Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: SE and CI by mrtab
From
Nick Cox <[email protected]>
To
[email protected]
Subject
Re: st: SE and CI by mrtab
Date
Tue, 15 May 2012 10:29:16 +0100
I made a request for pertinent literature references twice and Steve
underlined it. Evidently you have none to give.
It's not much of a recommendation for SPSS that it can supposedly
deal with this but you can't say what it does! This is important to
you because whatever you are doing -- a thesis, a paper, a report, a
presentation -- somebody may ask "How are those confidence intervals
calculated precisely?"
My wild guess (I've not used SPSS for some decades and I am not going
back even to look) is that SPSS just lets you pool several response
variables with similar possible values and collapse them into one
table. Continuing that wild guess, I imagine that it then treats them
as if they were a standard categorical set-up without special
attention to their multiple response character. (You can that in Stata
easily enough.)
See my reply to Steve just sent.
Either way, you may need to give details of what SPSS command or
function you used to get better comments out of any people on this
list who are well-informed about SPSS.
Nick
On Tue, May 15, 2012 at 6:28 AM, Abu Camara <[email protected]> wrote:
> I could not find such in SPSS as well.
On 15 May 2012 00:43, Steve Samuels <[email protected]> wrote:
>> Each "percentage" has the form P = (mentions of category X)/(number of mentions). áNumerator and denominator are random for each person, so the percentages are actually
>> ratios:
>>
>> ******************************
>> áuse http://fmwww.bc.edu/RePEc/bocode/d/drugs.dta, clear
>> ámrtab inco1-inco7, include title(Sources of income) width(24)
>> áegen sumi = rowtotal(inco*)
>> áratio inco1/sumi
>> *****************************
>>
>> Since Abu is knowledgeable about SPSS, I'd appreciate a reference to the confidence interval formulas that SPSS uses when percentages add to more than 100%. á(I couldn't find one in the SPSS 16 algorithms manual.) áI'd appreciate it also if he would compare the calculation above to the one that SPSS reports.
On May 14, 2012, at 8:40 AM, Nick Cox wrote:
>> If a program is counting mentions, they are not people. Either way, I stand by what I said. I don't think even "sample size" is well defined for such data, so I don't see how inference is well defined.
>>
>> I can't comment on what SPSS does, but I repeat my request. I would be grateful for literature references showing that SPSS, or anybody else, really has a solution for this problem. Just counting mentions regardless of where they come from sounds somewhere between dubious and fallacious to me.
>>
>> The author of -mrtab- is Ben Jann, who is not a member of Statalist. If you want his answers, you need to write to him directly.
Abu Camara
>> Thanks Nick. Consider the table below and you want to get the "se"
>> & "ci"for the responses variable which are in percentages. I was able
>> to do this for
>> other survey questions which are not multiple responses. Perhaps the
>> author might consider
>> including standard errors & confidence interval generation in his
>> program. I will have to turn to
>> SPSS which has the facility.
>>
>> --------------------------------------------------------------------------------------------------------.
>>
>> mrtab inco1-inco7, include title(Sources of income) width(24)
>>
>> Pct. of á á Pct. of
>> Sources of income á á á Freq. á responses á á á cases
>> -------------------------------+-----------------------------------
>> inco1 á á á á áprivate support á á á á 226 á á á 12.83 á á á 23.25
>> (partner, family,
>> friends)
>> inco2 á á á á á public support á á á á 607 á á á 34.47 á á á 62.45
>> (unemployment insurance,
>> social benefits)
>> inco3 á á á á á á drug dealing á á á á 293 á á á 16.64 á á á 30.14
>> inco4 á áhousebreaking, theft, á á á á á50 á á á á2.84 á á á á5.14
>> robbery
>> inco5 á á á á á á prostitution á á á á á82 á á á á4.66 á á á á8.44
>> inco6 á á á "mischeln"/begging á á á á 151 á á á á8.57 á á á 15.53
>> inco7 á á á á legal occupation á á á á 352 á á á 19.99 á á á 36.21
>> -------------------------------+-----------------------------------
>> Total á á á á1761 á á á100.00 á á á181.17
>>
>>
>> On 14 May 2012 14:35, Nick Cox <[email protected]> wrote:
>>> I don't really have further comments. I was half-assuming that you know exactly what you seek, but if so you are not spelling it out.
>>>
>>> As I see it, you would need to specify what data generation process you expect to apply and e.g. how confidence intervals are to be defined and calculated.
>>>
>>> For example, if the question is mode of transport to work and the answers look like
>>>
>>> Car
>>> Car, train, walk
>>> Walk
>>> Yak
>>> Horse
>>> Camel
>>> Personal helicopter
>>> ...
>>>
>>> it is not clear to me what meaning there could be to a standard error around the percent of people who say "walk". If the principle is that people can specify a variety of answers, the associated data generation process seems elusive to me. You can always count "mentions" rather than "people" but the inference for that I don't think is obvious.
>>>
>>> So, I don't think you can blame Stata for neglecting this area unless you can point to literature in which the logic is explained.
Abu Camara
>>> Thanks for the reply.
>>> I have no idea of writing my own program for "mrtab" to compute "se" &
>>> "ci". Further help/suggestion would be appreciated.
>>> Official Stata appears to be weak in complex tabulation.
>>> Abu.
>>>
>>> On 14 May 2012 12:18, Nick Cox <[email protected]> wrote:
>>>> SJ-5-1 ást0082 á. . . . . . . . . . . . . . . Tabulation of multiple responses
>>>> á á á á(help _mrsvmat, mrgraph, mrtab if installed) á. . . . . . . . áB. Jann
>>>> á á á áQ1/05 á SJ 5(1):92--122
>>>> á á á áintroduces new commands for the computation of one- and
>>>> á á á átwo-way tables of multiple responses
>>>>
>>>> You are correct, I think. -mrtab- doesn't provide these, so you may
>>>> need to write your own program.
>>>>
>>>> Nick
>>>>
>>>> On Mon, May 14, 2012 at 10:09 AM, Abu Camara <[email protected]> wrote:
>>>>
>>>>> I am running one and two way tables of multiple response using the
>>>>> user-written command "mrtab" (Stata 11.2). I tried to generate both
>>>>> standard errors and
>>>>> confidence intervals for tables of percentages but I could not find
>>>>> this as an option.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/