I'm doing a (statistical) survey on whether or not international students at my college receive satisfactory grants. Here's the deal, I don't expect them to say yes it is satisfactory. I need a way to formulate questions to tricking them into giving me an accurate questions. I've already got my formalities written but how can I approach them with these questions? Bonus would be to tell how much more they would need (if at all).
$\begingroup$
$\endgroup$
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
3
In my opinion, test-retest method would be suitable to your request.
This procedure is quite simple: individuals are asked to take the test and then take the same test again at a later date. The scores are then compared. The closer the scores are, the more reliable the test.
This methodology is appropriate for instruments such as IQ tests and surveys because there is little chance of people experiencing a sudden jump in IQ or suddenly changing their opinions.
I wish i have been helpful.
REFERENCES
PICONE L., Pezzuti L., Ribaudo F., (2013), "Teorie e tecniche dei test", Roma: Carocci.
-
1$\begingroup$ Isn't this impractical? Why would someone want to take a survey again? $\endgroup$ Commented Mar 8, 2018 at 2:50
-
$\begingroup$ @ReeshabhRanjan It is a method which increases the probability that a test reflects the behaviour. I don't understand your question $\endgroup$– FilCommented Mar 8, 2018 at 16:47
-
$\begingroup$ I want to take a survey of around 500 people. Now once a person has filled in the survey form, why would he like to do the same thing again? $\endgroup$ Commented Mar 9, 2018 at 10:22
$\begingroup$
$\endgroup$
1
Shuffle the questions to make at least two survey versions, which you then statistically compare to filter out any "leading" questions. this to compensate for the (mental state) influence that certain earlier questions will have on the following questions.
-
$\begingroup$ Can you please give an example? $\endgroup$ Commented Mar 8, 2018 at 2:50