COR 822 Graduate Telephone Internet Module Decision Rules Created: 01/19/2005 Updated: 09/09/2005 Contents: I. Overview II. Variable Descriptions A. Brief Variable Descriptions B. Detailed Variable Creation Descriptions III. Problems A. Individual Cases B. Survey/Replicate Updates IV. Coding and Data Editing V. People VI. References -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- I. Overview -------------------------------------------------------------------------------- The Internet module of the telephone interview includes items regarding the use of personal computers and the Internet by the respondent and members of the respondent’s household. Information about current or past household computer ownership and current or past household Internet access is included in this module. Main topics of interest for respondents who have access to the Internet from home include how long ago his or her household first got Internet access, how much time he or she spends using the Internet from home, and whether or not the respondent's spouse connects to the Internet from home. To gain a better understanding of why some respondents do not have a personal computer or access to the Internet from home, some items within the module ask the main reason why the respondent's household does not have these things. Previous waves of the Wisconsin Longitudinal Study did not include an Internet module. Although computer networks similar to the Internet existed much earlier, the public version of the Internet became available in the 1990's. At the time of the 1992 WLS study, very few households had Internet access, and an Internet module would have been impractical at that time. As the Internet continues to become a more popular tool for communication, entertainment, financial management, and learning, understanding our respondents' computer and Internet use trends can provide important insights about computer and Internet use among the aging population. In addition to obtaining information about computer and Internet use, the Internet module will also be used as an experiment to test different wordings of the same questions. Five questions in this module have alternate wordings and response choice orders. -------------------------------------------------------------------------------- II. Variable Descriptions -------------------------------------------------------------------------------- A. Brief Variable Descriptions ---------------------------------------------------------------------------- gm001he Is there currently a computer in the respondent’s household? gm002he Was there a computer in the respondent’s household in the past five years? gm003he-gm004he Main reason that the respondent’s household does not have a computer. gm005he Did anyone in the respondent’s household use the computer to connect to the Internet? gm006re Did the respondent use the computer to connect to the Internet? gm007re Did the respondent use the computer to do anything besides use the Internet? gm008he-gm009he Main reason that the respondent’s household no longer has a computer. gm010he Does anyone in the respondent’s household use the Internet from home? gm011he Does the respondent use the computer for anything besides the Internet? gm012he Did the respondent’s household ever have access to the Internet from home? gm013he-gm014he Main reason that the respondent’s household does not have Internet access. gm015he-gm016he Main reason that the respondent’s household no longer has Internet access. gm017he Century month when respondent’s household first obtained Internet access. gm018he-gm019he Type of connection that the respondent’s household uses to connect to the Internet. gm020re Does the respondent ever use the Internet from home? gm021re Does the respondent ever use the computer from home? gm022re Estimated time per week that the respondent spends connected to the Internet. gm023sp Does the respondent’s spouse use the Internet from home? gm024sp Does the respondent’s spouse ever use the computer from home? SUMMARY VARIABLES gms01re Fluflag - sampled at 50%. gms02re Dichotomous variable - the respondent has ever had a computer. gms03re Dichotomous variable - the respondent currently has a computer. gms04he Dichotomous variable - the respondent has ever had access to the Internet. gms05he Dichotomous variable - the respondent currently has access to the Internet. gms06he Dichotomous variable - Form of the question. gms07he Dichotomous variable - Form of responses. gms08re Dichotomous variable - Replicate that the respondent is in. gms09he Dichotomous variable - Respondent has a high-speed Internet connection. ---------------------------------------------------------------------------- B. Detailed Variable Creation Descriptions ---------------------------------------------------------------------------- The following notes offer a more detailed description of each variable, including the raw variables used to create each variable and notes about sampling, skip patterns, and variable creation. SAMPLING Code was included in the CASES instrument