This experiment explores the impact of the presentation of Don’t Know/Prefer Not to Answer response options in electronic self-completion questions. Currently in Understanding Society, the initial presentation of a question in a self-completion mode presents the question with neither the Don’t Know/Prefer Not to Answer options showing. When a respondent tries to go to the next question without providing a response to the initial asking of the question, the question is represented with the Don’t Know/Prefer Not to Answer options available in a different colour (blue) with a prompt to select a response. There are no initial instructions in the survey as how questions will be asked in relation to giving Don’t Know/Prefer Not to Answer responses.
To explore the impact of presentation variations, three conditions are used in face-to-face CASI modules or the analogous web version. First is a control condition, presenting Don’t Know/Prefer Not to Answer options as is currently done in Understanding Society. Second is as is the control condition, but with a specific prompt on the initial screen to inform respondents how they can view additional options. Third is inclusion of Don’t Know/Prefer Not to Answer options as part of main response lists.
The experiment covers several modules contained in the face-to-face CASI module toward the end of the survey: the SF-12, GHQ, Young Adults, Child Development, and Non-resident relations. Additionally, there were two additional attitude questions on issues of nuclear energy and the United Nations that were also be included in the response presentation experiment. For these new additional questions, respondents were also asked follow-up questions about their self-rated level of knowledge on these topics. For those who provided a Don’t Know response but indicated little knowledge on the topic, a follow-up question on reasoning was asked. Similarly, those saying they have a lot of knowledge but provided a Don’t Know response on the attitude question were also asked a follow-up question about reasoning.
Individual respondents were allocated equally to these three conditions (i.e. 1/3 chance for each condition). The controlling variable for this experiment is as follows:
dkcond on record k_indresp_ip:
1 {Control, standard procedure}
2 {Instructions on DK procedure given first}
3 {DK/REF response options offered in the initial presentation of question}
The following variables used for this experiment are on record k_indresp_ip:
nuclen, untrust, knownuclen, knowun, whynucans1_code, whynucans2_code, whyunans1_code, whyunans2_code, whorupro, whoruedu, whorurac, whorupol, whorufam, whorusex, whoruage, scsf1, scsf2a, scsf2b, scsf3a, scsf3b, scsf4a, scsf4b, scsf5, scsf6a, scsf6b, scsf6c, scsf7, scghqa, scghqb, scghqc, scghqd, scghqe, scghqf, scghqg, scghqh, scghqi, scghqj, scghqk, scghql, ncrr1, ncrrm, ncrry4, ncrr3, ncrr4, ncrr5, ncrr6, ncrr8, , ncrr9, ncrr11, ncrr12
The following variables used for this experiment are on record K_chdev_ip:
cdvla, cdcond, cddis, cdlmt, cd3pera, cd3perb, cd3perc, cd3perd, cd3pere, cd3perf, cd3perg, cdvll, cdvlb, cdvlc, cdvld, cdvle, cdvlf, cdvlg, cdvlh, cdvli, cdvlj, cdvlk, chsdqc, cdvlm, cdvln, cdvlo, cdvlp, cdvlq, cdvlr, cdvls, cdvlt, chsdqa, chsdqb, chsdqn, chsdqd, chsdqe, chsdqf, chsdqg, chsdqh, chsdqi, chsdqj, chsdqk, chsdql, chsdqm, chsdqy, chsdqo, chsdqp, chsdqq, chsdqr, chsdqs, chsdqt, chsdqu, chsdqv, chsdqw, chsdqx, cdtvvidhrs, cdcread, cdoread, cdwread1, cdwread2, cdwread3, cdwread4, cdwread5, cderead, mealsreg, bedreg, cdtvvidhrw, cdcomp, cdpchrs, cdconsol, cdconstm, cdphsc, cdphscy1, cdphscy2, cdphscy3, cdphscy4, cdphscy5, cdphscy6, cdphscy7, chrisk, chpat, chdelay
