![]() ![]() Example: Quantitative researchYou investigate whether a new drug reduces the effects of fatigue. Improperly cleansed or calibrated data can lead to several types of research bias, particularly information bias and omitted variable bias. Using hypothesis testing, you find out whether your data demonstrate support for your research predictions. In quantitative research, you collect data and use statistical analyses to answer a research question. Frequently asked questions about data cleansing.I'm not sure if the routines in Solver will be stable enough, but you might try using Solver to perform the regression. Not knowing exactly how MS implemented this arbitrary 0, I don't know for sure if this will work, but it might be worth a shot. I factor out the 1e-18, so LINEST returns a value between 0 and 20 rather than 0e-18 to 20e-18. ![]() Is there a way to set up the regression so the parameters are "scaled?" For example, I perform a regression quite frequently where one of the parameters is on the order of 1e-18. I'm not familiar with this particular type of regression, but, if you can set it up so you can plot the pertinent data in a chart and use the chart trendline feature, MS didn't apply this "arbitrary 0" algorithm to the chart trendline routiines. If an earlier version of Excel can generate the correct coefficients, then that should solve your problem. I do not know the details of this algorithm (you know how secretive MS can be about their code).ĭo you have access to an earlier version of Excel? Earlier versions employ a reportedly less stable numerical algorithm for performing regressions, but they don't arbitrarily 0 coefficients. As I understand, when Microsoft reworked the algorithms for the spreadsheet regression routines (LINEST etc.) for the 2003 release, they included an algorithm that looks at each coefficient and arbitrarily sets to exactly zero (0) any coefficient that it believes should be 0. ![]() I think you may have just bumped up against one of the issues that has made me very leery of upgrading to Excel 2003 or higher. Involving very *large* numbers may not yield accurate results due to problems with the STDEV function, but that centering the data beforeĭoing the analysis will correct this problem in most cases. It is interesting to note that I just read an article found at which stated that regressions If anyone wants a sample of some test data I have used to troubleshoot these problems, you can download a file from: I was wondering if anyone had any insights on why I am experiencing these problems. When doing these multiple regressions with non-centered predictors, all regression coefficients are estimated accurately. Initially, I thought the problem might have to do with the cross-product of the centered predictors, but even just doing a regression with one of the centered predictors (for certain centered predictors) yields a regression coefficient of 0 (although it should be non-zero as per SPSS 14.0.2). I have spent many hours troubleshooting this problem (and searched many forums on the internet) and still do not know why this is happening. When I run the regression in Excel with the centered predictors, some of the regression coefficents in the output are estimated to be 0, although they are clealry *not* 0 as estimated by SPSS 14.0.2. This multiple regression equation is structured to test for interactions between the two continuous predictor variables (x and z) as prescribed by Aiken and West (1991) in their classic book. More specifically, I have a dataset with 18 observed data points containing a criterion (y), a centered predictor variable (x), another centered predictor variable (z), and the interaction of the two centered predictor variables (xz). Recently, I have encountered some problems related to getting accurate regression coefficients in Excel 2003 (all Office updates installed) from a dataset with *small* (standard) numbers, which contains centered predictors. ![]()
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |