Talk:Survey methodology/Archives/2012

Latest comment: 11 years ago by Nancydarling in topic Issues


Response rates

The response rates suggested in earlier versions of this article were very generous and not what is typically achieved in the field, using the AAPOR definitions of response rate.

  • The response rates vary wildly accross the different countries. In Finland, where I work as a statistical interviewer in Statistics Finland[[1]], we consistently achieve 70% response rate in telephone surveys with consumer respondents and over 90% when places of businessess are being surveyed. The relative response rates of different types of surveys stay comparatively same irregardless of country (i.e. surveys via mail get a lower response rate compared to telephone interviews). I'll put a link to those numbers on the talk page, for the case if someone thinks those points are worth being mentioned. It would be highly misleading to only include the situation in USA and to assume it's wide applicibility all over the world. Santtus 10:34, 29 December 2005 (UTC)
Response rates also vary greatly depending on the subject of the survey (e.g. people are often happier to talk about their health than their finances), how well it's designed (longer surveys = poorer response), and whether it's compulsory (as is the case with a lot of government surveys). I don't think giving unqualified response rates in the article is very useful, especially without cites to establish context, so I'll remove these. --144.53.251.2 (talk) 00:29, 3 December 2008 (UTC)

Graduate program listings

There are 3 programs in the US that specifically offer a PhD in "Survey Methdology", though training in survey research techniques/methods is available at others. The programs are at the University of Michigan (Michigan Program in Survey Methdology; http://www.isr.umich.edu/gradprogram/), the University of Maryland (Joint Program in Survey Methodology; http://www.jpsm.umd.edu/) and the University of Nebaska (Survey Research and Methdology Program; http://sram.unl.edu/). Each has a terminal Master's degree also. Michigan and Maryland offer short courses and summer courses, but I'm not sure about Nebraska. Mattjans (talk) 20:00, 7 May 2008 (UTC)


Are there really only a handful of graduate programs in this field as listed in this article? Or is this list a random collection of a handful of programs? If it's only a small sampling, I'd like someone to explain why these particular programs are listed. If there is no other reason other than "These are the ones random Wikipedians have added" then I would like to delete the entire section. It's unencyclopediac and smacks of self promotion on the part of these institutions. --ElKevbo 21:28, 17 May 2006 (UTC)

I believe there are very few graduate programs specifically in survey methodology worldwide. Of course there are many, especially in statistics or the social sciences, that offer a few courses/papers on the subject. From what I could tell (not being able to read Dutch), the Utrecht program was of this sort, so I have deleted it from the list. The one glaring omission (IMO) was Southampton's Social Statistics program, which I have added. -- Avenue 00:05, 18 May 2006 (UTC)
Sorry, it seems I was too hasty about the Utrecht program. Delving a bit deeper, I found these links [2] [3] which make it clear that the program does cover survey methodology in depth. I have restored the program to our list. -- Avenue 01:15, 18 May 2006 (UTC)
Here is a list of relevant degree programs from the American Statistical Association.[4] It has a few more than our list, but not many. The only programs they list from outside the US and Canada are Southampton and Hebrew University. -- Avenue 14:10, 29 May 2006 (UTC)

Can we get some criteria for this list? The Southampton course you listed is in social statistics. Should we also include courses at Essex (home of the UK Data Archive and many surveys and methods courses)[5], and Manchester (home of the Centre for Census and Survey Research[6])? Surrey[7]? City University[8]? All of whom are strong on statistics and surveys, but (like Southampton) don't have particular degrees in Survey. I think the list of degrees in social statistics could be quite long. -- zzuuzz (talk) 14:53, 29 May 2006 (UTC)

The only description of that list I can see is online here:
'A group of faculty involved in the various grad programs in survey statistics and methodology at American Association for Public Opinion Research (AAPOR) assembled a list of "Universities and Colleges with Degree Programs in Survey Methodology."'
I think Southampton's MSc in Official Statistics is clearly relevant, and I think there could be a good case for the Statistics Pathway in their MSc in Social Statistics, but its Research Methods Pathway seems less clear. Of its nine compulsory units, only five seem clearly survey or statistics related from their titles.
I don't think every social statistics program would be relevant to this article, although the Manchester one looks good. But there seems to be a danger of original research creeping in here. Should we decide on some criteria we can use to expand the list, and make it a separate page if it gets too long? Or should we just try to find a European equivalent of the ASA's list and link to both, then delete our list? By the way, there is a similar list in our Sampling (statistics) article. -- Avenue 23:23, 29 May 2006 (UTC)


With no criteria or evidence for why these courses are on the list, this section is blatant advertising; Wikipedia is not a directory of acadmeic courses and unless a course is of paticular note i.e. we have sourced statements saying that it is so, or it is the only one of its kind, a section like this has no place on the article. --Pretty Green (talk) 11:30, 22 February 2008 (UTC)

Attention Editors I've removed this **Example Online Surveys are here. Looks like spam for a commercial survey software. - there is free survey software available. ie surveymonkey.


Page

I combined a few sections that should not really have been separate. The article has been shortened considerably but If you don't want it to look like some ones notes, the page should be re-written. I would do it, but hardly feel qualified.

Mundokiir (talk) 09:50, 14 January 2008 (UTC)

Example

Here's what an actual online research survey looks like. Feel free to complete it if you wish. http://survey.ubcomm.org/surveys/nd_survey_a.html

Purpunknat (talk) 19:57, 9 April 2008 (UTC)

Question wording and ordering contradiction?

"Every respondent should be presented with the same questions and in the same order as other respondents."

Really? In the opinion poll article question rotation and "split-sampling" are mentioned as methods often used to control for bias. Destynova (talk) 15:00, 13 July 2010 (UTC) Agree with you, Destynova. Also, the "Same questions" part isn't possible in moderately complex surveys where sequencing/filtering happens. Jusque (talk) 19:55, 19 January 2012 (UTC)

Issues

The article has been tagged with "rewrite" and "cleanup" for some time now. Are there any specific issues to be dealt with? Please note them here so that someone can think about them. Melcombe (talk) 15:58, 24 August 2010 (UTC)

The "Advantages and Disadvantages of Surveys" section is unclear and lacks citations. I wonder if it belongs at all in an article about methodology.
I'm particularly curious about this bullet point: * When the test is of destructive nature, sampling is only the way out. In such cases the population survey is not possible. In my years as a market researcher, I've never had to destroy a respondent. :)
Katharine908 (talk) 22:53, 17 January 2012 (UTC)
This article is (now) the main article for Category:Survey methodology and so needs to be fairly general (the whole of statistical surveys). Note that this article was recently renamed ... see article history for a reason. On the second point, the article needs to (and the lead section indicates that the article does) cover market surveys but also other types of survey. For example, various types of quality control of products involve inspection steps that would leave product-items unuseable. Melcombe (talk) 23:19, 17 January 2012 (UTC)
IMO, this article should be merged with Survey Research. They completely overlap and both could use material from the other.:::: — Preceding unsigned comment added by Nancydarling (talkcontribs) 21:13, 17 December 2012 (UTC)