Microdata User Guide - University of Waterloo
Contents
1. w ww wwwemmwemnwmmwmmwa 28 7 0 BF ea BAS KA NAE iia en el ena 29 7 1 Response Rates aueia E e E ES ESOT ini TETTE tS Eh 29 Id SUPVEY Biti eee hada hak Make lei chsh ia dad stad sh choked abohathobetel habench cas 31 8 0 Guidelines for Tabulation Analysis and Release cccccecesseeeseeeeeseeeesteeenseeeees 32 8 1 Rounding Guide eceeccccessccssssceessceessscesseeeeseeeesteeeeseeeenseeeeeeeesseeeeseeeeeaeees 32 8 2 Sample Weighting Guidelines for Tabulation wwwwmwwwo 32 8 3 Definitions of Types of Estimates Categorical and Quantitative 33 8 3 1 Categorical Estimates 0 cccccccscccssscceeseccsssecesseeeesseseeseeeeeseeeesseeenseeeeeaes 33 8 3 2 Quantitative EU maini haseded hasered shed akaka kaka kaka bakaba kabsa 33 8 3 3 Tabulation of Categorical EstimateS lt w m wmmewemwmmnemwnwe 33 8 3 4 Tabulation of Quantitative Estimates cccccccsssccesessecceeessseeeeessseeees 34 8 4 Guidelines for Statistical Analysis ccccccescccsssccesseeeeseeesseeeesteeeeseeesseeeee 34 8 5 Coefficient of Variation Release Guidelines lt w wemewmeea 35 8 6 Release Cut off s for the 2004 2005 Youth Smoking Survey Data 37 2004 2005 Youth Smoking Survey User Guide 1 0 Introduction The Youth Smoking Survey is undertaken with the cooperation and support of Heal2. Definitely yes friends was to offer 2 Probably yes you a cigarette would 3 Probably not you smoke it 4 Definitely not 99 Not Stated Y_ql6 Atanytime during the 1 Definitely yes next year do you think 2 Probably yes you will smoke a 3 Probably not cigarette 4 Definitely not 99 Not Stated e IfY ql4 5 and Y_ql5 4 and Y_q16 4 then D SUSCEPTIBLE 1 e If Y_ ql4 1 2 3 or4 And Y ql5 1 2 or 3 And Y_q16 1 2 or 3 then D_SUSCEPTIBLE 2 e IfY ql4 99 Or Y_ql5 99 Or Y_ql6 99 then D SUSCEPTIBLE 99 DVSELF The objective of this variable is to measure the student s overall self esteem and is based on the following item Y_q8 Choose the answer that best describes how you feel a In general I like the way I am b Overall I have a lot to be proud of c A lot of things about me are good d When I do something I do it well e I like the way I look 1 False 2 Mostly False 3 Sometimes False Sometimes True 4 Mostly True 5 True 2004 2005 Youth Smoking Survey User Guide 17 The above scale for the 5 parts of Y 98 was recoded as follows 0 False 1 Mostly False 2 Sometimes False Sometimes True 3 Mostly True 4 True Then the scores were added up and averaged across the questions that were answered by the student giving an overall score for variable DVSELF The next five derived variables use the following questions from the survey Think back over the last 7 days F
3. Non Smoker Has smoked fewer than 100 cigarettes in his her lifetime Experimental smoker beginner Has smoked between 1 and 99 cigarettes in his her lifetime and has smoked in the 30 days preceding the survey Past experimenter Has smoked between 1 and 99 cigarettes in his her lifetime but has not smoked in the 30 days preceding the survey Puffer Has smoked less than one whole cigarette in his her lifetime but has tried smoking Never tried Has never tried smoking not even just a puff The YSS collects data from students in grades 5 through 9 or approximately ages 9 to 14 Given that smoking rates as they are typically characterized are very low in this age group users may want to consider focusing on indicators such as ever tried and susceptibility refer to section 6 4 for a derived variable for susceptibility 2004 2005 Youth Smoking Survey User Guide 3 4 0 Survey Methodology The 2004 2005 Youth Smoking Survey YSS was administered to a sample of children in grades 5 through 9 primary 5 6 and secondary I to III in Qu bec by sampling from all public and private schools in Canada 4 1 Population Coverage All 10 provinces were selected to participate in the study by Health Canada Sampling frames for each province began with a list of all school boards alternatively called school divisions and school districts in each of the provinces Each provincial sampling frame consisted of a range of info
4. Manitoba 32 28 4 88 Saskatchewan 32 22 10 69 Alberta 60 30 30 50 British Columbia 115 32 83 27 Canada 513 281 232 55 2004 2005 Youth Smoking Survey User Guide Table 3 Sample Size by Grade Grade5 Grade6 Grade7 Grade8 Grade 9 Total Newfoundland and Labrador 54 561 59 41 423 2518 Nova Scotia 517 622 590 530 523 2782 Prince Edward Island 656 781 394 360 294 2485 New Brunswick 420 588 603 580 390 2581 Qu bec 851 915 731 743 404 3644 Ontario 784 817 834 800 915 4150 Manitoba 619 706 610 638 437 3010 Saskatchewan 379 378 397 330 562 2046 Alberta 606 604 620 464 331 2625 British Columbia 505 645 596 988 668 3402 Canada 5884 6657 584 584 447 29243 2004 2005 Youth Smoking Survey User Guide 7 5 0 Data Collection Data collection was conducted from February to June 2005 with school board and school recruitment beginning in October 2004 Students were surveyed in their classrooms Parent interviews were completed by telephone within 3 weeks following the school data collection 5 1 Questionnaire Design Several key considerations guided the design of the student guestionnaire e Comparability the basis of the questionnaire was the 2002 YSS with most items unchanged to allow for comparisons with the 2002 and 1994 data e Responsiveness to meet the needs of users of the data provincial collaborators were given an opportunity to contribute topics items e Relevance to ensure value added for participating schools items importan
5. There were several ways in which the actual grade of the student was determined The primary method was to check the grade assigned on the student consent form If this grade was an appropriate grade then this value was used as the student s actual grade If the grade on the student s consent form was inappropriate or missing then an attempt was made to link the student s identification number to the questionnaire data to see what the student filled out on the questionnaire If it was not possible to link to the questionnaire data then the actual grade was considered to be the grade assigned to the class Once the actual grade was determined a table was generated to compare the actual grades to the grades assigned to the student s class Adjustments were then made to the data by grade for the number of students enrolled the number of returned consent forms and the number of positive consents where appropriate During the process of determining consent form return rates and consent to participate rates several students had filled out a grade on the survey that was inconsistent with the grades represented in the school Ifa student filled out a grade that did not match the relevant grades in the school it was recoded to match the nearest relevant grade 17 records were so recoded W is then calculated for each grade in each school as W Number of eligible students with consent number of eligible students Number of eligible stu
6. interviewers conducted the brief interview usually within 3 weeks of the student data collection 2004 2005 Youth Smoking Survey User Guide 9 Interviewers made up to 25 call backs over a 3 to 4 week period and provided a 24 hour toll free number for those who wished to call at a time convenient to them 2004 2005 Youth Smoking Survey User Guide 10 6 0 Data Processing The main output of the Youth Smoking Survey YSS is a microdata file This chapter presents a brief summary of the processing steps involved in producing this file 6 1 Data Capture Student questionnaires were machine scanned using Optical Mark Read OMR technology Several quality control measures were used to ensure the accuracy of the scan data First all questionnaires were visually scanned and marks that were too light or incomplete e g check marks instead of filled in circles were darkened to ensure that they would be recognized by the scanner During this process the perforated booklets were separated and oriented in preparation for the OMR scan Secondly standards were used to ensure that the calibration of the scanner remained constant Finally all bundles of questionnaires were scanned twice and discrepancies were investigated Staff were trained to make decisions according to strict criteria For example they had to distinguish between true uncodeable responses not to be corrected e g where the respondent chose two answers and those which were ma
7. were then rank ordered based on their adult smoking rates and each board was assigned to one of two strata so that approximately half the total student enrollment in any province was assigned to each stratum The boards with the higher smoking rates formed the Upper Stratum and boards with the lower smoking rates formed the Lower Stratum Within each stratum in each province boards were randomly selected with probability proportional to the total enrollment in the board in the numbers described below After a first set of boards was selected a second set of substitute boards was selected in case any board refused to participate 2004 2005 Youth Smoking Survey User Guide 4 Stage 2 Selection of Schools not including private or independent schools Within each selected board schools were stratified into two strata The Senior stratum contained schools with students in the senior elementary or high school grades specifically schools with grades 5 8 5 9 6 8 6 9 7 7 8 7 9 8 and 9 The Junior stratum contained those schools with students in grades 5 6 5 6 5 7 and 6 7 To accommodate the larger number of students in Senior stratum schools there was an over selection of Junior stratum schools where possible i e the board included adequate schools in the Junior stratum A second set of schools was also chosen to act as substitute schools in case any of the originally selected schools refused to participate Selectio
