Understanding Society User Support: Issueshttps://iserredex.essex.ac.uk/support/https://iserredex.essex.ac.uk/support/support/favicon.ico?15995719382017-10-25T14:46:38ZUnderstanding Society User Support
Redmine Support #869 (Closed): Unexpectedly strong gender homophily in Understanding Society compared wit...https://iserredex.essex.ac.uk/support/issues/8692017-10-25T14:46:38ZTill Hoffmann
<a name="Summary"></a>
<h1 >Summary<a href="#Summary" class="wiki-anchor">¶</a></h1>
<p>I am interested in homophily effects that can be derived from the social network questions in waves B, D, F, H, J, L, N, P, R of the BHPS and waves C and F of Understanding Society. Homophily with respect to occupational status is relatively constant across all waves, and age homophily only changes slightly. However, homophily with respect to gender increases sharply from the last wave of the BHPS (R) to the first wave of Understanding Society (C). Do you happen to have any ideas whether a data error (or misuse on my part) might cause such an effect?</p>
<a name="Details-and-code-for-reproducing-the-observations"></a>
<h1 >Details and code for reproducing the observations<a href="#Details-and-code-for-reproducing-the-observations" class="wiki-anchor">¶</a></h1>
<p>For each wave of the survey I compute the contingency table of the sex of the ego (respondent) and alter (nominated friend) and apply the Fisher exact test (<a class="external" href="https://en.wikipedia.org/wiki/Fisher%27s_exact_test">https://en.wikipedia.org/wiki/Fisher%27s_exact_test</a>) to compute the odds ratio of association and test for statistical significance.</p>
<p>To account for effects of the survey design I use the weight <pre>#xrwght</pre> for the BHPS and <pre>#_indscub_xw</pre> for Understanding Society (the results are similar when omitting the survey weights). The results below include survey weights but the contingency tables have been rounded to the nearest integer. The code to reproduce the results can be found here: <a class="external" href="https://gist.github.com/tillahoffmann/e48dbc3546c418561263d181b2bda694">https://gist.github.com/tillahoffmann/e48dbc3546c418561263d181b2bda694</a></p>
<p>Thank you very much for your help!</p>
<pre>
BHPS wave: b
Contingency table
[[ 8862 3011]
[ 2202 11176]]
Odds ratio: 14.947, p-value: 0.000000
BHPS wave: d
Contingency table
[[ 8435 2817]
[ 2053 11044]]
Odds ratio: 16.116, p-value: 0.000000
BHPS wave: f
Contingency table
[[ 8500 3253]
[ 2195 11149]]
Odds ratio: 13.277, p-value: 0.000000
BHPS wave: h
Contingency table
[[ 8328 3034]
[ 1980 10994]]
Odds ratio: 15.249, p-value: 0.000000
BHPS wave: j
Contingency table
[[ 7936 2935]
[ 1941 10541]]
Odds ratio: 14.697, p-value: 0.000000
BHPS wave: l
Contingency table
[[ 7625 2791]
[ 1888 10157]]
Odds ratio: 14.696, p-value: 0.000000
BHPS wave: n
Contingency table
[[7320 2663]
[1827 9780]]
Odds ratio: 14.713, p-value: 0.000000
BHPS wave: p
Contingency table
[[7099 2597]
[1714 9436]]
Odds ratio: 15.052, p-value: 0.000000
BHPS wave: r
Contingency table
[[ 7682 2910]
[ 1964 10312]]
Odds ratio: 13.865, p-value: 0.000000
Understanding Society wave: c
Contingency table
[[36818 9400]
[ 6777 44476]]
Odds ratio: 25.710, p-value: 0.000000
Understanding Society wave: f
Contingency table
[[30372 7348]
[ 5004 36279]]
Odds ratio: 29.972, p-value: 0.000000
</pre> Support #843 (Closed): w_englang is not part of xwavedat, but available for wave a, e, f https://iserredex.essex.ac.uk/support/issues/8432017-08-26T14:13:20ZNico Ochmannnico.ochmann@postgrad.manchester.ac.uk
<p>Dear Alita,</p>
<p>if I understand the variable description correctly, englang is unfortunately not part of xwavedat. Whether English is first language is asked both immigrants and natives in wave one, and the IEMB in wave 6. <br />For wave 5, the englang variable appears as well. I want to use all six waves in USoc, for the years 2009-2015. For that purpose, I need to impute the information I have on englang to the missing waves/years. <br />So far, I have used the information from all three waves and generated a new variable englang_dv which I then merged m:1 pidp with the rest of the data set. <br />In broad terms, am I doing this correctly? <br />Thanks a lot! <br />Best wishes.</p>