to randomly assign each respondent to one of two conditions for the following raw variables: fluflag, z3rand, z7rand, z11rand, z12rand, and z13rand. fluflag (gms01re) 50% of respondents = 0 50% of respondents = 1 z3rand 50% of respondents = 1 50% of respondents = 2 z7rand 50% of respondents = 1 50% of respondents = 2 z11rand 50% of respondents = 1 50% of respondents = 2 z12rand 50% of respondents = 1 50% of respondents = 2 z13rand 50% of respondents = 1 50% of respondents = 2 GENERAL RULES When considering the data from this module, it is imperative that the user take into account the various reasons that an item was inapplicable to the respondent, and these reasons can be found in the "Inap." category of each variable. These reasons include general skip patterns, sampling strategies, and randomized questions. It is best to use the Internet flowchart in conjunction with the data to understand how respondents move through the Internet module. A response of "don’t know" or "refused" followed the same skip pattern as a response of "no", and it may be assumed that "don’t know" and "refused" should be treated as a response of "no" when tracing any skip patterns unless otherwise noted. VARIABLES gm001he - Is there currently a computer in the respondent’s household 1. Components in this variable: A. z1 2. This variable can be used to describe current personal computer ownership among respondents. Included in personal computers are desktop computers, laptop computers, and WEBTV. gm002he - Was there a computer in the respondent’s household in the past five years? 1. Components in this variable: A. z2 2. Respondents who were assigned to FLUFLAG = 0 and who did not currently have a personal computer did not receive this item and went to the end of the Internet module. For respondents who do not currently have a personal computer at home and who were assigned a FLUFLAG value of 1, this variable can be used to describe past ownership of personal computers among respondents. "Past ownership" refers to personal computer ownership within the past five years, but not current ownership. Included in personal computers are desktop computers, laptop computers, and WEBTV. gm003he - Main reason that the respondent’s household does not have a computer. 1. Components in this variable: A. z3b B. z3b 2. Respondents who have not had a computer in the last five years received this item. Respondents received one of two different wordings of this question with different response choice orders. The version that the respondent received depended on his or her z3rand value. See gms06he and gms07he for the wording of the question and response orders. gm004he - Main reason (other specify) that the respondent’s household does not have a computer. 1. Components in this variable: A. c_z3a B. c_z3b 2. Respondents who gave some other reason as a response to why their household did not have a computer received this item. gm005he - Did anyone in the respondent’s household use the computer to connect to the Internet? 1. Components in this variable: A. z4 2. Respondents who reported that they had a computer in the household in the last five years received this item. gm006re - Did the respondent use the computer to connect to the Internet? 1. Components in this variable: A. z5 2. Respondents who had Internet access in the last five years, but not currently, received this item. gm007re - Did the respondent use the computer to do anything besides use the Internet? 1. Components in this variable: A. z6 2. Respondents who did not personally connect to the Internet when they had it in the past received this item. gm008he - Main reason that the respondent’s household no longer has a computer. 1. Components in this variable: A. z7a B. z7b 2. Only respondents who had Internet access in the last five years, but not currently, received this item. gm009he - Main reason (other specify) that the respondent’s household no longer has a computer. 1. Components in this variable: A. c_z7a B. c_z7b 2. Respondents who gave some other reason about why their household no longer had a computer received this item. gm010he - Does anyone in the respondent’s household use the Internet from home? 1. Components in this variable: A. z8 2. Respondents who reported that they currently have a computer in the household received this item. gm011re - Does the respondent use the computer for anything besides the Internet? 1. Components in this variable: A. z9 2. Respondents who reported that no one connected to the Internet from home received this item. gm012he - Did the respondent’s household ever have access to the Internet from home? 1. Components in this variable: B. z10 2. Respondents who were assigned to FLUFLAG = 0 and who did not currently have an Internet connection from home did not receive this item and went to the end of the Internet module. Respondents who reported that no one connected to the Internet from home and who were assigned to FLUFLAG = 1 received this item. gm013he - Main reason that the respondent’s household does not have Internet access. 