8. 7 days in row 2 No 2004 2005 Youth Smoking Survey User Guide 14 5 Experimental Smoker Beginner An experimental smoker is a person who has smoked in the last 30 days but has not smoked 100 or more cigarettes Y g18 Have you ever smoked a whole cigarette 1 Yes And Y 920 Have you ever smoked 100 or more whole cigarettes in your life 2 No i An Y_q24 On how many of the last 30 days did you smoke one or more cigarettes 2 1 to 5 days or 3 6 to 10 days or 4 11 to 20 days or 5 21 to 29 days or 6 30 days every day 6 Past Experimental Smoker A past experimental smoker is a person who has smoked a whole cigarette but did not smoke in the last 30 days and also did not smoke 100 cigarettes in his her life time Y_q18 Have you ever smoked a whole cigarette Yes And Y_q20 Have you ever smoked 100 or more whole cigarettes in your life 2 No And Y_q24 On how many of the last 30 days did you smoke one or more cigarettes None 7 Puffer A puffer is a person who has tried smoking but has never smoked a whole cigarette Y_q11 Have you ever tried cigarette smoking even just a few puffs 1 Yes And Y_q18 Have you ever smoked a whole cigarette 2 No 8 Never Tried A person classified as never tried has never tried a cigarette not even just a few puffs Y_q11 Have you ever tried cigarette smoking even just a few puffs 2 No 2004 2005 Youth Smoki
9. Before board recruitment began project information packages were mailed to all provincial Ministries of Education In all provinces school boards were approached prior to contact with schools A standard recruitment package included an invitation letter a project summary sample questionnaires sample consent letters and forms and a template feedback report In addition formal application forms and procedures were adhered to as required by individual boards Active parental consent was required for participation in the student survey A parent information letter and consent form was sent home with students Parent information letters provided details about the project contact information for project staff and referral to the website for further details including copies of the questionnaires Ifa child s class was selected to participate in the parent interview additional explanation and consent fields were included in the package Parents were given a minimum of two weeks to return consent forms To improve consent form return rates schools chose to resend consent materials conduct phone follow up to parents and or provide verbal or written reminders to students Site coordinators worked with a school contact to arrange data collection at each school School contacts were asked to provide a list of classes for the eligible grades that included teacher name course name and or the classroom number grade room number optional and the numb
10. Y_q23sun to Y_q23sat were coded as 96 Valid Skip if Y_q18 Have you ever smoked a whole cigarette was 2 No or 96 Valid Skip Y 924 On how many of the last 30 days did you smoke one or more cigarettes and Y_q25 Thinking back over the last 30 days on the days that you smoked how many cigarettes did you usually smoke each day are only relevant if the respondent had smoked a whole cigarette Therefore Y_q24 and Y 25 were coded as 96 Valid Skip if Y_q18 Have you ever smoked a whole cigarette was 2 No or 96 Valid Skip Y_q26 What brand of cigarettes do you usually smoke is only relevant if the respondent had tried smoking cigarettes Therefore Y_q26 was coded as 96 Valid Skip if Y_q11 Have you ever tried cigarette smoking even just a few puffs was 2 No Y_q27ato Y_q27j Why do you smoke the brand of cigarettes that you do and Y 928 During the past 12 months have you switched cigarette brands are only relevant if the respondent reported that they have a usual brand Therefore Y_q27a to Y_q27j and Y_q28 were coded as 96 Valid Skip if Y_q26 was 1 I do not smoke 2 I do not have a usual brand 96 Valid Skip or 99 Not Stated 2004 2005 Youth Smoking Survey User Guide 20 Y 929 Where do you usually get your cigarettes and Y g30ato Y_q30f How do you go about buying cigarettes from a store are only relevant if the respondent had tried smoking cigarettes Therefore Y_q29 and Y_q30ato Y_q30f were cod
11. consistent with these established guidelines 8 1 Rounding Guide Users are urged to adhere to the following guidelines regarding the rounding of such estimates a Estimates in the main body of a statistical table are to be rounded to the nearest hundred units using the normal rounding technique In normal rounding if the first or only digit to be dropped is 0 to 4 the last digit to be retained is not changed If the first or only digit to be dropped is 5 to 9 the last digit to be retained is raised by one For example in normal rounding to the nearest 100 if the last two digits are between 00 and 49 they are changed to 00 and the preceding digit the hundreds digit is left unchanged If the last digits are between 50 and 99 they are changed to 00 and the preceding digit is incremented by 1 b Marginal sub totals and totals in statistical tables are to be derived from their corresponding un rounded components and then are to be rounded themselves to the nearest 100 units using normal rounding c Averages proportions rates and percentages are to be computed from un rounded components i e numerators and or denominators and then are to be rounded themselves to one decimal using normal rounding In normal rounding to a single digit if the final or only digit to be dropped is 0 to 4 the last digit to be retained is not changed If the first or only digit to be dropped is 5 to 9 the last digit to be retained is increased by 1 d Sum
12. important psychosocial and environmental factors and rarely address behaviours knowledge and attitudes in a way that allows evaluation of the impact of prevention efforts The main objective of the YSS is to provide benchmark data on national prevalence rates for students in grades 5 through 9 In addition to prevalence rates the YSS also offers a detailed snapshot of purchasing behaviour and information about the effect of continued tobacco marketing This information is critical to assessing the need for increased legislative controls on tobacco and bolstering public support for these policy options Without this type of monitoring the effectiveness of our prevention efforts cannot be gauged The YSS also provides a unique opportunity to advance our knowledge of the psychosocial correlates of smoking behaviour including initiation and cessation as well as individual differences in the influence of tobacco marketing and purchasing controls Questions regarding alcohol and other drug use permit similar reporting on a national basis specific to youth in grades 7 through 9 secondary I to III in Qu bec Finally the collection of data from parents at the same time as youth responses is unique in a national survey of this type The parent data will help in the investigation of social influence on youth smoking behaviour provide socio economic status items and describe potential for exposure to tobacco smoke in private homes and cars Partici
13. questionnaire Parent Interviews In randomly selected classes two per school per grade parents were invited to participate in a brief telephone interview If there were only one or two classes in an eligible grade these classes were automatically selected to participate in the parent interview In some provinces the number of grade 9 classes included in the parent interview condition was increased in order to obtain a satisfactory number of grade 9 parent interviews In Ontario a maximum of 4 grade 9 classes per school were selected and in the provinces of Alberta Qu bec and Prince Edward Island a maximum of 3 grade 9 classes per school were selected 2004 2005 Youth Smoking Survey User Guide 5 4 3 Sample Size The following tables show the number of school boards schools and students sampled for each province More detail regarding response rates can be found in section 7 1 Table 1 Board Recruitment Outcomes Approached Agreed Refused Reais Newfoundland and Labrador 4 4 100 Nova Scotia 5 5 100 Prince Edward Island 2 2 100 New Brunswick 4 4 100 Qu bec 11 10 1 91 Ontario 14 12 2 86 Manitoba 9 8 1 89 Saskatchewan 6 6 100 Alberta 12 8 4 67 British Columbia 19 9 10 47 Canada 86 68 18 79 Table 2 School Recruitment Outcomes Approached Agreed Refused Re Newfoundland and Labrador 24 24 100 Nova Scotia 31 24 7 77 Prince Edward Island 25 24 1 96 New Brunswick 28 20 8 71 Qu bec 60 36 24 60 Ontario 106 41 65 39
