Table with the Percentage of correct Answers

[Phi Dil]

Hello everyone,

I just started working with stata, so my knowledge is very basic.
My dataset includes several questions that respondents could answer on a scale from 1 - 4. The intention of these questions was to test the knowledge of respondents, so there is one correct answer, two incorrect answers and a category for "do not know".

I want to create a table that shows the percentage of correct, incorrect and do not know answers.

I want it to look like this:

	Question1	Question2	Question3
Correct	40%	50%	60%
Incorrect	40%	30%	30%
Do not know	20%	20%	10%

How do I do this? I know that I have to generate new variables for Correct, Incorrect and Do not know. I already thought of categorical variables which I generated like this for the first question, assume that option 1 was the correct answer:

gen Question1_1 = .
replace Question1_1 = 1 if Question1==1
replace Question1_1 = 2 if Question1 >1 & <4
replace Question1_1 = 3 if Question1==4

This more or less gives me what I want for one question, but I have no idea how I can get this for all of my questions to show up in one table.

Maybe anyone can help me out.

Thank you in advance,

Phi Dil

[Mike Murphy]

Try this:

ssc install tab_chi // installs the tabm command if you don't have it
tabm Question_1 Question_2 ... Question_N, row transpose

[Marcos Almeida]

Please read this FAQ, since it will probably be helpful to you.

[Phi Dil]

Thanks for your responses. I will check if that works for me.

[Phi Dil]

So the FAQ was helpful if I only had two variables I needed to match. But my problem occurs since I want to have an variable as an identifier for correct answers to 9 questions. How do I do that? I keep confusing myself with this.

[Mike Lacy]

Let's suppose that the codes for the correct answers are, in order from Question_1 to Question_9, "143112443". The variable with the correct answer is:

gen str9 answercodes = "143112443"

If you wanted these answers as numeric, and in separate variables, just do

gen byte answer1 = 1
gen byte answer2 = 3
.... etc.

A better solution to your situation would involved storing the correct answer codes in a local macro, rather than a variable, but that's not something that's likely to be useful to you unless you have some programming experience.

[Phi Dil]

Let me try to explain it differently: assume you have three products: Shoes, Shirts, Pants, each can be in one of the following colors: Red, Green, Blue

How do I make a table that shows me how big the share of each color for each product is?

	Shoes	Shirts	Pants
Red	60%	30%	20%
Green	30%	50%	20%
Blue	10%	20%	60%

I hope that this better explains my problem.

Again, thank you all in advance.

[Rich Goldstein]

unless I'm completely confused, the basis for what you want was given above in #2; first the data from #7 but using -dataex-:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(shoes shirts pants)
1 1 1
1 1 1
1 1 2
1 2 2
1 2 3
1 2 3
2 2 3
2 2 3
2 3 3
3 3 3
end
label values shoes color
label values shirts color
label values pants color
label def color 1 "red", modify
label def color 2 "green", modify
label def color 3 "blue", modify

now, after installing tabs as per #2 above, the result:

. tabm shoes shirts pants, col transpose nofreq nokey

           |             variable
    values |     shoes     shirts      pants |     Total
-----------+---------------------------------+----------
       red |     60.00      30.00      20.00 |     36.67 
     green |     30.00      50.00      20.00 |     33.33 
      blue |     10.00      20.00      60.00 |     30.00 
-----------+---------------------------------+----------
     Total |    100.00     100.00     100.00 |    100.00

[Phi Dil]

Thank you Rich! That table looks exactly like I want it to look!

I assume there is no way around tabm? When I try to run

ssc install tab_chi

Stata tells me that the file next.trk already exists, but still I get the error that tabm is unrecognized.

[Nick Cox]

Naturally there are alternatives to tabm, such as your writing equivalent code yourself.

The problem with next.trk can be worked around by direct installation. Run

adopath

and see what adopath calls PLUS. On my current machine it is

c:\ado\plus

while on your machine it may well be different. Go to that directory or folder with some variant on

cd "c:\ado\plus"  

and then go to the subdirectory where program files whose names begin with t will be installed.

cd t

If Stata complains that this is not possible, then

mkdir t

and try again. Now you can

ssc copy tabm.ado 
ssc copy tabm.sthlp

Now you can back to where you were and try tabm.

If the problem with next.trk persists, you may need to ask again for advice.