"To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of."
The same applies, with perhaps even more force, to analysis of survey based research.
Do you need to design a survey?
CALL ME FIRST!
Designing a survey seems simple. But it's not. There are lots of decisions to make and lots of traps for the uninitiated and unwary. I've designed lots of surveys and analyzed many others. Telling someone to "go back and start over" is not something I like to say and not something anyone wants to hear. But it happens. Sometimes, I can only say what the survey died of.
- How long should the survey be?
- Should it be web-based? Paper-based? Conducted by phone?
- Do you need to give participants assurances of confidentiality and how best to do that?
- What about translation into different languages?
- What response scale should be used?
- Should you offer a reward of some kind for finishing, and if so, what?
Do you know the answers to the above? (hint, many vary depending on the topic, the purpose and the audience).
What's wrong with this question?
Please say how much you agree or disagree with the following statement:
"Barack Obama is to blame for the current economic problems and should be impeached?" (I am not talking about its political content. Something is wrong with how the question wording).
How about this one:
Please state your age _____
Please state how much you agree with this:
I would never not obey the law
You're an expert in what you want to ask and who you want to ask. But you need an expert in how to ask. So, if you need to design a survey, call me first
If you are a PhD student you are probably aware that it is possible to buy a dissertation. There are lots of...
Macros can be a very complex topic, but some very simple macros can make life easier for a data analyst or...
Signal versus noise Description and inference That is statistics
Many times, researchers will categorize continuous variables. For example, birth weight of human infants is...
OK, there are lots of places where it’s written that using RUN statements makes code look cleaner, but that...
In statistics and research design, there are two types of study: Experiments and observational studies. Some...
When you have bivariate data - that is, data on two variables - either or both may be categorical or...
In ordinary regression, we are interested in modeling the mean of a continuous dependent variable as a linear...
Modus tollens in logic In logic, there is an argument style called modus tollens: If H0 then R Not R...
Question: What makes a good statistician? My answer: Curiosity, persistence, some anal tendencies, a...
I introduced this series the other day. Next up in the list is "getting help". In both SAS and R, there are...
Regression refers to a collection of techniques for modeling one variable (the dependent variable or DV), as a...