Currently, I'm on the process of making my research paper. I made my own satisfaction survey questionnaire. Is it okay to make my own? How can I measure the reliability of my questionnaire? Moreover, my variables are FUNCTIONAL PERFORMANCE and TECHNICAL PERFORMANCE of a building. Attached here are my evaluation instruments. Please, feel free to comment or suggest. Thank you so much.
P.S I want to validate my instruments if they are applicable for use in building performance evaluation. At the same time, I want to assess the building performance of the building. Just having a hard time on how to show relationships of functional and technical performance with the building performance.
You can design your own instruments based on your literature review. I would suggest you look for similar tests on the internet just as a guideline - generally you will never find one that exactly suits your study except maybe in the psychology field. You should test them using Cronbach's alpha for reliability e.g. https://statistics.laerd.com/spss-tutorials/cronbachs-alpha-using-spss-statistics.php
I always recommend including a core set of questions that others have asked in similar settings, so you have a basis for comparing your results to the findings of others. Give priority to questions with published reliability statistics. Also gather standard demographic information, again so you can compare your results to others. After that, go for it. Ask whatever you want. Be sure to keep the questions simple. And check your questions for reliability when the results come in.
The answer is a bit complicated. If your purpose is to develop a measurement instrument to guage the satisfaction of building facilities for future use, then you would typically go through a rigorous process of instrument development, outlined in DeVellis, R.F. (2011). Scale development: Theory and Applications. That would require reviewing the literature to identify scale domains, generating a large number of items, sampling a large number of people based on your items, using exploratory factor analysis to reduce the items etc. There is no sense calculating a Chronbach's alpha for a scale that has not been developed in a systematic fashion. If, however, you merely want to identify variables and design a a questionnaire for general bivariate statistics and percentages, you can write up questions and survey how ever many people you want. You just can't claim that the subscales are valid or reliable measures of the constructs you are proposing. Another alternative is to find an existing scale that measures satisfaction and has good reliability and validity data that is published. Then you could either use it with your questions or adapt it to your project. I have looked at your instrument and it seems very project-specific: That is, it is unlikely you would be developing this for wide-spread usage. For that reason, I would say it's find to use it without actually naming it as an instrument, sum the subscale areas, and run those satisfaction levels against the building metrics; you'll have to establish cut-scores for the building metrics so that size (big/medium/small or big/small) can be run against your satisfaction scores. Looking at your questions, they are a bit general and may mean different things to different people. For example, people may not like the color but there may not be anything objectively unsatisfactory about the color so what are you measuring. Similarly, how can people judge air quality? Whether there is an odor? Whether they have used a detector to ascertain whether the quality is poor despite smelling OK? I would give these questions to a few of your colleagues and ask them to identify questions where they are uncertain of the meaning then reword for specificity. Hope that helps.
It is possible to develop your own questionnaire provided your qurstions are aligned with your objectives of the study...if those questions will lead to what you would like to determine. Furthermore, in checking the validity and reliability, you must do pilot testing of your questionnaire. The purpose of this is actually to have a feedback of your questions are clear to your respondents and if all of them have the same perception and understanding about what is being ask...Morever you may know whether the objectives met with their answers...it really takes time to finally obtain validation of your questions...I've been there...it was hard but worth it. Goodluck in your endeavor.
I was about to adapt same questions used from previous researches. However, my professor told me to make my own evaluation instrument.
Hi Riselle. It is safe to adapt same questions used from previous researches. However, not all questions will determine the objective of your study. Again, my question to you, are your questions answering your objectives? So there should be pilot testing of your questionnaire to validate its clarity and reliablity, most especially to your target respondents.God bless:)
You've gotten some really good feedback from others. I'll just chime in a bit. While you can develop a new measure, keep in mind that the psychometric properties of that instrument may (actually probably will) be called into question when/if you try to publish the findings. The DeVellis text provides a really good overview of scale development. It sounds like you are using the Gronroos model for service quality (technical vs. functional quality). Generally speaking, you are better off using an existing instrument that you can adapt for your purposes. Still, creating a new one may be necessary. Just don't forget to actually speak with the people who are in the target population so that you can ensure that you are tapping the domains of importance to them (and creating items that reflect those). It's possible that the Gronroos model may not be the best fit for your particular instance. If you adapt an existing instrument or create your own, you can always at least provide a basic estimate of reliability (e.g., Cronbach's alpha) to appease some reviewers/readers.
Best of luck!
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers