing here?
Hi Mark,
Jajajajja You are absolutely right. And that is probably the reason why I finally did not send you that suggestion before. Sorry about that. In fact I see now that those indexes do not exist in GRIDCOR. I simply copied the notes I had collected months ago without double check nor thought. My mistake. Sorry.
To be honest… As I was writing the issue yesterday… I clearly saw that the calculation of the Total Polarization for all constructs (calculated as the count of all extreme values in all constructs divided by the total number of scores) would be the same as the Total Polarization for all elements (calculated as the count of all extreme values in all elements divided by the total number of scores), and also equal to the Total Polarization of the Grid, but whatever the reason I thought that would be different to the average of the construct’s / element’s polarizations.
So please forget about the suggestion and just keep the cosmetic indication.
Thanks, Mark
José
Hi Mark,
Just a cosmetic correction and a suggestion for improvement regarding...
indexPolarization
In the documentation for this function we read:
the same line is repeated three times ("polarization_total: Grid's overall polarization"); last 2 lines should be replaced by:
Polarization_constructs:
Constructs' polarizationsPolarization_elements:
Elemnents' polarizationsBesides this minor correction in the documentation, I'd like to suggest the addition of 2 new results for this function (both of them are already included in GRIDCOR):
Polarization_Total_Constructs:
average of all constructs polarization valuesPolarization_Total_Elements:
average of all elements polarization valuesSo far, the
indexPolarization
provides just the Total Polarization for the entire grid, and the individual polarizations for each construct and element.These 2 new summary polarization values I propose to be added (one for all constructs, and the other one for all elements) provide a usefull overview that will enrich the output of this function.
Thanks in advance, BR José