Converting Values (Access 2k)

Topic

New Reply

Don’t know if this is an Appropriate place to ask but here goes.
I have a survey where the answers for 8 questions range from 1 to 10.
The survey has been around for a long time and it works great.
Our company standard in Japan is a scale from 1 to 4.
I’m looking for ideas on how to convert the 1 to 10 into 1 to 4 values
and try to keep as close to the original relationship as possible.
Thanks,
Scott

Reply | Quote

Replies

You haven’t really provided enough detail to get any useful suggestions. I assume 1 through 10 represent a range of values, such as from Highly Satisfied to Highly Dissatisfied, right? Are you asking for help figuring out what the categories from 1 to 4 should be, how to adapt the existing survey to a different range of values, or what?

Reply | Quote

Charlotte,
I’m looking for a way to convert the existing survey results to a different range of values(1 to 4)
The existing survey with the 1 to 10 values won’t be changed since customers are familiar with it.
Hope this helps explain what I want to do.
Scott

Reply | Quote

Are you trying to report on existing results in a different scale (ratio 2.5:1)? Is this for a report, a graph or what?

Reply | Quote

Charlotte,
The results are produced on a bar chart that has the values for 8 different questions.
The results never go below a 5. (If they did the employee would probably be terminated)
Scott

Reply | Quote

Charlotte,
The results are produced on a bar chart that has the values for 8 different questions.
The results never go below a 5. (If they did the employee would probably be terminated)
Scott

Reply | Quote

Are you trying to report on existing results in a different scale (ratio 2.5:1)? Is this for a report, a graph or what?

Reply | Quote

Hi,

I’ve had to deal with this situation before… I just used a mapping table and then used this to group the results rather than the existing column. I made sure that all records defaulted to 0 in the first instance. Created the mapping table (example below) then join on the field in your data table to the ID filed in this table. Group on the descrip field and count the KeyID from your original table. The same theory applies to making age grouping etc..

ID Descrip
0 Non Response
1 1
2 1
3 2
4 2
5 2
6 3
7 3
8 3
9 4
10 4

What you have to acknowledge is that you are changing the statistical signifigance of the outcome…4 doesn’t go into 10 as such.

Cheers
Tony

Reply | Quote

Hi,

I’ve had to deal with this situation before… I just used a mapping table and then used this to group the results rather than the existing column. I made sure that all records defaulted to 0 in the first instance. Created the mapping table (example below) then join on the field in your data table to the ID filed in this table. Group on the descrip field and count the KeyID from your original table. The same theory applies to making age grouping etc..

ID Descrip
0 Non Response
1 1
2 1
3 2
4 2
5 2
6 3
7 3
8 3
9 4
10 4

What you have to acknowledge is that you are changing the statistical signifigance of the outcome…4 doesn’t go into 10 as such.

Cheers
Tony

Reply | Quote

You should consult a professional statistician for this. There are statistical methods of converting one response scale to another one with minimal loss of information.

Reply | Quote

Don’t know if there is any interest but I brainstormed with a co-worker and we came up with this.
We divided 4 by 10 which is the highest value in both cases and got .4
We then multiplied .4 by all the values in the 1 to 10 scale. This still gave us a linear scale that mirrors the original
and keeps the same relationship.
10—4
9—3.6
8—3.2
7—2.8
6—2.4
5—2
4—1.6
3—1.2
2—.8
1—.4

Reply | Quote

Don’t know if there is any interest but I brainstormed with a co-worker and we came up with this.
We divided 4 by 10 which is the highest value in both cases and got .4
We then multiplied .4 by all the values in the 1 to 10 scale. This still gave us a linear scale that mirrors the original
and keeps the same relationship.
10—4
9—3.6
8—3.2
7—2.8
6—2.4
5—2
4—1.6
3—1.2
2—.8
1—.4

Reply | Quote

You should consult a professional statistician for this. There are statistical methods of converting one response scale to another one with minimal loss of information.

Reply | Quote

Charlotte,
I’m looking for a way to convert the existing survey results to a different range of values(1 to 4)
The existing survey with the 1 to 10 values won’t be changed since customers are familiar with it.
Hope this helps explain what I want to do.
Scott

Reply | Quote

You haven’t really provided enough detail to get any useful suggestions. I assume 1 through 10 represent a range of values, such as from Highly Satisfied to Highly Dissatisfied, right? Are you asking for help figuring out what the categories from 1 to 4 should be, how to adapt the existing survey to a different range of values, or what?

Reply | Quote