<p>Nico</p> Support #837 (Closed): Industrial Injury Disability Benefit - who is askedhttps://iserredex.essex.ac.uk/support/issues/8372017-08-18T11:34:57ZRay Storryraystorry@incomesdataresearch.co.uk
<p>I have been running a query vis a cross-tabulation between _ff_bentype09 BY w_bendis8 filtering on w_ff_bentype09 = 0 (they said they were not in receipt of the IIDB at the previous interview).</p>
<p>For Waves 3-5 I am getting a total population figure of between 12,700 and c15,000. But in Wave 6 the population jumps to c30K. Why is this the case.</p> Support #804 (Closed): Missing Value Analysishttps://iserredex.essex.ac.uk/support/issues/8042017-06-28T14:10:56ZDaniel Okoli
<p>I'm currently attempting missing value analysis of merged data which comprises HHRESP and INDRESP data from waves 1 - 6. The MVA is stuck on 'running' and has been on that status for several hours. Since this is time consuming, would you advice I run MVA for each wave of data before merging the files?</p> Support #803 (Closed): Tracking parental death https://iserredex.essex.ac.uk/support/issues/8032017-06-26T13:42:12ZEmily Lowthianlowthianem@Cardiff.ac.uk
<p>Hi there,</p>
<p>I'm looking to create a variable of parental death. However, I am having trouble of finding cases where individuals are both parents (using wave 4) to children in wave 6. I currently am using mnpno, fnpno, finloc, mpid and ivfio. However, it seems to be that no parents have deceased?</p>
<p>Any thoughts or guidance to where I could be going wrong, or whether there is no way of coding parental death would be greatly appreciated.</p>
<p>Many thanks,</p>
<p>Emily</p> Support #799 (Closed): Parent sample for research projecthttps://iserredex.essex.ac.uk/support/issues/7992017-06-19T14:39:44ZDaniel Breslindaniel.breslin@actionforchildren.org.uk
<p>Hello,</p>
<p>I was wondering if you might be able to help me. I am trying to get a better understanding of the responders to understanding society to gauge whether it is the right data set to use for a prospective project.</p>
<p>It is possible to know how many respondents to the wave 1 and wave 6 survey are: <br />· Parents<br />· Parents with a child aged 16 or under<br />· Parents who themselves aged 25 or under</p>
<p>Any help you could offer with this query would be greatly appreciated.</p>
<p>Best wishes,</p>
<p>Dan</p> Support #798 (Closed): Variable a_frmnthimp.dv in the protected file a_income_protect.dtahttps://iserredex.essex.ac.uk/support/issues/7982017-06-16T14:30:05ZMyera Rashidmr4028@nyu.edu
<p>In the data file a_income_protect.dta, what is the format of the data for the variable a_frmnthimp.dv? I am asking this as we are getting a mean value of 313 pounds for the total income from source and this seems like a terribly low value for a mean total income.</p>
<p>Any help or explanation would be greatly appreciated!</p> Support #796 (Closed): #794 continuedhttps://iserredex.essex.ac.uk/support/issues/7962017-06-14T09:56:48ZJulia Borodinajul1bor35@me.com
<p>Thanks! So as far as I understood 'feend' gives an age when a respondent is supposed to leave further education while still in it. Is it correct?<br />If it is so, how can I find out the age at which all respondents completed their full-time education. Any help would be appreciated!</p> Support #792 (Closed): Merging datasetshttps://iserredex.essex.ac.uk/support/issues/7922017-06-07T14:21:30ZEmily Lowthianlowthianem@Cardiff.ac.uk
<p>Hi there,</p>
<p>My name is Emily - I am a MSc student, currently working on a dissertation looking at how socioeconomic status moderates the impacts of trauma on adolescent well-being.</p>
