Date:         Wed, 2 Oct 2002 02:41:41 -0500
Reply-To:     Ben Earnhart <bearnhar@blue.weeg.uiowa.edu>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Ben Earnhart <bearnhar@blue.weeg.uiowa.edu>
Subject:      Re: your mail
Comments: To: rangarajan <b.rangarajan@pgrad.unimelb.edu.au>
In-Reply-To:  <4.1.20021002155634.009833f0@cassius.its.unimelb.edu.au>
Content-Type: TEXT/PLAIN; charset=US-ASCII

Get to know your data.  Odds are the median value for the tenth percentile
really *is* 0, especially becuase judging by your variable names, it looks
like you're getting frequencies on wages.  Read up on "truncation"
"censored data" and "floor effects" and it will make a lot of sense; basic
idea is that since the value of the variable can't go below 0, you'll have
a bunch of people clustered there.

On Wed, 2 Oct 2002, rangarajan wrote:

> Hi:
> When i ran following syntax, i always get  the median value as zero at the tenth percentile. How could i f fix this. Hope
> somebody can guide me.
>
> FREQUENCIES
>   VARIABLES=hwage85 hwage86 hwage87 hwage88 hwage89  /FORMAT=NOTABLE
>   /NTILES= 10
>   /STATISTICS=STDDEV MEAN MEDIAN
>   /ORDER  ANALYSIS .
>
>
>       Frequencies
>
>
>

*==========================================================*;
* Ben Earnhart                                             *;
* Department of Management Information Sciences (student), *;
* Department of Sociology (work)                           *;
* University of Iowa                                       *;
* (319) 335-2887                                           *;
*==========================================================*;