What happened on SPSSforum.com in March

April 3rd, 2009 andris

SPSSforum.com is forum for users of SPSS. It is owned and set up by SPSS users. We have 4392 registered users who have posted a total of 6945 articles. In this post I will give you the highlights of what was posted on the forum in the last month.

linkedinBut first I want to announce the start of an SPSS user Group on LinkedIn. Are you a LinkedIn user and an SPSS User? Please join our group!

Statistics for beginners
Moderator Fierce responds to a question by Leanne about calculating a mean from age ranges.  Leanne has conducted a survey where an age range is asked with an open top range (75+). Fierce explains to create a new variable based on the range mean and take the mean of the new variable. For the open ended he advises to take the lowest value (75).

Kate used SPSSforum.com to ask advice on alternative test to the Chi Squared test. She wanted to use the chi squared test on some data, but has some categories which have a value less than five. Also some of the expected values for some categories are zero. Fierce explains her Fisher’s exact test is an alternative (ANALYZE > DESCRIPTIVES > CROSSTABS > EXACT), and statman adds it needs to be installed, this is an add-on. Statman also explains Kate that she can use the results but to note that there are case count issues.

Heavy Statistics
User blondestats stated that a Pearson’s chi-square was significant for a 2 x 4 matrix. The user than calculated std. residuals for each cell to determine what cells contributed significantly to the model. None of the std. residuals were greater than +/- 1.96 even though chi-square was significant. Blondestats ask for explanation on this. User EJGumbel explains that all the Chisq test is doing is testing for deviation away from independance. What the Chisq test is testing for is non-”flatness”, which user Blondestats has. That doesn’t mean that any one cell is particularly non-flat, just that collectively they are not flat.

Coding SPSS Syntax
Fierce explains user Tana which Syntax can be used the get the earliest data in a range of dates.

In another post Heidi asks how to recode email addresses into a new variable based on the domain name. She is explained by Fierce how she should first use Syntax to create a new variable with the domain name in it.

SPSS for beginners
User somaye has a big SPSS file (over two million records). He can only see the values of the first 100 records, the values for the other records are just question marks. Even when computing a new variable, the values in that variable are mostly question marks. He asks for advice. Fierce explains him that this might be possible because the computer / SPSS are still loading. As long as he can run analyses he should not worry. The analyses will work. If that isn’t working as well, SPSS support should be contacted.

As you can see a lot of questions are solved in the forum, on many different subjects in SPSS. If you need any help, or if you want to help out others yourself, please join in on http://www.spssforum.com!

Entry Filed under: SPSSforum.com updates


April 2009
« Mar    

Most Recent Posts