<p>I essentially need to merge 2 datasets in wave 4 - d_indsamp and d_indresp. I am using stata V14.2 for this. <br />However, despite trying to use the pidp and d_pno as the unique identifiers Stata keeps flagging up an error that the ID variables do not uniquely identify observations in the master dataset (indsamp).<br />I'm not entirely sure why this is happening - please could you help?</p>
<p>Additionally - I would like to merge on the wave 4 hhresp dataset on to this (I will just duplicate the hh responses on to each indivdiual). Then wave 6 youth dataset will be appended on based using hh ID so the children sit in the parental household and it can be submerged later on.</p>
<p>Any help will be greatly appreciated as I have hit a wall here!</p>
<p>Emily</p> Support #788 (Closed): Why so many "inapplicables" when tabulating on two specific variableshttps://iserredex.essex.ac.uk/support/issues/7882017-05-19T22:14:28ZJohn Haskeyjohn.haskey@uwclub.net
<p>There are two Tables I've derived from Wave 6 of Understanding Society from which I wish to draw some conclusions - but the number of cases which seem to have to be disregarded is so large that I wonder whether I can draw any conclusions at all from the small number of cases in each of the cells in which I’m interested. The two key questions which are the bases of these Tables are:</p>
<p>1. Likelihood of marrying partner in future, f_ncrr12 ;<br />2. Intention to live together within three years, f_ncrr11</p>
<p>I've tabulated each of these two variables by age, by sex and by de facto marital status. (I can send these 2 tables as an Excel file - I've tried adding them here but they come out distorted.)</p>
<p>The first table contains 26245 cases which are “inapplicable” – out of a total of 28945 ie 91% of the total.<br />And the second Table contains the same number of “inapplicables”, but out of a slightly different total of 28939.</p>
<p>I've written the paragraph below (in a paper I've been drafting), based on the data that seemed useable (i.e. was not "inapplicable"): (I've added the sample numbers concerned in brackets)</p>
<p>"In another survey, Understanding Society, respondents who weren’t married were asked the likelihood of marrying their partner. With only small sample numbers representing respondents in England and Wales, around one half of men and women in their forties (for men 30 out of 56; for women 48 out of 101)said it was likely, or very likely, that they would, but, of those aged fifty or over, only about one quarter of men (50 out of 202), and one eighth of women (22 out of 170) , expressed the same degree of likelihood of marrying their partner. Another question asked whether the respondents intended to live together with their partners within three years; just under two thirds of men (61 out of 99) and women in their forties (111 out of 182) replied Yes, but amongst those aged 50 and over, less than one third thought so (for men 84 out of 293; for women 56 out of 243). These tentative results are consistent with rates of marriage and the prevalence of cohabitation declining with older ages, and with cohabitation being considered more likely than marriage."</p>
<p>I would be grateful for your guidance and advice, as I obviously wish to derive accurate results.<br />Also, I've just heard that the paper has been accepted for publication, so there's all the more reason for checking whether my interpretation is valid.</p>
<p>With many thanks!</p>
<p>John Haskey<br />Dept of Social Policy and Intervention<br />University of Oxford</p>
<p><a class="email" href="mailto:john.haskey@uwclub.net">john.haskey@uwclub.net</a><br /><a class="email" href="mailto:john.haskey@spi.ox.ac.uk">john.haskey@spi.ox.ac.uk</a></p> Support #779 (Closed): BHPS file KLIFEMSThttps://iserredex.essex.ac.uk/support/issues/7792017-05-09T15:20:35ZJulia Borodinajul1bor35@me.com
