QUALITATIVE SOFTWARE LAUNCHED
 |
Observant LLC has launched QUESTRILL® (patent pending), an online software tool that drastically improves the speed of qualitative research. By its nature, qualitative research results in the production of voluminous data, often in the form of un-coded hand written or typed observational notes and comments. This data must then be processed and coded by the researcher before and during in-depth analysis. This time-consuming activity can limit qualitative research in circumstances that r equire rapid analysis and dissemination to address critical issues. The problem is particularly acute in studies which would involve multiple sites, multiple researchers, or are international in scope. The use of QUESTRILL® increases Observant LLC’s qualitative capabilities in four ways: (a) minimizes the time burden for all processes (including training, project setup, notes and observations, ad-hoc edits to research parameters, textual coding, queries, reporting and dissemination); (b) maintains the flexibility to accommodate varying researcher preferences for discussion guide creation, note taking, analysis and reporting methods, etc.; (c) accommodates alternative and iterative investigation objectives both “on the fly” and post hoc; and (d) makes raw data and annotated coding accessible on-line to researchers and clients in real time. QUESTRILL is a trademark of Observant LLC |
 |
PUBLICATIONS AND WEBSITE UPDATE
|
| Please visit our website for more information on recently published articles on ethnography and research methods. Also, our website includes new examples of recently completed work in the vaccine and sexual health fields. |
|
NEW PERSONNEL
|
 SCOTT JANKO, JD/MBA
Scott joined the Observant LLC team this past winter and brings his experience in both qualitative and quantitative research from a broad range of industries.
|
|
 DYLAN JOHNSON, Ph.D.
Dylan joins Observant LLC from P\S\L Research in Montreal where she managed quantitative and qualitative market research projects for pharmaceutical clients and has an extensive background in qualitative policy research. |
|
 JASON FOGLER, Ph.D.
Jason is an NIMH clinical psychology fellow with expertise in qualitative and quantitative research techniques that uncover latent drivers of behavior, particularly expressed emotion. |
|
NOTE FROM THE PRESIDENTS:
MAINTAINING A CENTER OF EXCELLENCE
We recognize that in order to best meet our client needs, it is critical the firm maintains an in-house research program. Our internal R&D efforts focus on (a) the development of new methods and tools for both qualitative and quantitative research; and (b) conducting in-house studies, independent of client funding, on issues central to the work of our clients. This program has resulted in the patent pending software application QUESTRILL® (above), and the firm presenting in the ‘Research Excellence’ series at the Pharmaceutical Business Intelligence Research Groups’ 2006 conference. Our paper prepared in collaboration with Asst. Professor James McQuivey at BU’s College of Communication, “Repositioning Pharmaceutical Companies in the Minds of Consumers”, integrated projective qualitative methods, and latent statistical modeling, to suggest a strategy for overcoming negative impressions of the industry. Please be in touch if you would like to know more about these and other efforts, or would like to suggest issues for research where independent data could be helpful to you.
Best wishes for the summer,
|
|
|
RESEARCH COMMENTARY:
Is the Difference Statistically Significant?
The most common criterion for determining statistical significance is p < .05. While this is usually appropriate, under certain circumstances it is advisable to use a lower threshold (e.g., p < .10 or higher).
Most of us have learned (erroneously) that a p-value of .05 or less means that there is a 95% chance that the difference is “real” and can be treated as a “significant difference.” But the case is actually more complicated arising from the fact that there are two kinds of mistakes in statistics: 1) False Positives - accepting a finding as significant when it is not real, and 2) False Negatives - failing to accept a finding as significant that is in fact real. Importantly, p-values and confidence intervals assess only false positives.
Unfortunately, by myopically focusing on avoiding false positives many researchers significantly increase the likelihood of false negatives. This may lead us to fail to recognize the potential of a marketing strategy that would significantly increase sales, thereby leaving millions of dollars on the table.
So what should we market researchers do? We should balance the risks of false positives and false negatives to minimize the overall chance that we’ll make either mistake. While p<.05 is often the best criterion, p<.10 may be more appropriate in the following situations:
- Any time the number of responses per group analyzed is 50 or less. False negatives rates can be very high when sample cost or availability restrict sample sizes.
- When comparing stimuli that are very similar (e.g., visual aids, details, ad concepts, positioning options), because in such conditions any real differences are likely to be very small.
- Anytime the cost of false negatives is especially high
Erik Coats, Ph.D.
Vice President
ecoats@observant.biz
|
|
printable
version 
