I always try to get as much information as possible from responses, so the more categories you have the richer the data is. This, however, has to considered against the possibility that people may find it harder to answer questions with a large number of response categories. At a basic level there is evidence that for many forms if statistical analysis using 5 or more response categories allows you to treat your variables as continuous, and this makes things easier.
Please notice that when there are more than 5 response categories you may not consider such a DV as continuous. See the material on ordinal logistic regression in the notes attached. Best wishes, David Booth
Weijters, B., Cabooter, E., & Schillewaert, N. (2010). The effect of rating scale format on response styles: The number of response categories and response category labels. International Journal of Research in Marketing, 27(3), 236–247.http://doi.org/10.1016/j.ijresmar.2010.02.004
You might also want to consider the labelling of scale-points (e.g. whether you label all points or only the end-points) when selecting the number of response-options. In an experimental study, Menold et al (2014) found that labelling all scale-points increases reliability of a measure, however, as number of scale-points increase, the respondents' attention to scale-points decrease, suggesting a potential trade-off between quantity of information and quality of information when number of scale-points increase beyond 5.
Study: Menold, N., Kaczmirek, L., Lenzner, T., & Neusar, A. (2014). How do respondents attend to verbal labels in rating scales?. Field Methods, 26(1), 21-39. (be mindful however that the study used a German sample and assessed the effect of scale-length and labelling using scales that measured national socialism and attitudes to the EU which might be a confounding variable). Best of luck with your study!
As with all scaling, it depends on the question/construct under study. The two scales I use most frequently are:
0-10 so you can use never/not at all/none and always/completely, with the rest non-labeled points as radio buttons, to take advantage of familiarity with the base 10 number line in those countries where it is in use.
Note -- I do not label 0 or 10 with numbers, only labels. It is a common error to label with both words and numbers, just extra reading and potentially biasing. You can enter your numbers as the reporting value for your stat package, of course.
Also, quite frequently, I will use a categorical scale that is either 3 or 5 points -- Disagree/Somewhat agree/Strongly agree or when assessing frequency, which I do a lot, Never/Occasionally/About half the time/Usually/Always and variants of that depending on the question.
I have never used 1-5 or 1-7 as these are not naturalistic -- they have no real-life semantic relationship, apart from days of the week or fingers on one hand, which aren't usually relevant.
In terms of measurement scales, the most common way is Likert type with 5, 7 or 9 divisions. Although the 5 division scale is most often used, the use of 7 or 9 divisions leads to measurements with a higher degree of accuracy.
For studies on General population and\or urban areas, I tend to find 1-5 scales more convenient, easier to deal with and easier to be labelled.
I generally use 1-5 scale when I need to know a general idea about the attitude\opinion of the population regarding a certain topic. (For ex. Do you think caffeine is good for health?)
While I use 1-10 scale when I need a more specific response regarding a specified behavior\phenomena or from more specified population. (For ex. On a scale of 10, how much are you dependent on caffeine?)
This is of course very general and very dependent on the type of question\population.
Bottom line is, Larger scale=More information \\\ Smaller scale=Easier interpretation
First of all, thank you for answering my question. In fact, the interpretation of your reply is really clear and very useful to me to understand the function of the scale.
چالاک جان سلام معمولا وقتی از 5 مقیاس بیشتر میشه افرادی باید پاسخ بدند که به مساله یا متغیر تسلط و آگاهی کامل داشته باشند چون رنج کوچکتر میشه و معمولا مقیاس بالای 7 مربوط به متخصصین اون زمینه است. موفق باشید