14. stage a weight Wi was created based on the board selection scheme In the second stage a weight W2 was developed to account for the school selection A third weight W3 was calculated to adjust for student non response At the second and third stages the survey weights were adjusted to remove undue variability and outlying weights that would have an undue influence on the resulting estimates and their estimated variances The weights were also calibrated to the provincial gender and grade distribution so that the total of the survey weights by gender grade and province would equal the actual enrollments in those groups Finally bootstrap weights were generated to attach to the data file Stage 1 Calculation of W In Newfoundland Nova Scotia and Prince Edward Island all boards were selected and hence their first stage weight W 1 For the remaining provinces all the boards in a given province were divided into two strata based on their sizes and related adult smoking rates see section 4 2 Within each stratum in each province boards were randomly selected with probability proportional to the total enrollment in the board For such a board j W has been computed as Wi 1 7 j where 7 is the probability of inclusion for board j at stage 1 and where except in British Columbia in this expression Mj Total enrollment for board j 1 Number of boards selected in the stratum at the first stage of sampling and L Total number
15. the following references Emmanuelle Pi rard Neil Buckley and James Chowhan Bootstrapping made easy A STATA ADO file The Research Data Centres Information and Technical Bulletin Statistics Canada Spring 2004 vol 1 no 1 pp 20 36 can be downloaded at http www statcan ca english freepub 12 002 XIE 12 002 XIE2004001 pdf James Chowhan and Neil J Buckley Using mean bootstrap weights in Stata A BSWREG revision The Research Data Centres Information and Technical Bulletin Statistics Canada Spring 2005 vol 2 no 1 pp 23 37 can be downloaded at http www statcan ca english freepub 12 002 XTE 12 002 XIE2005001 pdf 2004 2005 Youth Smoking Survey User Guide 35 First the number of respondents who contribute to the calculation of the estimate should be determined If this number is less than 30 the weighted estimate should be considered to be of unacceptable quality For weighted estimates based on sample sizes of 30 or more users should determine the coefficient of variation of the estimate and follow the guidelines below These quality level guidelines should be applied to weighted rounded estimates All estimates can be considered releasable However those of marginal or unacceptable quality level must be accompanied by a warning to caution subsequent users Table 9 Quality Level Guidelines Quality Level of Estimate Guidelines 1 Acceptable Estimates have a sample size of 30 or more and low coefficients of variation
16. Mail Smoking Sunyey Microdata User Guide kef anea Sawa www yss uwaterloo ca Table of Contents Dra SOE Contents akakaa kaaba kakake EASE ana aa EEOAE ana E Oaai i 1 0 Introd ctioMisr renn a ene a a aa ate aaaea 1 2 0 Baki ccctcsrie cc tt te ee tt ie te erin ce cee cin ce cies cer kel 9 2 3 0 Concepts and Definitions ua as vn bu ees Wu bo a ees 3 4 0 Survey Methodology ccccccccsssccesseeesseeceseecesseeeesseeceseeeeseeesseeeseeeeseeeesesesseeeeseees 4 4 1 POpulatiom Coverage cts t teh sk oe Moat hatin natin hal nk oe on eee 4 4 2 Sample DeSronie s cc esieisceisistsvdetstovstsieteieeesfessteletssatesetscotela tose Eo nt irentsia 4 4 3 SAMpleS1Ze wea E E code olotel solos wahaha mhaya ovate oa ce 6 5 0 Data Collecion UA aana ann anaa aaa aa AE ina naed AAA 8 5 1 Questionnaire DSi en eere e rE 8 5 2 Data Collection Protocols ccecscecssscessseceesseceeseeeesecesseeceeseeeeseseeseeseeseeeess 8 6 0 Data Processing WA AA LEE aT 11 6 1 Wata Capture KA 11 6 2 Editing and Imputation cccscccsssscessnecessnecesceessnecensnecsneceesseeeesenersnanens 11 6 3 Coding of Open ended OptionNs sewesemewemenwmmsumewimwwmawemwemwwwa 12 6 4 Creation of Derived Variables ccccccccessseceeseeeeeseeeeseeeeseeceseeesseeeesseeene 12 6 5 Skip T a n MA AAA Aen Ae TAA 19 6 6 Weighing ta eraat anaana a a a a a aa E E a E E A Era o aora ah 23 6 7 Suppression of Confidential Information
17. abrador Nova Scotia Prince Edward Island New Brunswick Qu bec Ontario Manitoba Saskatchewan Alberta British Columbia Canada Approached 4 5 Agreed 4 5 68 Refused 10 18 Response rate 100 100 100 100 91 86 89 100 67 47 79 2004 2005 Youth Smoking Survey User Guide 29 Table 7 School Recruitment Outcomes Approached Agreed Refused pein Newfoundland and Labrador 24 24 100 Nova Scotia 31 24 7 77 Prince Edward Island 25 24 1 96 New Brunswick 28 20 8 71 Qu bec 60 36 24 60 Ontario 106 41 65 39 Manitoba 32 28 4 88 Saskatchewan 32 22 10 69 Alberta 60 30 30 50 British Columbia 115 32 83 27 Canada 513 281 232 55 The third level of response rate is based on individual student consent The response rate at the student level is derived based on the number of eligible students as provided by school contacts for participating classes Non response at the student level can be attributed to several factors Some parents guardians refused to allow their child to take part in the survey and even with parental consent some students refused to participate or the student was absent from class on the day of collection Finally some observations were removed because they did not contain sufficient information and could not be considered usable They were however considered as valid responses in the calculation of the response rates The final response rates at the studen
18. and drug use items are only intended for respondents in grade 7 8 or 9 Therefore Y_q73 to Y 9g88d age was coded as 96 Valid Skip if grade was indicated as 5 Grade 5 or 6 Grade 6 Note that the module of the questionnaire distributed to grade 5 and 6 classes did not include these items Y_q74 How old were you when you first had a drink of alcohol that is more than a sip is only relevant if the respondent had ever had a drink of alcohol Therefore Y_q74 was coded as 96 Valid Skip if Y_73 Have you ever had a drink of alcohol that is more than just a sip was 2 No 96 Valid Skip or 99 Not Stated Y_q75 In the last year how often did you drink alcohol is only relevant if the respondent had ever had a drink of alcohol Therefore Y_q75 was coded as 96 Valid Skip if Y_73 Have you ever had a drink of alcohol that is more than just a sip was 2 No 96 Valid Skip or 99 Not Stated Y_q76 Have you ever had 5 drinks or more of alcohol on one occasion is only relevant if the respondent had ever had a drink of alcohol Therefore Y_q75 was coded as 96 Valid Skip if Y_73 Have you ever had a drink of alcohol that is more than just a sip was 2 No 96 Valid Skip or 99 Not Stated 2004 2005 Youth Smoking Survey User Guide 21 Y g77 How old were you when you first had 5 drinks or more of alcohol on one occasion is only relevant if the respondent had ever had 5 drinks or more on one occasion Therefore Y_q77 w
19. as coded as 96 Valid Skip if Y_q76 Have you ever had 5 drinks or more of alcohol on one occasion was 2 No 96 Valid Skip or 99 Not Stated Y_q78 In the last year how often did you have 5 drinks of alcohol or more on one occasion is only relevant if the respondent had ever had 5 drinks or more on one occasion Therefore Y_q77 was coded as 96 Valid Skip if Y_q76 Have you ever had 5 drinks or more of alcohol on one occasion was 2 No 96 Valid Skip or 99 Not Stated Y_q81 How old were you when you first used marijuana or cannabis is only relevant if the respondent had ever tried marijuana Therefore Y_q81 was coded as 96 Valid Skip if Y_q80 Have you ever used or tried marijuana or cannabis was 2 No 96 Valid Skip or 99 Not Stated Y_q82 In the last year how often did you use marijuana or cannabis is only relevant if the respondent had ever tried marijuana Therefore Y_q81 was coded as 96 Valid Skip if Y_q80 Have you ever used or tried marijuana or cannabis was 2 No 96 Valid Skip or 99 Not Stated DVAMTSMK The average number of cigarettes smoked per day in the past week should only be determined if the respondent had smoked a whole cigarette Therefore DVAMTSMK was coded as 96 Valid Skip if Y_q18 Have you ever smoked a whole cigarette was 2 No or 96 Valid Skip DVCIGWK Total number of cigarettes smoked in the past 7 days prior to the survey should only be determined if the respo
20. ation estimates for each province based on coefficients of variation for estimates of population totals For example the table shows that the quality of a weighted estimate of 1 206 people possessing a given characteristic in Newfoundland and Labrador should be flagged as marginal Table 10 Release Cut Off s by Province Province Acceptable Marginal Unacceptable CV 0 0 16 5 CV 16 6 33 3 CV gt 33 3 Newfoundland and Labrador 4 442 8 over 1 206 to lt 4 442 under 1 206 Prince Edward Island 538 amp over 136 to lt 538 under 136 Nova Scotia 1 333 amp over 329 to lt 1 333 under 329 New Brunswick 4 939 amp over 1 301 to lt 4 939 under 1 301 Quebec 43 586 amp over 11 342 to lt 43 586 under 11 342 Ontario 38 286 amp over 9 647 to lt 38 286 under 9 647 Manitoba 16 043 amp over 4 676 to lt 16 043 under 4 676 Saskatchewan 15 046 8 over 4409 to lt 15 046 under 4 409 Alberta 8 886 amp over 2 222 to lt 8 886 under 2 222 British Columbia 19 879 amp over 5 156 to lt 19 879 under 5 156 Canada 29 995 amp over 7 360 to lt 29995 under 7 360 2004 2005 Youth Smoking Survey User Guide updated April 2006 37
21. ative school boards for provinces in which these are administratively separate from public boards Stratum 3 also includes the three provinces NL NS PE in which all boards were selected Table 4 Number of boards selected by stratum Stratum Total 1 2 Newfoundland and Labrador 0 0 4 4 Nova Scotia 0 0 5 5 Prince Edward Island 0 0 2 2 New Brunswick 2 2 0 4 Qu bec 4 5 1 10 Ontario 5 6 1 12 Manitoba 3 3 2 8 Saskatchewan 2 3 1 6 Alberta 5 2 1 8 British Columbia 4 4 1 9 Canada 25 25 18 68 2004 2005 Youth Smoking Survey User Guide 24 Stage 2 Calculation of W2 Within each selected board schools were stratified into two strata Stratum 1 the Junior stratum contained schools with student in grades 5 6 5 6 and 6 7 Stratum 2 the Senior stratum contained schools with students in grades 5 8 5 9 6 8 6 9 7 7 8 7 9 8 and 9 Schools were sampled by stratified random sampling without replacement For the private schools simple random sampling was used to select the required number of boards The number of private schools selected in a province was proportional to the number of students enrolled in private schools in that province Because the differences in sampling fractions in the two secondary strata within a board were administrative in origin a raising factor from school to board was computed for each grade as follows Z N g 2 in g e 9 where n g is the number of students responding in grade g in
