AUSSA 2011 USER GUIDE - Australian Data Archive
Contents
1. amp tempage amp retired gen age gp replace age gp 1 if tempage gt 18 amp tempage lt 29 replace age gp 2 if tempage gt 30 amp tempage lt 39 replace age gp 3 if tempage gt 40 amp tempage lt 49 replace age gp 4 if tempage gt 50 amp tempage lt 64 replace age gp 5 if tempage gt 65 amp tempage lt 105 tab age gp m label define age gp Page 10 1 18 29 2 30 39 3 40 49 4 50 64 5 654 label values age gp age gp EDUCATION ATTAINMENT categories match ABS ones capture drop edimp gen edimp replace edimp 1 if highed replace edimp 2 if highed replace edimp 3 if inlist highed 3 4 replace edimp 4 if inlist highed 5 amp h6 replace edimp 5 if inlist highed 5 amp inlist h6 1 2 3 4 5 label define edimp 1 Year 12 modify label define edimp 2 Year 12 modify label define edimp 3 Certificate or diploma modify label define edimp 4 Bachelors modify label define edimp 5 Postgraduate modify label values edimp edimp 1 Give each person with missing education the most common occupational group irrespective of tempage tab edimp occupation l1 col nofreq These are the managers 5 changes replace edimp 3 if occupation l1 1 amp edimp These are the professionals 4 changes replace edimp 5 if occupation 11 2 amp edimp These are the technicians until sales workers 22 changes2. KKK KK KKK Weight for age group gender and education attainment preserve capture drop gender clonevar gender hl Gen tempage group capture drop tempage gen tempage ag replace tempage if tempage 1 If age is missing estimate based on average for that sex retired capture drop retired gen retired replace retired 1 if h19 place retired 0 if h19 6 amp hl9 an tempage over gender retired place tempage 48 if gender 1 amp retired 0 amp tempage place tempage 73 if gender 1 amp retired 1 amp tempage Kh 3 Kh H place tempage 47 if gender 2 amp retired 0 amp tempage replace tempage 71 if gender 2 amp retired 1 amp tempage If retirement missing use partner s retirement gen pretired replace pretired 1 if h28 mean tempage if pretired 1 over gender replace tempage 72 if gender 1 amp pretired 1 amp tempage replace tempage 69 if gender 2 amp pretired 1 amp tempage Otherwise just use average mean tempage over gender replace tempage 56 if gender 1 amp tempage replace tempage 54 if gender 2 amp tempage mean tempage over retired replace tempage 48 if gender amp tempage amp retired 0 replace tempage 71 if gender amp tempage amp retired 1 mean tempage replace tempage 55 if gender
3. The Law and Authority AuSSA 2011 also includes demographic and behavioural categories Personal Background that survey sex year born income education employment union membership country of birth household composition and religion There are also questions about the partner of the respondent employment highest level of education and income Page 2 Sample information Over 6 000 individuals were sampled from the 2011 Electoral Roll Within each state individuals were chosen using systematic random sampling The table below displays some information regarding the response rates and the dates of fieldwork To minimise sample losses several actions were taken One week before the guestionnaires were mailed out a pre letter introducing the survey was sent Around one and a half weeks after the guestionnaires were mailed a reminder postcard was sent Replacement guestionnaires were sent to respondents who had misplaced or lost their original guestionnaires but were willing to participate in the survey A second reminder postcard was sent approximately one month after the first mail out of the guestionnaires TABLE 1 SAMPLE INFORMATION AND FIELDWORK DATES AuSSA 2011 Sample size 6 250 Final number of respondents 1 946 AAPOR Response Rate 1 31 96 Cooperation Rate 1 68 96 Contact Rate 1 46 96 Less than half of questionnaire filled in 2 Cannot complete due to language 5 Blank questionnaire returned 794 Known respondent refusal 3
