I am a student brand-new to clinical research. I read 10 chapters of Andy Field's Discovering Statistics Using SPSS to prepare myself to perform the statistical analysis on my research paper, and followed his steps closely. Do you have any suggestions on how I might best check my work? Is consulting a professional or faculty necessary, or are there rigorous methods to assure that I've correctly calculated everything?
Thank you!
Discuss with your guide it is very important and will be helpful also.
Congratulations on reading the 10 chapters! Field has an entertaining style. It's hard to hold all that information in without concurrent hands-on data analysis. It wouldn't hurt to check with a faculty member or similar person. The main things to check would be if you are meeting the assumptions required for each statistical test, and if your data generally "make sense" in that it measures what it is supposed to measure. Think of construct validity. And possible sources of bias in your data or your analysis. Do you have outliers in your data. Consider if non-parametric tests could also be used (although you often lose some precision of results). Are your data normally distributed when it should be; would transformations be necessary. Recognize any limitations of your data and your methods. And certainly take a random look at published articles in your field and see what they did statistically, including their tables and graphs. Think about what might increase your likelihood of a Type I error or Type II error. I also recommend a book called "100 Statistical Tests" by Gopal Kanji. It describes some generally more obscure tests out there that could be useful. And I highly recommend Laird Statistics, a subscription-based website on SPSS data analysis. But keep things mainstream when you can and don't over-complicate it for yourself. Good luck!
Dear Lauren Bockhorn
I think you need the out put of your senior colleague
Jerry Miller's advice is spot-on! And thanks Jerry for the Gopal Kanji book recommendation. But I must disagree (in part) with Mohammed Yousif Abbas Mohammed - rather use an experienced researcher in your own field than a professional statistician, unless they are familiar with your field. Too often, pure statisticians will take a purist line that ties the beginner in knots. You need someone who knows what will be accepted in the field in question, and what will be publishable.
I agree with that,, Don. I'm in the medical field and I once had to help someone present a medical article written by some kind of economist. It was filled with the most convoluted, bizarre statistical procedures that were apparently used in economics and (mis)applied to medical data. We basically had to decipher it and still used guesswork. An article like that would never be cited or advance the field in any way because no one understands it and wouldn't have the time or desire to figure it out. Statististics are statistics, one might think, but they are very different from field to field. It is far better to use the guidance and experience of someone in one's field of study. A professional statistician is probably concerned with more complicated things than a novice's research, anyway.
I agree with what has been written about checking with a general statistician. Since you are in a Health Science Center, I'd check with the statisticians there. They will understand stats to use with medical research. Also check with them before you collect any data. It can be very frustrating to collect data and then talk with the statistician who might tell you the data you collected is only amenable to analysis by lower level stats.
Make sure to show them what you think should be done statistically and then have him or her explain why you are correct or would recommend alternative approaches.
Good luck. Hope you enjoy your research project.
It depends from you required research. We can use Cronbach index, correlation and presentation in an exploratory research and fit index within confirmatory research.
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers