Our team consists of three researchers, we will be designing a questionnaire having both open ended as well as close ended questions. After that, we will be emailing the questionnaire to our medical colleagues, most of them being situated in different parts of cities and states of India. Further, after they complete their questionnaire they will be instructed to email the questionnaire to their colleagues.
In respect to this , I have the following queries
Q1. What type of study design is this - is it a cross-sectional?
Q2. What type of sample technique id being used? Is it snowball technique ?
Q3. Is it a different type of non-probability sampling technique?
Q4. Is it neccessary to compute sample size here?
Q5.Will a pilot study be suffecient to validate this study or something else in addition has to be done to strenghthen the validity and reliability of study?
Q6.Can we still analyze using statistical tests in such a study?
Q7. Should prior to conduct of study should we compute the non-response/non-refusal rate to fill the questionnaire - if yes how will that be useful and what is the importance of that?
Q8 What will be the inherent limitation in this type of Study or due to this procedure?
if there is any pdf please do attach it thankyou
cross sectional I think, but is it random, in that do you have representativeness from different postcode regions, snowball sounds to be true but if just general email to say email to colleagues also sounds haphazard in that not specific enough in terms of you knowing then whom and how many will get this, validation question is difficult to answer unless someone knows content of questionnaire, are you using a repeatable validly tested apparatus for measurement, statistical tests can be used for analysis if you have multichoice questions, you can do frequency curves, other then that qualitative analysis, non response rate should be calculated how many don't send forms back and that will probably be high, other then that I am unsure of answers, good luck. and I've not really answered this well as in a hurry, family shouting at me to get in the car so hurrying.
Deborah J Hilton - Thnks for responding but mam if you have any more ansers to my queries then please do add on .
Thankyou very much
You will not be able to calculate the not response rate, other than to your initial shipment. For the snowball from there, you would need to know how many surveys were forwarded by your initial list, and how many each subsequent list forwarded to get your denominator.
I suggest you get the full list of relevant colleagues, take a random sample from those, and then you have a more powerful study that you can accurately calculate response rate.
Peter M Dore Thankyou sir for your answer but sir can you give me a reference or attach a pdf related to electronic media
I am sorry, but I don't understand what you are asking.
I suggest Qualtrics or maybe Survey Monkey, those are easy tools for doing electronic surveys. Qualtrics easily manages the e-mail list, and can even send reminders to those that have not responded after a given amount of time.
Peter Dore - Thankyou sir
What i meant is that how do we conduct electronic survey in medical research via email? Is there any link or pdf which can be helpful for me?
Thankyou
There are ways to create forms that people can complete and return with an e-mail, but I strongly suggest you email a link to an online survey, and have them respond to that. The online survey tool will do a lot of the work for you.
http://www.qualtrics.com/ is a good one, my favorite.
https://www.surveymonkey.com/ is good, and possibly cheaper, though qualtrics is free to everyone at my institution.
Good luck.
Hi Ashish,
Let me answer you question to question:
A1. This is definitely a cross sectional study as it finds out the 'knowledge/ perception' of a sample in a 'definite point in time'. In other words, there is just a one point data collection.
A2 & A3. This seems to be a snowball sampling (non probability) due to 'chain referral'. However, just to add on, snowball referral is recommended in rare diseases where recruitment may be a problem. I do not actually see the need of a snowball sampling here. Why don't you collect data from the MCI or state medical councils about the doctors and then contact them? A little more tedious process, but your study will become more unbiased.
A4. I would back question you here, How will you calculate the population in this kind of study? (That is one of the reason why snowball sampling is recommended in 'rare' conditions. I feel being a doctor is not that rare ;)!). In case you have a population it is always better to compute a sample size if you are seriously thinking of not wasting your time doing some research that doesn't actually matter (unless you are actually doing it in whole of the population!).
A5.Please validate the questionnaire. It is a must! Please read the first article where they have developed and validated a questionnaire.
A6 & A7: A face validity is needed. Calculate cronbach's alpha. Refer to the second article.
A8. Yes response rate is important, but then the same question. What is the study population? How do you know non response? Refer to the third and fourth articles.
I feel that I have already stated the limitations of the study as I have answered the previous questions.
A6. I am attaching a file about what statistical test to use and when.
Hope it helps.
http://www.ncbi.nlm.nih.gov/pubmed/25638047
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3650590/
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2384218/
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2254236/
Electronic data collection, or I'd better say Electronic data capture is much more than just emails or computer based survey forms. It is from capturing source data and integrating it into data management system. It does save substantial time spent in research, specifically in clinical trials; especially in multicentric trials with an integrated management system. Its quite a wide field in itself.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire