Hi. I'm new in statistics and I have a question regarding questionnaire design.
I want to design questions of one variable with True- False scale and the rest of the variables questions with 7 point Likert scale.
I should note that my true-false scale in this regard can not turn into Likert scale. For example suppose I want to measure prior knowledge of individuals toward a subject so I ask some true- false questions to know if they have prior knowledge.
Is it make any problem when I want to analyze by SPSS?
The true or false questions would be like a screening that allows to point out a characteristic of the sample to be surveyed, you should consider only the multiple questions as part of the instrument.
No, read about the scale of measurement. you should know the types of scales and where each belongs. true/false is nominal and cannot be combined with a Likert type scale. except if the true/false will be used for the descriptive part of the analysis.
of course you can have binary /dichotomous scale in your study.
I suspect that you do not mean likert scale, but items with a likert-type response scale?
As mentioned by Alireza shabani shojaei you can have items with differing response scales in a study. The response scales can differ both in the meaning anchoring of the response categories and with regard to the number of categories. You can have items with all the same response categories in a scale, but you can also have items with differing response scales, in the same scale. It is crucial that the item text and the response categories are designed well so that the possible responses actually fit the items questions or statement. It is all to common in many disciplines to apply the same response scale to all items (and usually strongly disagree, disagree, neutral, agree, strongly agree) without considering whether these response make sense in relation to the item texts. Likewise it is all too common to then interpret higher scores as the persons having more of the trait measured, when in fact it only means more agreement with the statements.
Hope this is usefull
Tine
Yes, you can create two scales, one using T/F and the other using a multi-point Likert. I would suggest that for the knowledge test you use a multiple choice format rather than just true-false. Multiple choice does a better job of assessing knowledge.
Defiantly yes, please get idea from below article, author uses different scales for every article and article is published in Q1 journal (Top journal of business & management)
If you are open to an Item Response Theory approach, the Partial Credit Model (Masters, 1982) can handle items of varying scales in a unidimensional environment. In a multidimensional IRT setting, the multidimensional random coefficients multinomial logit model or MRCMLM (Adams, Wilson, & Wang, 1997) can be applied.
- Badges
- Science topic
- Similar topics
- Correction