My thesis on " Using Mixed Methodology to assess sexual behaviour among Truck drivers at Tranport Nagar in Meerut, India
We have used Mixed METHOD DESIGN - In-depth interview and close ended questions
hence quantitative data was triangulated with qualitative data
So the queries are
1) First i had used a convenience sampling and sample size was 100 and non- respondents was 121 so does this pose a threat to external validity despite using a mixed method study design
2) What are the advantages in using mixed method study design what type of edge does it give , advantage of triangulation
Kindly please give references
Thankyou
It's not the non-response rate that drives the external validity but the representativeness of the sample you studied. In quantitative research, you are limited to trying to get basic demographics on non-respondents, but with a qualitative component you can ask participant about the possible reasons why someone would decide not to participate and develop more of a "feel" for the people you cannot interview.
Response rates are primarily relevant with regard to survey data, so your response rate of approximately 45% is reasonably good for a face-to-face survey on a sensitive topic.
As for the advantages of mixed methods, that really depends on what your goals are. If you want to compare what people said in the survey to what they said in the open-ended questions, that would be a classic use of triangulation (convergence in the results), and to the extent that they do converge, that increases your faith in the accuracy of the responses (i.e., validity).
In more general terms, the advantages of mixed methods would be covered in any number of introductory textbooks on that topic, such as Creswell & Plano-Clark's Designing and Conducting Mixed Methods Research, or my own Integrating Qualitative and Quantitative Methods (both from Sage).
It's not the non-response rate that drives the external validity but the representativeness of the sample you studied. In quantitative research, you are limited to trying to get basic demographics on non-respondents, but with a qualitative component you can ask participant about the possible reasons why someone would decide not to participate and develop more of a "feel" for the people you cannot interview.
What Ronan said using different words:
The issue with non respondents is that they could bias your survey results. You can postulate that there is some trait that is more prevalent in non-respondents, and that you do not know what that trait is. Maybe in this survey truck drivers are less likely to take the survey if they travel more than 50,000 miles per week, or drive Volvo trucks, or have families, specific religious groups, or just any factor you did not measure. I would suspect that some personality types will be less likely to submit to the survey, whether the interviewer was male or female, how questions were asked, ... and so forth. You can view this as a biased survey because the survey is not a random sample of the population of interest, or you can view it as a good survey from a population that is different from that which you originally intended. The conservative solution is to just remember that your results only apply to the population of drivers willing to fill out the survey.
Ashish:
If at all possible, please try for a better response rate so that your chances of publishing the research would be better. Here's an online excerpt on benefits and challenges of mixed methods:
https://mixed-methods-research-nnu-group-project.wikispaces.com/Benefits+and+Challenges+of+Mixed+Methods+Research
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size