The K6 consists of six questions that ask subjects to rate how often they felt (i) nervous, (ii) hopeless, (iii) restless or fidgety, (iv) so depressed that nothing could cheer you up, (v) that everything was an effort, and (vi) worthless over one of two recall periods: the past-month (respondents are were asked to rate how often the symptoms occurred in the 30 days before the survey) and the worst-month (respondents are asked about the 30-day period during the past 12 months when they had the most severe psychological distress). Some WMH surveys used only one of these recall periods while others used both. The decision about which recall period to used hinged on whether the investigators were interested in calibrating SMI point prevalence (most useful for screening in clinical settings), 12-month prevalence (most useful for estimating prevalence in surveys used for health policy planning purposes, as the year is the usual health policy planning period), or both.
The surveys that used both recall periods began by administering the past-month questions and then asked respondents a single question about whether there was any other 30-day period in the past 12 months when they had these symptoms more frequently than in the past 30 days. If not, the past-month responses were also used as the worst-month responses. If the respondent reported that there was a worst month, though, he or she was asked to think about that time in answering the six questions a second time. The six K6 questions had to be repeated for about 20% respondents when this two-part approach was used. That is, about 80% of the time respondents reported that there was no other 30-day period in the past 12 months that was worse than the last 30 days. The response options, which were identical in the two recall periods, were all of the time, most of the time, some of the time, a little of the time, and none of the time. These were coded 4-0, which means that the unweighted summary scale has a 0-24 range. However, it is also possible to weight either the scale items (as in a factor analysis factor-weighted scale) (Kim, 1993 ) or to weight the item responses within each item (as in an analysis of nested dichotomous items in an item response theory [IRT] modeling approach) (Embretson and Reise, 2000 ).
The surveys that used both recall periods began by administering the past-month questions and then asked respondents a single question about whether there was any other 30-day period in the past 12 months when they had these symptoms more frequently than in the past 30 days. If not, the past-month responses were also used as the worst-month responses. If the respondent reported that there was a worst month, though, he or she was asked to think about that time in answering the six questions a second time. The six K6 questions had to be repeated for about 20% respondents when this two-part approach was used. That is, about 80% of the time respondents reported that there was no other 30-day period in the past 12 months that was worse than the last 30 days. The response options, which were identical in the two recall periods, were all of the time, most of the time, some of the time, a little of the time, and none of the time. These were coded 4-0, which means that the unweighted summary scale has a 0-24 range. However, it is also possible to weight either the scale items (as in a factor analysis factor-weighted scale) (Kim, 1993 ) or to weight the item responses within each item (as in an analysis of nested dichotomous items in an item response theory [IRT] modeling approach) (Embretson and Reise, 2000 ).
