All-
While I can't be sure if my subject line is entirely accurate, I hope to expand on my problem/ question in some detail. I am currently working with an aggregated .sav which was created from a long (40k cases) and wide (several hundred cars) where my aggregating factors were year and state. Given that my original cumulative file was mostly a public opinion file- at least this is what I am concerned with- I know have a dataset of the same information only my unit of analysis has gone from Respondent to State. There are several things I am attempting to do with this but I should hope that I can re-outfit syntax to address the multiple issues I have because they are relatively similar. Let 'state' be a variable over the 50 states, 'year' be a variable over 60 (approx) years and each case is a unique combination of 'state' X 'year' so that based on the numbers above, I have 300 cases. Each of these 300 cases has values (or sysmis) on many variables which reflect the descriptives for a particular public opinion item. Let 'opin1' be a question of public opinion in the original cum .sav (as opposed to my agg'd file). Thus, there exist the vars 'opin1_mean' 'opin1_median' '..._sd' '..._min' '..._max' and a var N which is valued according to the weighted number of original cases in each aggregate case (i.e. respondents in each state each year, after weighting). I am attempting to compute an O'brien reliability coefficient (Sociological Methods and Research, 1990) which is of no real signifigance here, just suffice it to say it is the appropriate measure and I am looking to compute it by hand, that is telling SPSS to manipulate certain values I have in certain ways. As a point of information, this measure accounts for the reliability of agg'd measures that are calculated from nested units of analysis. This stands in contrast to something like cronbach's alpha which accounts for the reliability of a particular respondent. Now the heart of it. I need to calculate xbar where xbar= the summation of {(nj X xbarj) / n}, where xbar is the weighted mean public opinion over aggregates 1 through k where 1<j<k (INCLUSIVE) , nj is the weighted n in the jth aggregate, xbarj is the mean in that jth aggregate. Generally, I want to calculate this xbar measure and dump it back into my agg.sav, but I do not know how to compute such a variable where the value in any given case reflects more than just the values specific to that case. Notice that if xbar is dumped back into agg.sav it will have the same value for many of these already agg'd cases. This may or may not seem peculiar. The reason is that I will be computing something single factor ANOVA-esque on these agg'd values where time (or state) will be the factor where I compare a within value to a between value. I recognize that this can be done via the AGGREGATE command but I then need to manipulate these new agg'd values and as a result of there being several agg'd values will manipulate to compute the coefficient for each case, I am probably capable of punching it out sans SPSS just as quickly. Is there some way to compute the variables and use them but only dump out the desired, manipulated values? So, I ask, does anyone know how to calculate such a thing via syntax. I have searched here and elsewhere and have not found anything. Hopefully I have not overlooked anything and am now wasting a bunch of 1's and 0's. Thanks for your time! Pursuers of knowledge, myself included, thank this group for everything. As a note, I am not new to statistics, but very new to SPSS by way of syntax. Go easy on me please. -Jonathan Olmsted Jonathan Paul Olmsted University of Delaware Political Science 2007 |
Jonathan,
I think this is the thing to do because, as i understand you, you want to multipy your state-year stats (e.g., mean for item 13) by the n associated with that state-year. You then want to sum up those n(j)*xbar(j) products and divide that sum by the sum of the n(j)s. If so, then Compute tot=n*xbar. comute tag=1. aggregate /outfile=*/break=tag/tot n=sum(tot n). compute xbar=tot/n. Gene Maguin |
Free forum by Nabble | Edit this page |