GHQ-12
1. Able to concentrate2. Lost much sleep3. Playing a useful part4. Capable of making decisions5. Under stress6. Could not overcome difficulties7. Enjoy normal activities8. Can face up to problems9. Feeling unhappy and depressed10.Losing confidence11.Thinking of self as worthless12.Feeling reasonably happy
INSTRUCTIONS
We should like to know if you have any medical complaints, and how your health has been in general, over the past few weeks…..
We want to know about present and recent complaints, not those you had in the past.
12 item GHQ (6 positive items)• Item 1
– able to concentrate• Better than usual BTU• Same as usual SAU• Less than usual LTU• Much less than usual MTLU
• Item 3 – playing a useful part
• More so than usual MSTU• Same as usual SAU• Less useful than usual LUTU• Much less than usual MTLU
• Item 4– felt capable of making decisions
• More so than usual MSTU• Same as usual SAU• Less so than usual LSTU• Much less capable MLC
• Item 7 – able to enjoy day-to-day activities
• More so than usual MSTU• Same as usual SAU• Less so than usual LSTU• Much less than usual MTLU
• Item 8 – able to face up to your problems
• More so than usual MSTU• Same as usual SAU• Less able than usual LATU• Much less than usual MTLU
• Item 12– feeling reasonably happy all things considered
• More so than usual MSTU• Same as usual SAU• Less so than usual LSTU• Much less than usual MTLU
12 item GHQ (6 negative items)• Item 2
– lost much sleep over worry• Not at all NAA• No more than usual NMTU• Rather more than usual RMTU• Much more than usual MMTU
• Item 5 – playing a usefu part
• Not at all NAA• No more than usual NMTU• Rather more than usual RMTU• Much more than usual MMTU
• Item 6– felt capable of making decisions
• Not at all NAA• No more than usual NMTU• Rather more than usual RMTU• Much more than usual MMTU
• Item 9 – feeling unhappy and depressed
• Not at all NAA• No more than usual NMTU• Rather more than usual RMTU• Much more than usual MMTU
• Item 10 – been losing confidence in yourself
• Not at all NAA• No more than usual NMTU• Rather more than usual RMTU• Much more than usual MMTU
• Item 11– thinking of self as worthless person
• Not at all NAA• No more than usual NMTU• Rather more than usual RMTU• Much more than usual MMTU
ILLEGAL & WRONG !!!
Goldberg’s GHQ-12:BHPS VARIABLE & RESPONSE LABELS
ighqa : ghq: concentration 1: better than usual 2: same as usual 3: less than usual 4: much less thn usual
ighqb : ghq: loss of sleep 1: not at all 2: no more than usual 3: rather more thn usul 4: much more than usual
ighqc : ghq: playing a useful role 1: more than usual 2: same as usual 3: less so 4: much less
ighqd : ghq: capable of making decisions 1: more than usual 2: same as usual 3: less so 4: much less
ighqe : ghq: constantly under strain 1: not at all 2: no more than usual 3: rather more 4: much more
ighqf : ghq: problem overcoming difficulties 1: not at all 2: no more than usual 3: rather more 4: much more
ighqg : ghq: enjoy day-to-day activities 1: more than usual 2: same as usual 3: less so 4: much less
ighqh : ghq: ability to face problems 1: more than usual 2: same as usual 3: less so 4: much less
ighqi : ghq: unhappy or depressed 1: not at all 2: no more thn usual 3: rather more 4: much more
ighqj : ghq: losing confidence 1: not at all 2: no more than usual 3: rather more 4: much more
ighqk : ghq: believe in self-worth 1: not at all 2: no more than usual 3: rather more 4: much more
ighql : ghq: general happiness 1: more than usual 2: same as usual 3: less so 4: much less
Latent Goldberg: What kinds of models are required to ask what research questions?Latent Goldberg: What kinds of models are required to ask what research questions?
1. Latent variable model (IRT/LTA or CDFA)2. Discrimination & Difficulty params ICCs3. Information function (effective measurement)4. Scoring methods for -ve items (cGHQ)
1. Model-based vs simple sum score (linear models, interval (metric) scores)
2. Person (cluster) vs variable centered analyses3. IRT models with ordinal or nominal indicators
Duncan-Jones et al (1986), Psych MedGoldberg et al (1987) Psych Med
Whittington & Huppert (2003) SSMPloubidis et al (2007) P&ID
Focus on negative items• Normal ogive item response models WLS
• Logistic distribution item response models ML
Focus on GHQ positive items• Equal interval scoring as 2-1-0-0 Interval?
• Coding “more than usual” 1-0-0-0 + ve items
Q … the characteristics and performance of GHQ items as measures of distress, and the
estimated or assumed dist’n of the latent trait
Q … the potential for some items to be scored to capture a notion of positive mental health.
Does the GHQ “do” well-being?
Psychiatry Psychology
Standard statistical approach - under applied Novel scoring approaches, under-validated
Positively worded items scored 1000
F1
GHQPos
F2
GHQNeg
GHQ03 GHQ07 GHQ12GHQ01 GHQ04 GHQ08
GHQ05 GHQ09 GHQ11GHQ02 GHQ06 GHQ10
Orthogonal
Negatively items scored 0011
Setting the scene
• Software use: what do you use?– SPSS– SAS STATA – R/Splus MATLAB
– Specialist: Psychometric• BILOG MULTILOG TESTFACT PARSCALE• Rasch software
Psychometric papers
• What statistics/statistical analyses do you associate with a “psychometric” paper– What would you use a computer to calculate for you?
• Why are you doing that? Why only that?
• What else could you do and how?• Increasing your options and understanding of the range of
psychometric scaling approaches and statistic analyses• Psychometric analyses, and increasing the accuracy with
which you describe, execute and combine them
Example: GHQGeneral Health Questionnaire
See GHQ Users Guide and Goodchild-Duncan Jones (Ref pdf)
This instrument has no numbers on it
Scoring begins from assignment of
1, 2, 3, 4 (weights)
i.e. consecutive integersto the ordered levelsof the graded responses
[Scoring methods alter weights/Scaling models estimate weights]
Be careful with question wording
GHQ is balanced in terms of + & - question wording [what does this mean?]
Why – and what implication(s)?
NB: If response options are consistently (re)-worded, this avoids the need for reverse scoring]
GHQ-30 data from a population survey: HALS (Baseline)
GHQ-30 data from a population survey: HALS (9 items)
New portable C:\HALS_CGHQ_negativeonly.dta
Time to share out well tested
Data preparation pipeline(semi-automated)
Take it away Jon