1. Components in this variable: A. z11a B. z11b 2. Respondents who currently have a personal computer at home but who have never had access to the Internet from home received this item. Respondents received one of two different wordings of this question with different response choice orders. The version that the respondent received depended on his or her z11rand value. See gms06he and gms07he for the wording of the question and response orders. gm014he - Main reason (other specify) that the respondent’s household does not have Internet access. 1. Components in this variable: A. c_z11a B. c_z11b 2. Respondents who gave some other reason as to why their household does not have Internet access received this item. gm015he - Main reason that the respondent’s household no longer has Internet access. 1. Components in this variable: A. z12a B. z12b 2. Respondents who currently have a personal computer at home and had Internet access in the past, but not currently, received this item. Respondents received one of two different wordings of this question with different response choice orders. The version that the respondent received depended on his or her z12rand value. gm016he - Main reason (other specify) that the respondent’s household no longer has Internet access. 1. Components in this variable: A. c_z12a B. c_z12b 2. Respondents who gave some other reason as to why their household no longer has Internet access received this item. gm017he - Century month when respondent’s household first obtained Internet access 1. Components in this variable: A. z13a_a B. z13a_b C. z13ac_a D. z13ac_b E. z13b_a F. z13b_b G. z13bc_a H. z13bc_b I. adate 2. Only respondents who currently have Internet access received this item. Respondents received one of two different wordings of this question with different response choice orders. The version that the respondent received depended on his or her z13rand value. Respondents were able to respond with either an exact date (e.g. January 2000) or a relative date (e.g. two months ago). Respondents who gave an exact date were assigned a value of 99 in z13a_a or z13b_a. This variable is reported in century months, or months since January 1, 1900. Century months were calculated as follows: A. Exact dates: Months were assigned a value of 1 to 12 representing January to December, respectively. 1900 was subtracted from the year given, and the quotient was multiplied by 12. The value assigned to the month was then added to that value to obtain the century month. Example: Response given was March, 1995 1995 years - 1900 years = 95 years * 12 months / year = 1140 March is the third month of the year, so March = 3 1140 + 3 = 1143 century months B. Relative dates: A raw variable named z1date gave the date when the respondent received the first item of the Internet module. This date was in SAS date format, which is given in days since January 1, 1960. Respondents could give their answers in weeks, months, or years. Three different equations were used to calculate the century month depending on the unit given in the respondent’s answer. 1. Weeks - There are 7 days in a week. The quantity associated with the unit was multiplied by 7 and subtracted from the value in adate. The quotient was divided by 365, the number of days in a year, and then multiplied by 12 to get the number of months AFTER January 1st, 1960. 720, the number of months between January 1st, 1900 and January 1st, 1960, was added to the number to obtain the century month. Final century months were rounded to the nearest whole month (0.5 month and above were rounded to the next highest month, less than 0.5 month was rounded down to the next lowest month). Example 1: Response given was 3 weeks ago. The respondent received item z1 on July 9th, 2003, so adate = 15895. 15895 days - (7 days/week * 3 weeks) = 15874 days / 365 days / year = 43.49 years * 12 months / year = 521.88 months + 720 months = 1242 century months. 2. Months - There are 30.4 days in a month (365 days / 12 months = 30.4). The quantity associated with the unit was multiplied by 30.4 and subtracted from the value in adate. The quotient was divided by 365, the number of days in a year, and then multiplied by 12 to get the number of months AFTER January 1st, 1960. 720, the number of months between January 1st, 1900 and January 1st, 1960, was added to the number to obtain the century month. Final century months were rounded to the nearest whole month (0.5 month and above were rounded to the next highest month, less than 0.5 month was rounded down to the next lowest month). Example 2: Response given was 4 months ago. The respondent received item z1 on July 9th, 2003, so adate = 15895. 15895 days - (4 months * 30.4 days/month) = 15773.4 days / 365 days / year = 43.21 years * 12 months / year = 518.52 months + 720 months = 1239 century months. 3. Years - There are 365 days in a year. ). The quantity associated with the unit was multiplied by 365 and subtracted from the value in adate. The quotient was divided by 365, the number of days in a year, and then multiplied by 12 to get the number of months AFTER January 1st, 1960. 