22. chine errors that were to be corrected e g where the respondent erased one mark and chose another answer but the scanner picked up the erased mark too There were a total of 712 cases of discrepancy checked and appropriate corrections made during the scanning process Logbooks and a quality control record were kept to track the number of corrections made and to monitor the progress of merging files to create a school level file 6 2 Editing and Imputation The following standard codes are used in the microdata file Valid skip 6 96 and 996 Don t know 7 97 and 997 Refused 8 98 and 998 Not stated 9 99 and 999 The YSS 2004 2005 student dataset had 30 761 records One questionnaire was scanned twice and the duplicate file was removed Students who were not in grade 5 6 7 8 or 9 were removed from the data set resulting in 29 570 eligible records Based on incomplete records particularly related to the determination of smoking status 327 individuals were removed The final number of records is 29 243 Qu bec grades secondary I II III were converted to grades 7 8 9 respectively The variable Y_q1 was changed to the converted grade values as D grade The variable Y_ql was left for users who want to do analyses separately for Qu bec The original dataset included an additional 1 190 students sampled in Ontario in grades 10 11 and 12 for a related project 2004 2005 Youth Smoking Survey User Guide 11 The followi
23. d in the original sample design iii Then within each re sampled school all eligible students who had consent to participate are selected iv The weights for re selected units are recalculated and adjusted for the re sampling inference based on the method of Rao and Wu 1988 v Finally the new weights are recalibrated to the provincial enrollment figures using the administrative datasets Six thousand 6000 bootstrap samples have been computed The average of sets of twelve bootstrap weights have been used to create a set of 500 averaged bootstrap weights The formula for the weight adjustment is obtained as follows Let w be the smoothed calibrated main weight for student k in schoolj of board i Nn Nn where N is the number of boards in the primary stratum for board i and n is the number of boards actually selected in that stratum n M mM 5 j 4 1 _ where M is the number of schools in the secondary stratum N M m l i Let 4 within board i for schoolj and m is the number of schools chosen in that secondary stratum within board 7 Rao J N K amp Wu C F J 1988 Resampling inference with complex survey data Journal of the American Statistical Association 83 231 241 2004 2005 Youth Smoking Survey User Guide 27 The bootstrap weight Wit would then be given by w l1 4 number of times board i has been resampled A 2 number of times sch
24. dents However among the W computed this way there were some extreme weights due to low consent rates from certain schools Using the rationale described by Potter 1988 the lower level of response has been set to 0 25 for a given grade and school and the weight components were recalculated The final un calibrated weight is based on Weight W W2 W3 Potter F 1988 Survey of procedures to control extreme sampling weights ASA Proceedings of the Section on Survey Research Methods 453 458 2004 2005 Youth Smoking Survey User Guide 26 Stage 4 Calibration of survey weights The weights just described are then calibrated using school administrative datasets that include the total student enrollment by gender and grade grades 5 through 9 for each province Province grade and gender calibration are used to adjust the sampling weights so that estimated numbers of students in these domains reproduce known population numbers ezactly Stage 5 Construction of Bootstrap Weights The bootstrap weights for each province have been constructed separately as follows i Within each primary stratum board stratum the same number of boards is selected by simple random sampling SRS with replacement as has been selected in the original sample design 11 Within each re sampled board and within each school stratum the same number of schools is selected by simple random sampling SRS with replacement as has been selecte
25. ed as 96 Valid Skip if Y_q11 Have you ever tried cigarette smoking even just a few puffs was 2 No Y_q35ba to Y_q35bd Do you sometimes buy single cigarettes If you do where do you buy them are only relevant if the respondent had purchased single cigarettes Therefore Y_q35ba to Y_q35bd were coded as 96 Valid Skip if Y_q35a was 1 I do not smoke 2 I do not buy single cigarettes or 99 Not Stated Y_q36 to Y 940 cessation items are only relevant if the respondent had tried smoking cigarettes Therefore Y_q36 to Y_q40 were coded as 96 Valid Skip if Y_q11 Have you ever tried cigarette smoking even just a few puffs was 2 No Y_q41 to Y 946 addiction scale are only relevant if the respondent had smoked in the last 30 days Therefore Y_q41 to Y_q46 were coded as 96 Valid Skip if Y_q24 was 1 None or 96 Valid Skip Y_q50 How does your father or the person who is like your father feel about your smoking is only relevant if the respondent had tried smoking cigarettes Therefore Y_q50 was coded as 96 Valid Skip if Y_q11 Have you ever tried cigarette smoking even just a few puffs was 2 No Y 952 How does your mother or the person who is like your mother feel about your smoking is only relevant if the respondent had tried smoking cigarettes Therefore Y_q52 was coded as 96 Valid Skip if Y_q11 Have you ever tried cigarette smoking even just a few puffs was 2 No Y_q73 to Y_q88d_age alcohol
26. er of students enrolled This information was used to prepare consent materials and was entered along with other school particulars e g address data collection date etc into a database Upon receipt of consent materials student information was entered into this database and questionnaire IDs were assigned Questionnaires were bundled by classroom and couriered to the school contact for distribution to classroom teachers 1 to 2 days prior to the data collection date On the data collection date teachers administered the survey according to detailed instructions during a designated class period The survey took on average 30 to 40 minutes to complete To protect confidentiality teachers were asked not to circulate among the students Completed questionnaires were placed in a classroom envelope A project staff member site coordinator or data collector was required to attend each school data collection The staff member set up a station in front of the school office or another central location The data collector was available to answer questions and receive classroom envelopes at the end of the data collection period Within a few days of data collection the site coordinators shipped the completed questionnaires organized by school and classroom to the coordinating centre the Population Health Research Group at the University of Waterloo for processing Parent interviews were conducted through a sub contract with Crawford Canada Trained
27. ever smoked 100 or more whole cigarettes in your life 1 Yes And Y 924 On how many of the last 30 days did you smoke one or more cigarettes 6 30 days every day 2 Current Occasional Smoker A current occasional smoker is a person who currently smokes cigarettes but not every day Y 920 Have you ever smoked 100 or more whole cigarettes in your life 1 Yes And Y 924 On how many of the last 30 days did you smoke one or more cigarettes 2 1 to 5 days or 3 6 to 10 days or 4 11 to 20 days or 5 21 to 29 days 3 Former Daily Smoker A former daily smoker is a person who smoked at least 100 cigarettes in his her life time and smoked at least seven days in a row but did not smoke in the last 30 days Y_q20 Have you ever smoked 100 or more whole cigarettes in your life 1 Yes And Y 924 On how many of the last 30 days did you smoke one or more cigarettes 1 None And Y_q21 Have you ever smoked everyday for at least 7 days in row Yes 4 Former Occasional Smoker A former occasional smoker is a person who smoked at least 100 cigarettes in his her life time and did not smoke for at least seven days in a row and also did not smoke in the last 30 days Y 920 Have you ever smoked 100 or more whole cigarettes in your life 1 Yes And Y 924 On how many of the last 30 days did you smoke one or more cigarettes 1 None And Y_q21 Have you ever smoked everyday for at least