4. replace edimp 3 if inlist occupation 11 3 4 5 6 amp edimp These are the lower occupations 8 changes replace edimp 1 if inlist occupation 11 7 8 amp edimp 2 For people with unstated education and occupation give lowest level of education replac dimp 2 if dimp amp h5 1 replac dimp 1 if dimp 3 More treatment of missing data below after weights are tab edimp m Set up matrices for population and sample counts Rows i age gp 1 5 then gender 1 2 Cols j edimp 1 5 gen rindex 2 age gp 1 gender gen cindex edimp mat input ABS 277721 479757 419457 168877 24134N 205638 486886 381275 252856 39508 207646 198392 441204 213950 81514 200797 227990 359292 307909 105931 325065 154443 486846 179355 921381 We separate bachelor and postgraduate qualifications because we can although that means combining ABS categories but we combine trade certificate diploma because we are uncertain that AuSSA for their them the calculated Page 11 406376 200985 376797 228709 102908 511216 203430 626392 212093 118190 766073 216740 414506 230665 117222 506622 122333 372363 87074 43686 878947 155307 162661 83535 27974 matrix list ABS Construct an equivalent matrix of sample values tab gender edimp if age gp 1 matcell N1 tab gender edimp if age gp 2 matcell N2 tab gender edimp if age gp 3 matcell N3 tab gender edimp if age g
5. 7 Respondent unavailable during field period 15 Death 16 Physically or mentally unable incompetent 34 Addressee Not Known at Place of Address 515 Nothing ever returned 2881 Duplicate listings 5 Dates of fieldwork Pre letter sent 1 May 12 Questionnaire mailed out 8 May 12 Reminder postcard 15 May 12 Replacement questionnaires mailed out 30 May 12 Second reminder postcard sent 6 Jun 12 Close 8 Aug 12 Page 3 Missing codes All missing codes in the dataset are assigned to negative numbers The following missing codes are used 1 Missing Used when the respondent was did not provide an answer to a guestion 2 Skip Used when the respondent was not asked a particular guestion because it did not apply to them and they had therefore been filtered away from it 4 No partner Used in guestions relating to partner s characteristics in cases where the respondent did not have a partner A respondent was classified as not having a partner if they were not living with a partner or if the question on living with a partner was missing h20 5 Multiple response Used when the respondent was meant to give one response only but instead gave multiple responses For example in variable a7 respondents were asked to choose who they thought was the most important issue for Australia today People who gave multiple responses chose more than one option e g both health care and education 6 Illegible or impossible value Used when the responden
6. AUSTRALIAN SURVEY OF SOCIAL ATTITUDES 2011 USER GUIDE This purpose of this user guide is to give a brief explanation of the Australian Social Survey AuSSA 2011 dataset Contents Abstract EE 2 sample information eek a susu ee 3 Missing COC ES i Li eon e eka ES Ee p et Rude 03 4 Data quality 4 bi Nana 4 Weighting 3 menata aan asas sa ae Nan anemia 6 Citing the SUrV ey da e ege NN MI a E bia BE 9 APPENDIX A STATA code used to generate weights o oooocoooo 10 Page 1 Abstract The 2011 Australian Survey of Social Attitudes AuSSA is the fifth in a biennial series that studies social attitudes and behaviour of Australian citizens for the Australian and international research community AuSSA provides cross sectional data on the social attitudes and behaviour of Australians repeating a core questionnaire for each cross section and fielding specific modules relevant to the changing needs of the social research community AUSSA is Australia s official survey in the International Social Survey Program and regularly includes ISSP modules AUSSA 2011 includes both the ISSP s Environment III 2010 and Health 2011 modules The 2011 Survey includes attitudes and behaviours that are organised into seven categories Issues facing Australia today Environment Health Government Policy and services Kindness Australians in History and