720, the number of months between January 1st, 1900 and January 1st, 1960, was added to the number to obtain the century month. Final century months were rounded to the nearest whole month (0.5 month and above were rounded to the next highest month, less than 0.5 month was rounded down to the next lowest month). Example 3: Response given was 5 years ago. The respondent received item z1 on July 9th, 2003, so adate = 15895. 15895 days - (5 years * 365 days/year) = 14070 days / 365 days / year = 38.55 years * 12 months / year = 462.58 months + 720 months = 1183 century months. 4. To convert from century months to an exact month and year, simply divide the century month value by 12 and add to January 1st, 1900. Some conversions may be affected due to rounding error, as in Example C above. gm018he - Type of connection that the respondent’s household uses to connect to the Internet 1. Components in this variable: A. z14pre 2. Only respondents who currently have Internet access and who are in replicates 5-9 received this item. See gms08re for replicate information. gm019he - Type of connection (other specify) that the respondent’s household uses to connect to the Internet 1. Components in this variable: A. c_z14pre 2. Respondents who reported a type of Internet connection besides cable, DSL. or dial up received this item. All "other specify" responses were collapsed into the general categories of dial-up, high-speed, and not sure/don’t know. gm020re - Does the respondent ever use the Internet from home? 1. Components in this variable: A. z14 B. z14a 2. Respondents who currently have an Internet connection from home received this item. z14 and z14a are the same question; respondents in replicates 5-9 received z14a, which was stored in z14. gm021re - Does the respondent ever use the computer from home? 1. Components in this variable: A. z15 2. Only respondents who do not personally use the Internet from home received this question. gm022re - Estimated time per week that the respondent spends connected to the Internet 1. Components in this variable: A. z16_a B. z16_b 2. Only respondents who personally connect to the Internet from home received this item. It is reported in minutes, though respondents could report their responses in hours per week or minutes per week. gm023sp - Does the respondent’s spouse use the Internet from home? 1. Components in this variable: A. z18 2. Only respondents who currently have Internet access from home and are currently married and living with their spouse received this item, unless they were assigned to FLUFLAG = 0. If the respondent met the first two criteria but FLUFLAG = 0, then the respondent skipped to the end of the Internet module. gm024sp - Does the respondent’s spouse ever use the computer from home? 1. Components in this variable: A. z19 2. Only respondents whose spouses do not use the Internet from home received this item. SUMMARY VARIABLES All summary variables were created from the combination of other variables. All summary variables take on values 0, 1, and -2. The -2 value is assigned if the summary variable cannot be coded because not enough information was available in the component variables for one or more of the following reasons: "don't know" responses, refusals, partial interviews, inapplicable questions, and questions that were not ascertained. gms01re - Fluflag - sampled at 50%. 1. Components in this variable: A. FLUFLAG 2. This variable indicates the value of FLUGLAG that the respondent was assigned to. A 1 indicates that FLUFLAG = 1 and a 0 indicates that FLUFLAG = 0. For more information about FLUFLAG sampling, see memo 157 and COR 770. gms02he - Dichotomous variable - the respondent has ever had a computer. 1. Components in this variable: A. z1 (gm001he) B. z2 (gm002he) 2. This variable collapses gm001he and gm002he into one dichotomous variable to summarize computer ownership. Respondents who have ever had a computer in the past five years, including currently, are represented with a value of 1; respondents who have never had a computer in the past five years are represented with a value of 0. gms03he - Dichotomous variable - the respondent currently has a computer. 1. Components in this variable: A. z1 (gm001he) 2. This dichotomous variable summarizes current computer ownership. Respondents who currently have a personal computer at home are represented with a value of 1; respondents who do not currently have a personal computer at home are represented with a value of 0. gms04he - Dichotomous variable - the respondent has ever had access to the Internet. 1. Components in this variable: A. z4 (gm005he) B. z8 (gm010he) C. z10 (gm012he) 2. This variable collapses gm005he, gm010he, and gm012he into one dichotomous variable to summarize if the respondent has ever had Internet access from home. Respondents who have ever had Internet access from home in the past five years, including currently, are represented with a value of 1; respondents who have never had Internet access from home in the past five years are represented with a value of 0. gms05he - Dichotomous variable - the respondent currently has access to the Internet. 