28. examples of non sampling errors Over a large number of observations randomly occurring errors will have little effect on estimates derived from the survey However errors occurring systematically will contribute to biases in the survey estimates Considerable time and effort were taken to reduce non sampling errors in the survey Quality assurance measures were implemented at each step of the data collection and processing cycle to monitor the quality of the data These measures included a the use of protocols that have been validated in previous studies of school based data collection around youth smoking 2 detailed instructions for teachers 3 extensive training of project staff with respect to the survey procedures 4 procedures to ensure that data capture errors were minimized and 5 coding and edit quality checks to verify the processing logic 2004 2005 Youth Smoking Survey User Guide 31 8 0 Guidelines for Tabulation Analysis and Release Please note that this chapter is adapted from the 2002 Youth Smoking Survey User Guide written by Statistics Canada It details the guidelines to be adhered to by users tabulating analyzing publishing or otherwise releasing any data derived from the survey microdata files With the aid of these guidelines users of microdata should be able to produce the same figures as those produced by any statistician and at the same time will be able to develop currently unpublished figures in a manner
29. imates Estimates of the number of people with a certain characteristic can be obtained from the microdata file by summing the final weights of all records possessing the characteristic s of interest Proportions and ratios of the form X Y are obtained by 2004 2005 Youth Smoking Survey User Guide 33 a summing the final weights of records having the characteristic of interest for the numerator X b summing the final weights of records having the characteristic of interest for the denominator Y then c dividing estimate a by estimate b X Y 8 3 4 Tabulation of Quantitative Estimates Estimates of quantities can be obtained from the microdata file by multiplying the value of the variable of interest by the final weight for each record then summing this quantity over all records of interest For example to obtain an estimate of the total number of cigarettes smoked in the past seven days prior to the survey by students in grade 9 secondary III in Qu bec multiply the value reported in the derived variable DVCIGWK number of cigarettes smoked in the past seven days prior to the survey by the final weight for the record then sum this value over all records with DVCIGWK lt 996 For example to estimate the average number of cigarettes smoked in the past seven days prior to the survey by students in grade 9 a estimate the total number of cigarettes smoked in the past seven days prior to the survey by students
30. in grade 9 X as described above b estimate the number of students in grade 9 secondary III in Qu bec Y in this category by summing the final weights of all records with DVCIGWK lt 996 then c divide estimate a by estimate b X Y 84 Guidelines for Statistical Analysis The 2004 2005 YSS is based upon a complex sample design with stratification multiple stages of selection and unequal probabilities of selection of respondents Using data from such complex surveys challenges analysts because the survey design and the selection probabilities affect the estimation and variance calculation procedures that should be used In order for survey estimates and analyses to be free from bias the survey weights must be used While many analysis procedures found in statistical packages allow weights to be used the meaning or definition of the weight in these procedures differ from that which is appropriate in a sample survey framework with the result that while in many cases the estimates produced by the packages are correct the variances that are calculated are less precise For example suppose that analysis of all male respondents is required The steps to rescale the weights are as follows 1 select all respondents from the file who reported GENDER male 2004 2005 Youth Smoking Survey User Guide 34 2 calculate the AVERAGE weight for these records by summing the original student weights from the microdata file for these
31. in the range of 0 0 to 16 5 No warning is required 2 Marginal Estimates have a sample size of 30 or more and high coefficients of variation in the range of 16 6 to 33 3 Estimates should be flagged with the letter M or some similar identifier They should be accompanied by a warning to caution subsequent users about the high levels of error associated with the estimates 3 Unacceptable Estimates have a sample size of less than 30 or very high coefficients of variation in excess of 33 3 It is not recommended to release estimates of unacceptable quality Such estimates should be replaced with the letter U or some similar identifier and the following statement Unreleaseable due to low sample size 2004 2005 Youth Smoking Survey User Guide 36 8 6 Release Cut off s for the 2004 2005 Youth Smoking Survey Data In the YSS 2002 Micro User s Guide approximate sampling variability tables were supplied in order for the user to determine approximate Coefficients of Variation C V s for estimates of different types For the 2004 2005 Youth Smoking Survey data users are encouraged to use bootstrap weights as described earlier to calculate the variance of all estimates Coefficients of variation are derived using the variance formula for simple random sampling and incorporating a factor that reflects the underlying characteristics of the sampling design The table below provides rough release cut off s for popul
32. ind yesterday on the wheel and fill in the number of cigarettes that you smoked Then follow the wheel backwards and fill in the number of cigarettes you smoked on each of the last 7 days a Sunday Yo 0 0 Cigarettes smoked b Monday Y_q23mon 1 36 Cigarettes smoked c Tuesday Y_q23tue ees 96 Valid skip d Wednesday Y_q23wed 99 Not Stated e Thursday Y_q23thu f Friday Y q23fri g Saturday Y_q23sat Coverage Respondents where Y q18 1 Whole Cigarette DVAMTSMK The average number of cigarettes smoked per day in the past week as an integer value Y_g23sun Y g23mon Y q23tues Y g23wed Y q23thurs Y g23fri Y g23sat 7 All responses had to have valid responses for valid data If all responses have 99 or if any of the days are missing then DVAMTSMK 99 2004 2005 Youth Smoking Survey User Guide 18 DVCIGWK Total number of cigarettes smoked in the 7 days prior to the survey Y_q23sun Y_q23mon Y_q23tues Y_q23wed Y_q23thurs Y_q23fri Y_q23sat Not necessary for all to have valid responses Zero value has been treated as a valid response If all days have missing data then DVCIGWK 99 DVNDSMK Number of days on which respondent smoked in the week prior to the survey A count of Y_q23sun Y_q23mon Y_q23tues Y_q23wed Y_q23thurs Y_q23fri and Y_q23sat excludes the missing responses Zero has been treated as a zero response DVAVCIGD Average number of cigarettes smoked on the days tha
33. moker 2 Former Smoker 3 Never Smoker 1 Current Smoker A current smoker is a person who reports having smoked 100 cigarettes and has smoked in the past 30 days Y_q20 Have you ever smoked 100 or more whole cigarettes in your life 1 Yes And Y 924 On how many of the last 30 days did you smoke one or more cigarettes 2 1 to 5 days or 3 6 to 10 days or 4 11 to 20 days or 5 21 to 29 days or 6 30 days every day 2 Former Smoker A former smoker is a person who reports having smoked 100 or more cigarettes but did not smoke in the last 30 days Y 920 Have you ever smoked 100 or more whole cigarettes in your life 1 Yes And Y 924 On how many of the last 30 days did you smoke one or more cigarettes 1 None 3 Never Smoker A never smoker is a person who reports that they have not smoked 100 or more whole cigarettes in their life time but they might have smoked a whole cigarette Y_q20 Have you ever smoked 100 or more whole cigarettes in your life 2 No Or Have never smoked a whole cigarette DVTY2ST 1 Current Daily Smoker 2 Current Occasional Smoker 3 Former Daily Smoker 4 Former Occasional Smoker 5 Experimental Smoker Beginner 6 Past Experimental Smoker 7 Puffer 8 Never Tried 2004 2005 Youth Smoking Survey User Guide 13 1 Current Daily Smoker A current daily smoker is a person who reports currently smoking cigarettes every day Y 920 Have you
34. n of Private and Independent Schools Within each province except British Columbia lists of private schools were obtained A simple random sample of private schools was selected in each province from these lists The number of schools originally selected was roughly proportional to the number of students enrolled in private schools in that province Substitute private schools were selected as replacement schools in case any selected private school refused to participate In British Columbia private schools are integrated into public school districts and so were eligible for selection as described above Replacement of Boards and Schools If a selected board refused to participate it was replaced from the substitute list with the board from the same stratum whose adult smoking rate was closest to that of the original school board In British Columbia all original and substitute school boards were approached The high refusal rate meant a third round of board selections was required If a school refused it was replaced with the next substitute school in the same stratum Similarly if a private school refused it was replaced by the next substitute private school Selection of Students Within each selected school all students in the eligible grades were eligible for the final sample Consent for student participation was obtained from the parents of students who participated in the survey Students without consent were not allowed to complete a
