Stat Analysis: What do you know to not get lost in Statistics?
Often times I talk with customers and I inquire why they used our statistical services. My question to them is: "You have Stats analysis software, and more or less you know how to input the data, so, what is missing". And invariably, the answer is "I don't know how to interpret the outputs".
Conducting a Stats analysis is not a trivial endeavor, and it often times requires solid understanding of Statistical foundations in order to yield meaningful results. And we are talking about just analyzing and interpreting the data with some statistical software package (like SPSS).
We are not even considering the difficulties of coming up with a research design, with valid and reliable constructs and with data collection. Just the mere data analysis in itself could be feat.
|
Most Common Statistical Analysis
We find ourselves commonly doing stats consulting work for our customers who need detailed write up to summarize procedures and results. The most commonly used statistical software is SPSS, but sometimes we also work with customers who require more advanced tools like SAS and Stata.
Also, plain Excel is very common too, though Excel does not allow much flexibility and power on what you can do with it. It is also interesting the case of R, a free open source Stats software that has been gathering a fervent community, which has turned the software, based on its extensible nature, into a real beast.
The main disadvantage of R is that it command line based, without the typical graphical interface, with drop-down menus. Though, that has changed recently with several graphical R wrappers that contain R as the statistical engine and have user friendly menus to access most of its functionality.
In terms of the type of statistical analysis we conduct, the most common one is multiple linear regression, with full fledged model building , regression assumptions, etc. Factorial and One-Way ANOVA are also very common.
Can you conduct those statistical analyses yourself?
You certainly could, but I would say it could be tricky. As the statistical softwares have become more sophisticated, uninitiated users feel more tempted to use these softwares themselves, but they come to the harsh realization that you need expertise to interpret the results the software provides.
Of course, mostly everyone can go in Excel and run a linear regression and get the regression model, but that is dangerous, because you should not interpret, or use for matter, a model that does not strictly follow the linear regression model assumptions.
Excel does not provide precise diagnostic for regression assumptions, other than basic residuals, so the more complex and more realistic (and more accurate) you want your analysis to be, the more you will need a statistician to work with you.
Is there any Stats analysis that I could start with?
Yes, for sure. Probably working with t-tests and one-way ANOVA are good places to start. Those procedures are parametric tests that still require the fulfillment of certain assumptions to be used, so you still need to be careful. But it is really easy to get started with those.
I would caution that you don't get started with more complex stuff like mixed design ANOVA, with within-subject and between-subject factors and with more complex designs. ANOVA can come with a world of subtleties, so it would be safe to start first with consulting with a statistician.
In the end, you always can play with the software and get results, but if you need professional quality in your results, make sure that you work with a statistician until you get the ropes of it.
You can send you Stats homework problems for a Free Quote. We will be back shortly (sometimes within minutes) with our very competitive quote. So, it costs you NOTHING to find out how much would it be to get step-by-step solutions to your Stats homework problems.
Our experts can help YOU with your Stats questions. Get your FREE Quote. Learn about our satisfaction guaranteed policy: If you're not satisfied, we'll refund you. Please see our terms of service for more information about this policy.