1. Components in this variable: A. z8 (gm010he) 2. This dichotomous variable summarizes current household Internet use. Respondents who reported that someone in the household currently connects to the Internet from home are represented with a value of 1; respondents reported otherwise are represented with a value of 0. gms06he - Dichotomous variable - Form of the question. 1. Components in this variable: A. z3a, z3b (gm003he) B. z7a, z7b (gm008he) C. z11a, z11b (gm013he) D. z12a, z12b (gm015he) E. z13a_a, z13b_a (gm017he) 2. This variable summarizes an experimental condition that was tested on five questions administered in the survey. Two different wordings of the same question were tested, one containing the word "would", and the other not containing "would". For exact wordings of the questions, see the flowchart or the codebook. If the word "would" appeared in the question, then gms06he = 1. If "would" does not appear in the question, then gms06he = 0. The wording of the question was determined by the value of z3rand, z7rand, z11rand, z12rand, and z13rand. gms07he - Dichotomous variable - Form of responses. 1. Components in this variable: A. z3a, z3b (gm003he) B. z7a, z7b (gm008he) C. z11a, z11b (gm013he) D. z12a, z12b (gm015he) 2. This variable summarizes an experimental condition that was tested on four questions administered in the survey. Two different orders of the same responses were tested. Form A of responses was: you don’t think you would find it useful, you think that it’s too expensive, you don’t think you would know how to use it, or you can use it somewhere else; Form B of responses was: you don’t think you would know how to use it, you can use it somewhere else, you don’t think you would find it useful, or you think that it’s too expensive. "Other" was the last response for both conditions. If Form B of responses was used, then gms07he = 1. If Form A of responses was used, then gms07he = 0. The order of responses was determined by the value of z3rand, z7rand, z11rand, and z12rand. gms08re - Dichotomous variable - Replicate that the respondent is in 1. Components in this variable: A. group91 2. Respondents in replicates 0-4 have a value of 0; respondents in replicates 5-9 have a value of 1. gms09he - Dichotomous variable - Respondent has a high-speed Internet connection 1. Components in this variable: A. z14pre (gm018he) B. c_z14pre (gm019he) 2. Only respondents in replicates 5-9 received this item. It collapses gm018he with the "other specify" variable, gm019he. -------------------------------------------------------------------------------- III. Problems -------------------------------------------------------------------------------- A. Individual Cases ---------------------------------------------------------------------------- WLS interviewers were specifically instructed to count reports of WebTV as having a computer. However, in some cases, this was either erroneously entered as "no" to item z1 (gm001he) or was only discovered in a later item after "no" had been entered to item z1 (gm001he). These cases have been recoded as "yes" for items z1 (gm001he) and z8 (gm010he), and their responses to items on the skip path corresponding to "no" for item z1 (gm001he) have been changed to "Inap." (-2). "Not ascertained" (-4) codes were assigned to any questions that were missed due to the interviewer error. No flag has been created for these cases because there are only four of them. These problems were mainly identified when processing notes or "other specify" reasons for not having a computer. PROGRAM UPDATED 6/2007 During completing variable creation for GINET I noticed that the program could be more sophisticated. In Erin's code, a missing raw variable is recoded to inappropriate. When I compared my save file to Erin's there few changes to the analysis variables. After an email correspondence with Jeremy Freese it was decided not to make any significant changes to how the code was written. Namely the other specified codes really should be in the same variable as original code. GMS03HE, GMS05HE, GMS08RE, and GMS09HE were removed from the program as the variables do not have any real value and ultimately make GINET more complex than necessary. ---------------------------------------------------------------------------- B. Survey/Replicate Updates ---------------------------------------------------------------------------- 1. Added items: a. Two items were added to the Internet module for replicates 5-9. Summary variable gms08re can be used to determine to which group of replicates (0-4 or 5-9) the respondent belonged. i. z14pre (see gm018he) & c_z14pre (see gm019he) - What type of connection does the respondent’s household currently have to connect to the Internet? ii. z14a - Does the respondent ever use the Internet from home? This question is identical to z14, and responses to z14a were stored in z14 (gm020re). 