7. e was created by excluding those who did not have Australian citizenship this included those who did not state their citizenship Individuals with postgraduate degrees and graduate diplomas were combined into one category Individuals with certificates or diplomas advanced diplomas were also combined into one category Those who were recorded as having their highest non school qualification as inadequately described not stated or not applicable were assumed not to have a post school qualification and were coded according to their highest level of school education Those who had had completed Year 12 were coded as having completed high school education Those who completed up to Year 11 or below included those who did not state their highest level of school education were coded as having not completed high school Page 6 Constructing sampling weights The sampling weights are designed to rebalance the sample so the weighted sample freguencies are egual to the expected freguencies in the three way tabulation by age group five groups sex and educational attainment five levels Only 1 834 of the 1 946 observations had complete information on all three of these variables Imputation of missing data was possible in some cases however Where educational attainment was missing but occupational category was available for the purposes of constructing the weights people were given the most common education level for their occupation irrespective
8. eek 11 9 2 or 3 times a month 3 3 Once a month 3 2 Several times a year 12 12 Once a year 7 7 Less frequently than once a year 17 16 Never 42 44 Missing 3 6 Total 100 100 Total may add up to more than 100 due to rounding For other variables more closely related to age sex or highest education the effect of weighting is more noticeable as with the example of home ownership H46 Own or rent dwelling Unweighted Weighted Own outright 46 34 Own paying off mortgage 31 34 Rent from private landlord 11 12 Rent from public housing authority 3 3 Other boarding living at home etc 6 12 Missing 3 6 Total 100 100 Total may add up to more than 100 due to rounding Page 8 Citing the survey To cite the survey please use the following Evans A Australian Survey of Social Attitudes 2011 The Australian National University Australian Demographic amp Social Research Institute 2011 For further information regarding the survey please visit the website http aussa anu edu au index php Page 9 APPENDIX A STATA code used to generate weights KK KKKEKKEKEKKEKEEKEKEKEKEKKEKEKKEEKKEK KEKE KK KK KEK KEKE KEK KK KK KKK KKK KKK Name weight final do Purpose Sampling Weights for AuSSA 2011 Author Anna Reimondos adapted from Trevor Breusch E weights used in 2003 AuSSA Date November 2012 Stata version 12 1 KKKK KKK KKK KKK KKK KKK KKK KKK KKK KKK KK KKK KKK KKK KKK KK KKK
9. ght tab weight aw weight gender m gender edimp m aw weight m edimp aw weight m table age gp edimp capture drop gender tempag edimp edimp edimp edimp edimp if if if if if age gp age gp age gp age gp age gp contents mean weight by gender keep formid weight save weight dta replace age gp educ dimp rindex cindex sumw Page 13
10. of their age When age was missing in some cases this was imputed from other information available in the questionnaire If there was information on the respondents main activity they were given the average age of people with the same activity e g if the respondent was retired they were given the average age of retired people For people who were missing on gender they had a weight allocated to them that was the average of the weights over gender for their education and age To avoid extreme weights the weights were trimmed at the 1 per cent and 99 per cent level before being rescaled so that they averaged to one across all cases in each subsample The average calculated weights are shown in the following table Highest education Not Complete Postgraduate complete high Certificate Bachelor degree or Sex and age high school school or diploma degree diploma 18 29 Male 5 73 3 92 1 94 1 56 0 84 Female 5 73 5 73 1 20 1 21 0 91 30 39 Male 5 73 2 75 2 36 1 41 0 67 Female 5 57 3 52 1 35 1 53 0 47 40 49 Male 1 80 2 68 1 06 0 96 0 35 Female 2 69 1 86 0 84 0 81 0 53 50 64 Male 1 08 1 01 0 73 0 87 0 33 Female 1 24 1 43 0 48 0 58 0 21 65 and over Male 0 53 0 68 0 61 0 48 0 16 Female 0 89 1 08 0 29 0 53 0 15 Page 7 The effects For most variables the effect of weighting is only moderate H31 Frequency of attending religious services Unweighted Weighted Several times a week or more often 2 2 Once a w