35. nded to all questions in order that all students regardless of behaviour would finish at about the same time A separate version of the questionnaire for grades 5 6 classes did not contain alcohol and drug items The 2004 2005 parent interview was also based on the 2002 survey A number of changes were made in consultation with Health Canada representatives including the addition of questions on parent marijuana use and on smoking in vehicles 5 2 Data Collection Protocols Prior to implementation all protocols and materials were approved by the University of Waterloo Human Research Ethics Committee Local institutional review boards affiliated with the institutions of consortium members also reviewed the project at the provincial level where applicable i e where collaborators were affiliated with universities 2004 2005 Youth Smoking Survey User Guide 8 Each provincial collaborator hired a site coordinator to be responsible for school board and school recruitment data collection preparation and implementation Site coordinators attended a two day training session participated in additional web based training sessions received a comprehensive manual and had ready access to the Student Data Collection Coordinator for advice regarding day to day issues Materials databases and protocols were centrally developed to ensure consistency across provinces Provincial site coordinators were responsible for all board and school recruitment
36. ndent had smoked a whole cigarette Therefore DVCIGWK was coded as 996 Valid Skip if Y_q18 Have you ever smoked a whole cigarette was 2 No or 96 Valid Skip DVNDSMK Number of days on which respondent smoked in the week prior to the survey should only be determined if the respondent had smoked a whole cigarette Therefore DVNDSMK was coded as 96 Valid Skip if Y_q18 Have you ever smoked a whole cigarette was 2 No or 96 Valid Skip DVAVCIGD Average number of cigarettes smoked on the days that the respondent smoked should only be determined if the respondent had smoked a whole cigarette Therefore DVAVCIGD was coded as 96 Valid Skip if Y_q18 Have you ever smoked a whole cigarette was 2 No or 96 Valid Skip DVSMKPTN Smoking pattern in the last 7 days should only be determined if the respondent had smoked a whole cigarette Therefore DVSMKPTN was coded as 96 Valid Skip if Y_q18 Have you ever smoked a whole cigarette was 2 No or 96 Valid Skip 2004 2005 Youth Smoking Survey User Guide 22 6 6 Weighting As described in section 4 2 the sampling of schools for the 2004 2005 YSS was conducted in two stages At stage one school boards were sampled within each province From the selected school boards schools were then sampled All students in the grades 5 through 9 in the selected schools were eligible for the final sample The development of the survey weights was accomplished in stages In the first
37. ng Survey User Guide 15 P SMOKE 1 At least 1 parent smokes 0 Neither parent smokes 1 At least 1 parent smokes If Y_q49 Does your father or the person who is like your father smoke cigarettes 4 He smokes now OR Y_q51 Does your mother or the person who is like your mother smoke cigarettes 4 She smokes now If both Y_q49 and Y 951 are not stated then P_smoke 99 Not Stated If both Y_q49 and Y 951 are do not know then P_smoke 97 Do not know All other combinations of responses have a value of 0 Zero SIB_SMOKE 1 At least 1 sibling smokes 0 No siblings smoke 1 At least 1 sibling smokes If Y_q53 Do any of your sisters smoke cigarettes 4 At least 1 of my sisters smokes now Or If Y_q54 Do any of your brothers smoke cigarettes 4 At least 1 of my brothers smokes now If both Y_q53 and Y 954 are not stated then Sib smoke 99 Not Stated If both Y_q53 and Y 954 are do not know then P_smoke 97 Do not know All other combinations of responses have a value of 0 Zero 2004 2005 Youth Smoking Survey User Guide 16 D SUSCEPTIBLE 1 No 2 Yes 99 Not Stated The susceptibility scale is based on the following three questions Y_ql4 Do you think in the 1 I have already tried smoking future you might try 2 Definitely yes smoking cigarettes 3 Probably yes 4 Probably not 5 Definitely not 99 Not Stated Y_ql5 fone of your best 1
38. ng items reguired specific editing and or imputation Question 1 Y_q1 During the process of determining consent form return rates and consent rates it was determined that several students had filled out a grade on the survey that was inconsistent with the grades represented in the school Ifa student filled out a grade that did not match the relevant grades in the school the variable was recoded to match the nearest relevant grade If the value of Y_ql was changed by this process the indicator variable IMP_q1 was set to 1 If Y_ql grade was missing uncodeable or improper for the province then the student s grade was found from another source The primary source was the student consent form This form was filled out by the parent or the student and then signed by the parent Grade was one of the fields The secondary source was the grade that was associated with the student s class identification number If multiple grades were listed in these fields then the first grade listed was used If the value of Y_ql was changed based on these sources then the variable imp g was set to 1 if no change was made IMP _q1 was set to 0 Question 3 Y_q3 If Y_q3 gender was either missing or uncodeable then gender was found from another source The primary source was the student consent form This form was filled out by the parent or the student and then signed by the parent Gender was one of the fields The secondary source was the st
39. ntario KIA 0K9 Telephone 519 885 1211 ext 6632 Telephone 613 957 7852 Fax 519 746 8171 Fax 613 954 2292 E mail r4ahmed uwaterloo ca E mail Alan_Diener hc sc gc ca This manual has been produced to facilitate the manipulation of the micro data file of the survey results Please become familiar with the contents of this document before publishing or otherwise releasing any estimates derived from the microdata file of the 2004 2005 Youth Smoking Survey 2004 2005 Youth Smoking Survey User Guide 1 2 0 Background The Youth Smoking Survey YSS is a Health Canada sponsored survey of students in grades 5 through 9 and their parents The YSS was first administered in 1994 and was the largest and most comprehensive survey on youth smoking behaviour since 1979 The YSS was repeated in 2002 in order to track changes in the attitudes and behaviour of Canadian children and adolescents with respect to tobacco At that time the YSS was planned as a biennial survey The YSS is a unique and important survey because reliable data on smoking prevalence rates among young Canadians are somewhat sparse this is especially true for adolescents in grades 5 through 9 who are most vulnerable to start smoking Some information on smoking prevalence is obtained routinely through school surveys however national level data are obtained only rarely Monitoring surveys which address a variety of psychoactive substances provide only limited information on
40. of boards in the stratum Because of a high number of refusals in BC a second round of first stage sampling was necessary This was accomplished by sampling additional boards to meet our minimum requirements For these additional boards the probability of inclusion was calculated as o T where 2004 2005 Youth Smoking Survey User Guide 23 LM 2 2 SIO p sy sj gt J JES and where l Number of boards selected at the first round of first stage sampling l Number of boards selected at the second round of first stage sampling pi s Probability of the sample s of school boards at the first round m Probability of inclusion at the first round of first stage sampling and 7 Probability of inclusion at the second round of first stage sampling All private schools in a province were assigned to a distinct board All the native school boards in Manitoba were assigned to one native school board Because of very large enrollment it was assumed that the Toronto District School Board would have been selected with certainty Therefore their first stage weight W1 1 The table below summarizes the number of school boards per stratum As detailed in section 4 2 stratum 1 consists of school boards in public health regions with above average adult smoking rates as measured by the CCHS Stratum 2 consists of school boards in public health regions with below average adult smoking rates Stratum 3 includes private French language and or n
41. ool j has been resampled A 6 7 Suppression of Confidential Information It should be noted that the Public Use Microdata Files PUMF may differ from the survey master files held by the Population Health Research Group These differences usually are the result of actions taken to protect the anonymity of individual survey respondents The most common actions are the suppression of file variables grouping values into wider categories and coding specific values into the not stated category Specifically the following variables have been removed from the PUMF respondent s age school board identifier school identifier class identifier stratum identifier postal code and responses from the parent interviews 2004 2005 Youth Smoking Survey User Guide 28 7 0 Data Quality There are various factors that influence data quality This chapter summarizes threats to data quality and steps taken to ameliorate these 7 1 Response Rates There were various levels of non response throughout the 2004 2005 YSS A description of these levels is presented below along with the appropriate tables First some degree of non response was noted among school boards and schools Replacements were found for the majority of school boards and schools who refused to participate in the survey The final response rates at the school board and school level are presented in the tables below Table 6 Board Recruitment Outcomes Newfoundland and L
