Paper vs CASI self-completion

This study was relevant to changes toward self-completion being via CASI (Computer-assisted self interviewing) in the main Understanding Society survey in Wave 3. It tests the effects of paper vs. CASI self-completion on substantive measures and attrition. Attrition cannot be evaluated until later waves. The experiment was initiated at Wave 4 and continued until Wave 6.

At Wave 4, half of the Innovation Panel sample received the CASI instrument whereas the other half received the paper instrument. The refreshment sample was included in this experiment.

The controlling variable d_ff_casiw4 is on record d_hhsamp_ip and is coded as follows:

            Group 1 = CASI

            Group 2 = Paper.

The variable d_scac on record d_indresp_ip records the actual use of the self-completion. It is coded as follows:

1          Accepted as self-completion

2          Accepted as self-completion but interviewer to complete due to reading or sight problems

3          Accepted as self-completion but interviewer or someone else to help translate

4          Refused self-completion

5          Not able to do self-completion

At Wave 5, households were randomly allocated to either receive the same self-completion mode as at Wave 4, or the other mode. At Wave 6, only those groups that switched mode at Wave 5 were randomly allocated to the opposite condition. This design will give the opportunity of looking at the effects of (i) different and (ii) changing modes of the self-completion instrument have on the reliability of longitudinal measures. The table outlines the experimental allocation at Waves 4, 5, and 6.

It should be noted that all respondents were allocated to an experimental treatment, but only face-to-face respondents received the actual implementation.

Table: Experimental allocation to self-completion instruments

The Wave 5 controlling variable e_ff_casiw5 is on record e_hhsamp_ip, while f_casiw6 on record f_hhsamp_ip controls the Wave 6 allocation.  Both variables are coded as follows:

            1          CASI

            2          Paper

As with Wave 4, the Wave 5 variable e_scac on record e_indresp_ip records the actual use of the self-completion. in wave 6, the variable is f_scac on record f_indresp_ip. Both items are coded as follows:

1          Accepted as self-completion

2          Accepted as self-completion but interviewer to complete due to reading or sight problems

3          Accepted as self-completion but interviewer or someone else to help translate

4          Refused self-completion

5          Not able to do self-completion

Note, due to a programming error at Wave 5 around 50 per cent of those eligible to receive the questions in face-to-face CASI mode did not get asked the experimental questions (313 people, based on unedited data). It should be noted that this does not confound the experiment (i.e. no respondents were asked questions in the wrong mode), but this error does reduce its power to detect mode differences. Cases affected by this programming error are flagged with the item e_casiflager on record e_indresp_ip.  Please see Section Known Data Issues for details concerning programming errors in the Wave 5 questionnaire.