<p>Dear Support Team,<br />I'm currently looking for a variable that could proxy an individual's working experience. I found the variable 'kleslen' (lenght of employment history spell (months)) in datafile KLIFEMST of BHPS. <br />However, I find it difficult to interpret the fact that PID variable repeats itself and gives different results for 'kleslen'. For example, the first 2 rows in the screenshot attached for pid 952000258 produce different figures for 'kleslen' : 45 and 609.<br />Could you please comment on that because I find it a little misleading.<br />Many thanks in advance!</p> Support #777 (Closed): Clarificationshttps://iserredex.essex.ac.uk/support/issues/7772017-05-08T12:50:15ZNiamh Shortt
<p>Dear Madam/Sir</p>
<p>I would like to enquire about the US survey. I have been exploring the website but would like clarification on a few points, I am in the midst of preparing an ESRC application and want to make sure that I have the details correct!</p>
<p>1. I have area level data that I would like to match to the individual respondents for 2012, 2016 and 2018. Can you please confirm the matching waves to these years? (ie. 2012 = waves 3 and 4?)</p>
<p>2. I am hoping to use questions on smoking and alcohol consumption. Can you please confirm the waves closest to the years 2012, 2016 and 2018 that will have data on alcohol consumption and smoking practices (for youth and/or adults)</p>
<p>3. I will be using data for Scotland only. Can you let me know the sample sizes (households and individuals) for Scotland for the waves in 2012, 2016 and projected sample for 2018?</p>
<p>4. Can you please confirm when 2018 data will be be released?</p>
<p>Many thanks in advance.</p> Support #774 (Closed): Merging BHPS and Understanding Society statahttps://iserredex.essex.ac.uk/support/issues/7742017-05-04T10:51:26ZSonia Turrerosonia.turrero@barcelonagse.eu
<p>Good morning,<br />we are working on our Master project using BHPS panel data, and we wanted to add the Understanding Society database to increase our panel. But we cannot find the way to merge both types of datasets. We though PID was an unique identificator and did not change in the two different surveys, but we have no single individual in wave 2 of Understanding Society that coincides with any individual in any wave 14 - 18 of BHPS. I am sure we are missunderstanding something, so it would be great if we could get some help on this. I am sending a file in which it is explained how to merge them: at the end of page 2 it says: <br /><em>It is very easy to identify the BHPS sample members in Understanding Society as the unique cross-wave person identifiers in the BHPS, pid, are provided with each UKHLS data file (in Wave 2).</em><br />We would be very grateful if we could get some advice on how to keep on with this.</p>
<p>Thank you so much in advance.</p> Support #770 (Closed): Social background indicators/ parents education levelhttps://iserredex.essex.ac.uk/support/issues/7702017-04-26T09:39:18ZNatasha Codiroli Mcmastern.codiroli.14@ucl.ac.uk
<p>Hi Understanding Society Team,</p>
<p>I have a question about possible social background characteristics, particularly x_paedqf x_maedqf which ask for fathers and mothers highest education level. I was hoping to use this in my analysis, however the majority of responses appear to be inapplicable (35,000/50,000 in the first wave), and I cannot work out a reason for this.</p>
<p>Would you be able to help, or point me to a better indicator of background?</p>
<p>Thanks for your help,</p>
<p>Natasha</p> Support #769 (Closed): number of siblings ever hadhttps://iserredex.essex.ac.uk/support/issues/7692017-04-21T13:20:01ZSait Bayrakdarsb2152@cam.ac.uk
<p>Hi,</p>
<p>I see that in BHPS there is a question about the total number of siblings respondents ever had (NSIBS). Is there an equivalent of this question in Understanding Society? If not, which strategy would you recommend to get sibling number for respondents (including those not in the household)?</p>
<p>Many thanks,</p>
<p>Sait</p>