42. pating schools received a customized feedback report The feedback report focuses on smoking and related behaviours and is intended to be relevant and prescriptive for schools and communities 2004 2005 Youth Smoking Survey User Guide 2 3 0 Concepts and Definitions See section 6 4 Creation of Derived Variables for more detailed information about the derived variables reflecting these definitions Currently smokes Has smoked at least 100 cigarettes in his her lifetime and has smoked in the 30 days preceding the survey Currently smokes daily Has smoked at least 100 cigarettes in his her lifetime and has smoked at least one cigarette per day for each of the 30 days preceding the survey Currently smokes occasionally Has smoked at least 100 cigarettes in his her lifetime and has smoked at least one cigarette during the 30 days preceding the survey but has not smoked every day Formerly smoked Has smoked 100 or more cigarettes in his her lifetime but has not smoked at all during the 30 days preceding the survey Formerly smoked daily Has smoked 100 or more cigarettes in his her lifetime but has not smoked at all during the 30 days preceding the survey and has at some time smoked every day for seven days in a row Formerly smoked occasionally Has smoked 100 or more cigarettes in his her lifetime but has not smoked at all during the 30 days preceding the survey and has never smoked every day for seven days in a row
43. pulation possessing certain characteristics or falling into some defined category The number of students who ever smoked a whole cigarette or the proportion of smokers who usually buy single cigarettes from a friend or someone else are examples of such estimates An estimate of the number of persons possessing a certain characteristic may also be referred to as an estimate of an aggregate Examples of Categorical Questions Q Have you ever smoked a whole cigarette R Yes No Q Where do you usually get your cigarettes R I buy them from a vending machine I buy them myself at a store 1 buy them from someone etc 8 3 2 Quantitative Estimates Quantitative estimates are estimates of totals or of means medians and other measures of central tendency of quantities based upon some or all of the members of the surveyed population They also specifically involve estimates of the form X Y where X isan estimate of surveyed population quantity total and Y is an estimate of the number of persons in the surveyed population contributing to that total quantity The only example of a quantitative estimate in the 2004 2005 YSS is the number of cigarettes smoked on each of the last seven days If users want to estimate the average number of cigarette in a week then the numerator is the total number of cigarettes smoked in the last seven days and the denominator would be the number of days smoked in last seven days 8 3 3 Tabulation of Categorical Est
44. records and then dividing by the number of respondents who reported GENDER male 3 for each of these respondents calculate a RESCALED weight equal to the original student weight divided by the AVERAGE weight 4 perform the analysis for these students using the RESCALED weight While this method produces reliable estimates of the coefficients under consideration in the analysis note that because the stratification and clustering of the sample s design are still not taken into account the variance estimates calculated in this way are likely to be under estimates The calculation of more precise variance estimates requires detailed knowledge of the design of the survey Such detail cannot be given in this microdata file to respect confidentiality However variances that take account for the sample design can be calculated from the bootstrap weights which are provided as a separate data file Health Canada employed Stata for all analyses of the 2004 05 YSS Variance estimates were conducted by using the BBWREG command This procedure creates reliable estimates of the variance for both simple estimates such as totals proportions and ratios and more complex analyses such as linear or logistic regression Another option is to use the Bootvar program available in both SAS and SPSS formats It is made up of macros that compute variances for totals differences between ratios and for linear and logistic regression The bootstrap program can be reque
45. rmation about each eligible board including board enrollment health region and number of schools 4 2 Sample Design The sampling of schools for the 2004 2005 YSS was conducted in two stages At stage 1 school boards were sampled within each province From the selected school boards schools were then sampled All students in grades 5 through 9 primary 5 6 and secondary I to II in Qu bec in the selected schools were eligible for the final sample Stage 1 Selection of School Boards In Newfoundland and Labrador Nova Scotia and Prince Edward Island the small number of boards meant that all boards were selected for the sample For the remaining provinces the procedure described below was followed The Canadian Community Health Survey CCHS sampled sufficient adults in each health region to allow for an estimate of the current smoking rate at the level of the health region Within each province each school board was assigned an estimate of the adult smoking rate from the Canadian Community Health Survey CCHS for the public health region that contained the school board In the case that a school board was located in more than one health region an average smoking rate was calculated based on the smoking rates for each of the relevant health regions Within provinces some school boards that had few schools were combined with school boards with a similar adult smoking rate and treated as a single board for sampling The school boards
46. s and differences of aggregates or ratios are to be derived from their corresponding un rounded components and then are to be rounded themselves to the nearest 100 units or the nearest one decimal using normal rounding e Under no circumstances are un rounded estimates to be published or otherwise released by users Un rounded estimates imply greater precision than actually exists 8 2 Sample Weighting Guidelines for Tabulation The sample design used for the Youth Smoking Survey YSS was not self weighting When producing simple estimates including the production of ordinary statistical tables users must apply the proper sampling weight If proper weights are not used the estimates derived from the Stats Canada 2002 Microdata User Guide Youth Smoking Survey 2002 Accessible at http www statcan ca english sdds document 4401_D2_T9_V2_E pdf 2004 2005 Youth Smoking Survey User Guide 32 microdata files cannot be considered to be representative of the survey population and will not correspond to those produced by Health Canada 8 3 Definitions of Types of Estimates Categorical and Quantitative Before discussing how the YSS data can be tabulated and analyzed it is useful to describe the two main types of point estimates of population characteristics which can be generated from the microdata file for the YSS 8 3 1 Categorical Estimates Categorical estimates are estimates of the number or percentage of the surveyed po
47. school j c g is the consent rate for grade g in school j N g is the enrollment in grade g in the board obtained from administrative data The table below summarizes the number of schools per stratum Table 5 Number of schools selected by stratum Stratum Total 1 2 Junior Senior Newfoundland and Labrador 12 12 24 Nova Scotia 11 13 24 Prince Edward Island 14 10 24 New Brunswick 8 12 20 Qu bec 24 12 36 Ontario 11 30 41 Manitoba 11 17 28 Saskatchewan 3 19 22 Alberta 10 20 30 British Columbia 16 16 32 Canada 120 161 281 2004 2005 Youth Smoking Survey User Guide 25 Stage 3 Calculation of W The adjustment for non response was corrected by school at the grade not class level The enrollment information collected by Site Coordinators was at the class level However some of the classes were split grades e g grade 6 and grade 7 students were in the same class The following method was used to determine the number of students enrolled in grades 5 through 9 the number of consent forms returned and the number of students who had parental consent to participate from the participating schools The enrollment number of forms returned and positive consents were tabulated by using the grade level assigned to each class A second table was generated to list all of the split class students to allow adjustments to be made if the actual grade of the student was different from the grade assigned to the student s class
