Data Issues

Waves 1 to 7
Data issues - Waves 1 to 7 – February 2019

31 Gambling data inconsistencies

In wave 7 the B and K cohort Parents and K cohort Study Children were asked about their participation in gambling activities over the past 12 months.

gse26a1a - gse26a1j (Parent 1 B)/ ise26a1a - ise26a1j (Parent 1 K)/ ise26c1a - ise26c1j (Study Child K) - "During the last 12 months have you spent money on any of the following (Scratchies, Bingo, Lotto, Keno, Private betting, Poker, Casino games, Poker machines, Horse or dog races, Sports betting)?"  Yes or No

For "Yes" responses to above question, participants were then asked about the not online and online frequency of the gambling activity they had spent money on.

gse26a2a - gse26a2j (Parent 1 B)

ise26a2a - ise26a2j (Parent 1 K)

ise26c2a - ise26c2j (Study Child K)

Thinking about the past 12 months how often have you participated in "Word Sub gambling activities specified in previous question" not online

gse26a3a - gse26a3j (Parent 1 B)

ise26a3a - ise26a3j (Parent 1 K)

ise26c3a - ise26c3j (Study Child K)

Thinking about the past 12 months how often have you participated in "Word Sub gambling activities specified in previous question" online?

There is some inconsistency in the data when comparing the "Yes" responses to the top question (gse26a1(P1B)/ ise26a1(P1K) / ise26c1(SCK)) with the subsequent questions regarding frequency of not online and online gambling.  153 B parent records, 137 K parent records and 112 Study Child K records have indicated at the top level question that they had spent money on particular gambling activities in the past 12 months, but then in the subsequent questions about the frequency of this activity in the past 12 months they have answered "never" or "not in the past twelve months" to both not online and online, which is inconsistent with their response to the first question.  As a result, the derived variables for "method gambled in last 12 months" gse26a5 (P1B) /ise26a5 (P1K) / ise26c5 (SCK) have been removed from release 7.1 as they were derived from the not online and online frequencies. There is also a difference in the response values for the not online and online frequency of gambling activity when looking at parent (B and K) and study child items (K). Parents had a value range of 0 to 6 whereas the Study Child had a value range of 0 to 8

Users should be cautious when working with the not online and online frequency data and be aware that some inconsistencies may be highlighted if data is compared to the other gambling questions.

Users should also note that some minor corrections have been made to the gambling data since release 7.0.  

  • The code frame values in the main dataset for gse26a2a - gse26a2j (P1B), ise26a2a - ise26a2j (P1K), gse26a3a - gse26a3j (P1B) and ise26a3a - ise26a3j (P1K) have been updated to 0 through 6 to match the data dictionary and the labelled questionnaires. The code frame for these items in first release was incorrectly showing 1 through 7 in the main dataset.
  • Two records in the B cohort parents dataset have also had some gambling response values corrected from -9 (n/a) to -3 (refusal) for variables gse26a2a - gse26a2j and gse26a3a - gse26a3j. This was to correctly reflect where the respondent had refused to answer the question in the instrument rather than not being asked the question.