Hi,
Would it be appropriate to convert a 5-point 7-point[Not important at all-Very Important] scale to dichotomous scale for analysis purpose? A reference would be appreciated if its okay.
Thanks,
Ali
In general, it is not appropriate to transform a scale to multiple points in a dichotomous scale, because it threatens to "flatten" the data too.
However, in the case of Likert scales, whose preferences would summable, it is possible to carry out the processing.
For example, in a 5-point scale:
- low
- bass
- acceptable
- good
- excellent
you can sum the answers to the first two items to get the item "unsatisfactory"; adding the other three items you would get the item "satisfactory", getting a new dichotomous variable.
Instead, in a 5-point scale of the type:
- strongly disagree
- disagree
- indifferent
- agree
- strongly agree
to create a new dichotomous scale with item "agree", "disagree", you should add the upper and lower pairs of the scale, leaving the central item (indifferent).
In this case there would be a loss of information, more or less according to the number of preferences (frequencies) obtained by the central item.
I hope everything is clear and I hope I was helpful.
Dear Ali
first of all I would ask you why you want to do that?
When you do that you must take on mind that you are loosing some important information so you can't know the severity of the agreement or disagreement you are just know if yes or no, so Likert type make you have more information and also it has the uncertainty choice. By dichotomizations you make your items kind of forced response items while respondents have to say yes or no while sometimes it's neither yes nor no.
Thanks for your reply.
First, I Agree with your opinion. That was the reason I used 7 point scales at first place.
Regarding why I want to do so; I have recorded imprortance of different threats on 7 point scale to verify a taxonomy. To further analysis the results to study relationship with other variables, i was looking for a way.
Gioacchino suggested two ways of doing it.
Best Regards,
- Badges
- Science topic
- Similar topics
- Psychometrics
- Measurement