2. Changed items or flow patterns: a. (01/10/05) The sibling instrument was altered to correct a flow pattern that collects redundant information. A response of "Yes" to item z4 (gm005he) will take the respondent to item z5 (gm006re); any other response to item z4 (gm005he) will lead the respondent to z6 (gm007re). -------------------------------------------------------------------------------- IV. Coding and Data Editing -------------------------------------------------------------------------------- OPEN-ENDED CODING All open-end response coding and data editing, except for that described under the "Individual Cases" in section III.A of this COR, was completed using a web interface program. Items requiring open-ended coding included: z3a and z3b (gm004he); z7a and z7b (gm009he); z11a and z11b (gm014he); z12a and z12b (gm016he); and z14pre (gm018he). Open-ended coding was only needed if the respondent gave some other response outside of the given response choices. The coding scheme for z14pre is as follows: 01 OTHER HIGH-SPEED - Includes DSL, cable modem, satellite, and services such as roadrunner. 02 DIAL UP - Includes WEBTV 03 NOT SURE/DON’T KNOW 04 OTHER Some respondents listed two types of internet connections that would fit into different categories. For example, suppose respondent X has a cable Internet connection and a dial up connection. In this case, the highest speed connection was coded. Therefore, respondent X would be coded as 01 for this item. "Other specify" responses for z14pre are available in variable gm019he. The coding scheme for z3a, z3b, z7a, z7b, z11a, z11b, z12a, and z12b is as follows: 01 PLAN TO GET ONE/HAVE NOT YET 02 NO INTEREST/DON'T WANT - Also includes dislike of computers or the Internet 03 SPACE CONSTRAINTS 04 TIME CONSTRAINTS - Also includes "waste of time" 05 SECURITY RISK - Includes risk of being robbed, identity theft, and personal information concerns. 06 SOME OTHER REASON - Includes age reasons, computer being broken 07 PROBLEM - Web TV cases. See section III.A. 08 YOU DON'T THINK YOU WOULD KNOW HOW TO USE IT - Also includes being too complicated 09 YOU CAN USE IT SOMEWHERE ELSE 10 YOU DON'T THINK YOU WOULD FIND IT USEFUL - Also includes not needing a computer/Internet or lack of use. 11 YOU THINK THAT IT'S TOO EXPENSIVE 12 PERSON WHO OWNED/USED COMP./INET. NO LONGER IN HH 13 INTERNET - DON'T THINK THAT COMPUTER/EQUIPMENT WOULD SUPPORT IT 14 TEMPORARILY DISCONNECTED - Includes computer being broken with a plan to get a new one and computer not being hooked up yet. 15 SPOUSE OR OTHER HH MEMBER CITED AS REASON - Includes reasons such as, "my wife won’t let me", or "my husband hates computers" 16 DISSATISFACTION WITH COMPUTER/INTERNET WHEN PREVIOUSLY HAD IT Codes 08-12 correspond to reasons that were given response choices, and were collapsed back into the original response choice. For example, if respondent X said that she didn’t have Internet access for some other reason, and then said, "I just can’t afford it", gm013he would be coded as "You think that it’s too expensive", and gm014he would be recoded to inapplicable. If a respondent listed more than one reason, only the first reason was coded, with the assumption that the first reason stated is the most salient. However, if a respondent listed more than one reason in a more ambiguous manner, for example, "all of the above", then it was coded as "other reason" in gm014he because the order of importance is not apparent. DATA CLEANING Data cleaning was completed by processing all interviewer notes within the Internet module. Items with special cleaning procedures are listed below. gm017he - Interviewers were instructed to translate responses such as "2.5 years" to months. In some cases, only a note was left, and a response of 2 or 3 years was entered. Data cleaners corrected these cases to months. Also, some respondents misunderstood the question and responded with when they last connected to the Internet (e.g. an hour ago). These cases were assigned a "Not Ascertained" (-4) code. gm022re - Notes indicating partial units of hours (e.g. 3.5 hours) were recoded to minutes. Respondents who said "not at all" were coded as 0 minutes. -------------------------------------------------------------------------------- V. People -------------------------------------------------------------------------------- Jeremy Freese - Help with problematic cases and COR for problematic cases, major decisions. Tess Hauser - Coding of the Internet module for the CASES instrument. Wes Taylor - Coding of the Internet module for the CASES instrument, UWSC. Erin Wicke - Checking notes, making corrections, developing open ended coding scheme, coding open ended responses, variable creation, and writing the COR. James Yonker - Developing open ended coding scheme, creating access database for variable creation. Conrad Warnke - Took over Erin's duties when she left the WLS. Most notes and corrections were done by Erin. Also did the update discussed in section III. -------------------------------------------------------------------------------- VI. References --------------------------------------------------------------------------------