11. on l4 Industry industry 11 industry I2 industry I3 industry 14 pindustry 11 pindustry I2 pindustry I3 pindustry 14 Industry of respondent s place of work Industry of partner s place of work ANZSIC Australian and New Zealand Standard Industrial Classification 2006 Revision 1 0 Page 5 Weighting Responses were biased toward women the more highly educated and older people To correct for these biases the AuSSA 2011 sample has been benchmarked against the Census 2011 data on Australian citizens following the procedure and the STATA code used by Trevor Breusch 2004 for the 2003 AuSSA data The weighted sample is designed to have the same proportions as the Census in a three way cross tabulation of age sex and educational attainment AuSSA 2011 Census Sample Expected Age 18 29 8 20 30 39 10 17 40 49 16 19 50 64 34 25 65 30 18 Missing 2 Total 100 100 Sex Male 47 49 Female 52 51 Missing 1 Total 100 100 Highest education Less than Year 12 23 32 Year 12 8 18 Trade certificate diploma 33 30 Bachelor degree 15 15 Postgraduate degree or diploma 16 6 Missing 4 Total 100 100 The cross tabulation by age sex and highest education level from the Census was done using the 2011 Census Table Builder A table was created using sex and age single years in the rows and Australian Citizenship Highest level of education and highest year of school completed in the columns The final tabl
12. p 4 matcell N4 tab gender edimp if age gp 5 matcell N5 mat R N1 N2 N3 N4 N5 matrix list R Calculate totals of all cells in matrices Sca sumABS 0 Sca sumR 0 forval i 1 10 forval j 1 5 sca sSumABS sumABSt tABS 1 j3 sca sumR sumR R i j Weight is ratio of population relative cell count to sample relative cell count gen weight forval i 1 10 forval j 1 5 qui replace weight ABS i j sumABS R i j sumR Xf rindex i amp cindex j amp R i j 0 More treatment of missing data 3 For those with missing gender give them the average weight for their other characteristics replace rindex 2 age gp 1 1 if gender forval i 1 2 10 forval j 1 5 replace weight ABS i j sumABS R i j sumR ABS i 1 j sumABS R i 1 j sumR 2 if rindex i amp cindex j amp weight Weights should average 1 0000 summ weight summ weight detail Trim at 1 and 99 percentile replace weight 5 9 if weight gt 5 9 replace weight 0 15 if weight lt 0 15 sum weight sum weight detail gen sumw sum weight replace weight weight 1946 sumw summ weight label var weight Sampling weight Display the weights by gender education and tempage group Page 12 tab tab tab tab tab tab gender s gender s gender s gender s gender s age gp age gp tab weight tab weight tab weight tab wei
13. t did provide some information but it was too faint to read or it was an impossible value For example a response of 300 for height c28 The purpose of this variable is to differentiate cases were something was written from those with unit non response 1 missing 7 Don t know This was used in a few instances where the respondent had written a note on the questionnaire indicating they did not know Also used for the height and weight variables c28 amp c29 Data quality Variables h42 and h43 income should be used with caution A large number of respondents had difficulty interpreting and answering these questions leading to poor quality of data Page 4 Australian Bureau of Statistics ABS classifications The following ABS classifications were used for selected variables J Variables ABS coding standard h360 Respondent country of birth h370 Mother s country of birth h38o Father s country of birth SACC Standard Australian Country of birth Classification of Countries 2011 ASCCEG Australian Standard Classification of Cultural and Ethnic Groups 2011 h39oa Ancestry other 1 Ancestry h39ob Ancestry other 2 occupation l1 occupation I2 Respondent s Occupation occupation I3 occupation l4 occupation ANZSCO Australian and New Zealand Standard Classification of poccupation I4 Partner s Occupations First Edition Revision 1 poccupation I2 occupation poccupation I3 poccupati
Download Pdf Manuals
Related Search