48. sted free of cost from Stats Canada with the documents explaining how to modify and use the program to meet users needs 8 5 Coefficient of Variation Release Guidelines Before releasing and or publishing any estimate from the 2004 2005 YSS users should first determine the quality level of the estimate The quality levels are acceptable marginal and unacceptable Data quality is affected by both sampling and non sampling errors as discussed in Chapter 7 However for this purpose the quality level of an estimate will be determined only on the basis of sampling error as reflected by the coefficient of variation as shown in the table below Nonetheless users should be sure to read Chapter 7 to be more fully aware of the quality characteristics of these data BSWREG is a Stata ado file For more information on the BSWREG command including the necessary ado files to run the command please refer to the following two papers Health Canada used the updated version of the BSWREG ado file that takes into account the fact that the bootstrap weights provided are actually mean weights In the case of the 2004 05 YSS data each of the 50 mean bootstrap weights is the average of twelve bootstrap weights Hence when using the BSWREG command one must set cmeansb 12 Note that users must merge the YSS PUMF with the bootstrap weights data file before proceeding This as well as the other necessary details required to conduct this type of analysis are provided in
49. t level are summarized in the table below 2004 2005 Youth Smoking Survey User Guide 30 Table 8 Student Level Response Rates Eligible Students with Completed Usable Response students consent guestionnaires guestionnaires rate a 4247 2773 2548 2518 60 Nova Scotia 5240 3067 2824 2782 54 Prince Edward Island 3637 2718 2509 2485 69 New Brunswick 4028 2937 2619 2581 65 Qu bec 7233 4022 3682 3644 51 Ontario 7386 4621 4188 4150 57 Manitoba 4739 3238 3034 3010 64 Saskatchewan 3280 2211 2070 2046 63 Alberta 4429 2874 2649 2625 60 British Columbia 7066 3701 3430 3402 49 Canada 51285 32162 29553 29243 58 based on completed questionnaires 7 2 Survey Errors The estimates derived from this survey are based on a sample of schools Somewhat different estimates might have been obtained if a complete census had been taken using the same questionnaire data collection staff processing methods and so on as those actually used in the survey The difference between the estimates obtained from the sample and those resulting from a complete count taken under similar conditions is called the sampling error of the estimate Errors which are not related to sampling may occur at almost every phase of a survey Teachers may misunderstand instructions respondents may make errors in answering questions the answers may be incorrectly entered on the questionnaire and errors may be introduced in the processing and tabulation of the data These are all
50. t had tried smoking cigarettes Therefore Y_q18 was coded as 96 Valid Skip if Y_q11 Have you ever tried cigarette smoking even just a few puffs was 2 No Y g19 How old were you when you smoked your first whole cigarette is only relevant if the respondent had smoked a whole cigarette Therefore Y g19 was coded as 96 Valid Skip if Y g18 Have you ever smoked a whole cigarette was 2 No or 96 Valid Skip Y_q20 Have you ever smoked 100 or more whole cigarettes in your life is only relevant if the respondent had smoked a whole cigarette Therefore Y_q20 was coded as 96 Valid Skip if Y g18 Have you ever smoked a whole cigarette was 2 No or 96 Valid Skip Y_q21 Have you ever smoked every day for at least 7 days in a row is only relevant if the respondent had tried smoking cigarettes Therefore Y_q21 was coded as 96 Valid Skip if Y_qll Have you ever tried cigarette smoking even just a few puffs was 2 No Y 922 How old were you when you first smoked every day for at least 7 days in a row is only relevant if the responded had smoked every day for at least 7 days Therefore Y_q22 was coded as 96 Valid Skip if Y_q21 Have you ever smoked every day for at least 7 days in a row was 2 No or 96 Valid Skip Y_q23sun to Y_q23sat Think back over the last 7 days Fill in the number of cigarettes you smoked on each of the last 7 days are relevant only if the respondent had smoked a whole cigarette Therefore
51. t the respondent smoked DVCIGWK DVNDSMK DVSMKPTN 1 Smoked every day 2 Smoked week days only 3 Smoked weekend days only 4 Did not smoke in the last 7 days 5 Other pattern 99 Not stated Calculated based on these variables Y_q23sun Y_q23mon Y_q23tues Y_q23wed Y_q23thurs Y g23fri Y_q23sat 6 5 Skip Patterns The youth questionnaire was intentionally designed with no respondent use skip patterns to avoid the identification of smokers by rate of completion during the classroom session Thus all smoking behaviour items included a response option such as J do not smoke However due to the logical flow of the questions a number of questions are extraneous based on the answer to a previous question In these cases a skip pattern has been imposed onto the data set If a question could have been skipped if this were allowable within the structure of the questionnaire it was coded as 96 or 996 The following explains each question that has a 96 or a 996 code and the logical reasoning for coding the question in that way 2004 2005 Youth Smoking Survey User Guide 19 Y g12 How old were you when you first tried smoking cigarettes even just a few puffs is only relevant if the respondent had tried smoking cigarettes Therefore Y g12 was coded as 96 Valid Skip if Y_q11 Have you ever tried cigarette smoking even just a few puffs was 2 No Y_q18 Have you ever smoked a whole cigarette is only relevant if the responden
52. t to schools were added in order to enhance school level feedback reports The 2004 2005 student questionnaire was adapted from the 2002 YSS A Content Committee which included Health Canada representatives and all provincial collaborators was formed to review and update the questionnaire The provincial collaborators provided input on areas of interest for their province The questionnaire was finalized through a series of reviews and meetings In October 2004 a pilot test of the student questionnaire and the protocols for data collection was conducted As part of the pilot students were asked to complete the questionnaire according to normal protocols as well as provide feedback on any questionnaire items they found difficult to answer or did not understand Two focus groups were held with small groups of students in order to explore reactions to the survey in more depth Feedback was also collected from the school contact and teachers As a result of the pilot test refinements were made to six questions e g rewording additional response options and to explanatory text e g additional definitions repetition of assurances of confidentiality The student questionnaire was formatted to be machine readable so that data could be scanned directly into a computer For some items this necessitated the creation of response options to minimize the number of open ended responses The questionnaire contained no skip patterns all students respo
53. th Canada The 2004 2005 Youth Smoking Survey YSS was implemented under the leadership of Dr Steve Manske Principal Investigator with the Centre for Behavioural Research and Program Evaluation CBRPE at the University of Waterloo and Drs Steve Brown and Mary Thompson Co Principal Investigators from the Statistics and Actuarial Sciences Department at the University of Waterloo The 2004 2005 YSS was coordinated by staff from the Population Health Research Group PHR The investigators and staff were assisted by a consortium of university and non governmental organizations across the country Dr Shirley Solberg Memorial University Ms Meg McCallum Canadian Cancer Society Nova Scotia Division Ms Donna Murnaghan University of Prince Edward Island Dr William Morrison University of New Brunswick Dr Jennifer O Loughlin Institut national de sant publique du Qu bec Dr Dexter Harvey Canadian Cancer Society Manitoba Division Ms June Blau Saskatchewan Coalition for Tobacco Reduction Dr Cameron Wild University of Alberta Dr Chris Lovato University of British Columbia Any questions about the data set or its use should be directed to Population Health Research Group Health Canada Rashid Ahmed Dr Alan Diener Senior Data Analyst Office of Research Room LHN 2704 Surveillance and Evaluation 200 University Ave W Tobacco Control Programme Waterloo Ontario N2L 3G1 123 Slater Street University of Waterloo Ottawa O
54. udent s name If the name was unclear for gender then gender was left missing If the value of Y_q3 was changed based on these sources then the variable IMP_q3 was set to 1 if no change was made IMP 93 was set to 0 Question 23 Y_q23 This question asked how many cigarettes had been smoked on each of the last 7 days The range that was allowed for each day was 0 36 All responses between 37 and 99 have been set to 99 not stated Valid skips and not stated responses were set to 96 and 99 as in other variables 6 3 Coding of Open ended Options There were eight partially open ended items on the student questionnaire that included an Other specify option These write in answers were examined and recoded or retained as Other The recoding was done into existing categories or created answer categories For example 8 students wrote the kind my friend has or a very similar variation in the Other space for Q 27 even though My friends smoke the same brand was a response option In this instance 8 responses were recoded into the applicable response option 6 4 Creation of Derived Variables A number of variables in the microdata file have been derived by combining items on the questionnaire in order to facilitate data analysis Examples of derived variables include the average number of cigarettes smoked daily and the number of years the respondent smoked 2004 2005 Youth Smoking Survey User Guide 12 DVTYIST 1 Current S
Download Pdf Manuals
Related Search