nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

How do I do correlation on dissimilar data (Spearman's rho)

Jim101

Hi!

I am working on a faculty project that is quite complex. I have gotten all the usual statistics (mean, median, mode, SD, etc.) but need to do a correlation on dissimilar data.

The project is one on users of online versus face-to-face students; demographics and survey question (Likert scale) answers.

One example is, do people who have taken many online classes tend to prefer a pre-course orientation?

Another example is, do people who prefer face-to-face classes tend to prefer working in group activities?

I have researched some papers and I notice that Spearman's rho was quoted quite often.

I have attached an Excel file with representative information.

I don't know how to set this up; interactive explanation preferred.

Thanks!

Jim101 File.xlsx

Accepted answers

Jim101

I will update and run it this weekend. Thanks so much!

Jim101

All comments

txnelson

Do you have the individual responses? Your summary table isn't going to give you answerw to the questions you are asking.

Jim101

I do have the individual responses in an Excel worksheet. They were listed as text, such as "highly effective", "effective", "somewhat effective" and "not effective." These were converted to numbers and that gave the result of averages and such.

For example, in one question, I have the individual responses to "Expectations for assignments were.." with the four categories of answers.

Jim

txnelson

1. I have attached a JMP data table from the SAMPLE DATA provided with JMP.

Help==>Sample Data==>Customer Preferences.jmp

2. Your individual data needs to be setup in a form similar to it's structure

3. You data should be entered as you indicate in you Excel spreadsheet as

1=Highly Effectiver

2=Effective

3=Somewhat Efffective

4=Not Effective

Below is a display from the Consumer Preferences data table, showing the "Brush" data. The data are entered into the data table as numeric values, and then a "Value Labels" is setup to have the values displayed as "Almost Every Day", "Every Day"....etc.

Once the data are setup in this form, all of your Spearman correlations, etc. can be easily determined.

I suggest that you brush up on how to do this in JMP by reading the the JMP book, "Consumer Research"

Help==>Books==>Consumer Research

Consumer Preferences.jmp

txnelson

@Jim101

I am responding to you via the discussion forum rather than a direct response to your email. Your question, should really have been made through the Discussion Forum, because your question is a good question, and others need to see the question and the responses to the question.

Also, the direct messages do not have the ability to attach items, or to format code etc.

So, regarding your issue of having an Excel spreadsheet that you are having a laborious time working with, I suggest you give up on it. Just load the spreadsheet into JMP.....JMP can do the recoding much easier than Excel........given how you described the issue. Below is a simple script that generates a sample table in the format that I believe you stated your data looks like.

Names Default To Here( 1 );
dt = New Table( "bunch of data" );
dt << add rows( 500 );
theList = {"Not Helpful", "Not Very Helpful", "Helpful", "Very Helpful"};
For( i = 1, i <= 500, i++,
	dt << New Column( "Question " || Char( i ),
		character,
		formula( theList[Random Integer( 1, 4 )] )
	);
	Wait( 0 );
	
);
dt << delete columns( "column 1" );
For( i = 1, i <= 500, i++,
	Column( dt, i ) << delete formula
);

Run this script, and it will create a data table for you.

Now your request was how do you convert to 1,2,3 and 4 for the given character values. The script below can be run against the sample data table, and it will end up with numeric values of 1,2,3 and 4, for the character input.

Names Default To Here( 1 );
dt = Current Data Table();
For( i = 2, i <= N Cols( dt ), i++,
	Column( i )[dt << get rows where( As Column( i ) == "Not Helpful" )] = "1";
	Column( i )[dt << get rows where( As Column( i ) == "Not Very Helpful" )] = "2";
	Column( i )[dt << get rows where( As Column( i ) == "Helpful" )] = "3";
	Column( i )[dt << get rows where( As Column( i ) == "Very Helpful" )] = "4";
	Column( i ) << data type( numeric ) << modeling type( ordinal );
);

Jim101

I will update and run it this weekend. Thanks so much!

Jim101