How to make a scorecard (1-100) with different weights?

[danishussalam]

Hi everyone,

I have three variables each having a different scale

1) clubtime - indicating whether club activity started on time (1), started late (0.5) or if there was no club activity at all (0)
2) attendance - an integer value for the number of girls who attended the club on the day of the survey.
3) financial_literacy - a binary variable indicating whether the club as whole organized a financial literacy training or not.

I want to create a scorecard (1-100) by combining the above 3 variables such that clubtime and attendance each get 30% weight in the scorecard whereas financial_literacy gets 40% weight.

Any help in this regard will be highly appreciated.

Thanks
Danish

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte(survey_country ela_club) double clubtime byte(attendance financial_literacy)
1  2  0 20 0
2 11  1 11 1
1  1  0 20 0
2  9 .5  2 1
1  6  1  2 1
1  3 .5  3 0
1  5  1  6 1
1  4  0 20 0
2  7  1  8 1
2 10 .5  0 1
2 12 .5  4 1
2  8  1 11 1
3 15 .5  8 1
3 16 .5  8 1
3 14  1  4 1
3 13  1 16 1
end
label values survey_country survey_country
label def survey_country 1 "Uganda", modify
label def survey_country 2 "Tanzania", modify
label def survey_country 3 "Liberia", modify
label values ela_club ela_club
label def ela_club 1 "Masaanaa", modify
label def ela_club 2 "Muzana 1", modify
label def ela_club 3 "Bukejje", modify
label def ela_club 4 "Kikajjo", modify
label def ela_club 5 "Kampala Road", modify
label def ela_club 6 "Kirombe A", modify
label def ela_club 7 "Nsalaga", modify
label def ela_club 8 "Chemchan", modify
label def ela_club 9 "Chadurua", modify
label def ela_club 10 "Ipagala A", modify
label def ela_club 11 "Sandali", modify
label def ela_club 12 "Mamboleo", modify
label def ela_club 13 "Focus Girls", modify
label def ela_club 14 "Faith Girls", modify
label def ela_club 15 "Bright Future", modify
label def ela_club 16 "Opportunity", modify
label values financial_literacy financial_literacy
label def financial_literacy 0 "No", modify
label def financial_literacy 1 "Yes", modify

[Andrew Musau]

You have to specify the scoring rules for each of your variables, e.g.,

clubtime - indicating whether club activity started on time (1), started late (0.5) or if there was no club activity at all (0)

no activity=0/30, started late= 15/30, started on time=30/30

[danishussalam]

Hi Andrew Musau

Thanks for the prompt response. I'm not sure if I get what you're asking.

- attendance is a numeric value and can take any value from 0 to 30. Should I create my own rules here? For example if attendance is from 0-10 the score is = 0, if 10-20 = 0.5, and anything > 20 gets a score of 1

- Financial_literacy is a binary variable taking the value of 1 if they had the training, 0 otherwise.

Thanks.

[William Lisowski]

You want to rescale clubtime to 0-30, attendance to 1-30 (well that's done since by coincidence the largest possible value of attendance is 30), and financial_literacy to 0-40. Then you add those together for your score. The value of the sum will run from 1 to 100, and clubtime and attendance will each amount to 30% and financial_literacy to 40%.

That means financial_literacy is 0 or 40, and clubtime is 0, 15, or 30.

[Andrew Musau]

You can rescale as William suggests or simply do the following

gen score= (30*clubtime)+ attendance +(40*financial_literacy)

It makes sense to treat attendance as a score if for each of the clubs, the maximum number of possible attendees is 30. Otherwise, your scoring rule will be unfairly punishing smaller clubs.