W2 L3 Dev

Question 0

Epidemiology

Answer 0

Hinges on sampling and measurement, only as good as those

Ask: how COMMON is a given problem
And
Who has the given problem? Scattered, clumped?

Identify rates of a problem in the population, incidence. And incidents:
Prevalence all cases.
Incidence,

Must have random sample.
Want to see how common a given problem is.
6 months prevalence rates of psychiatric disorders for children of Ontario.

Must identify random sample, and get them to participate.
Then ask the questions must be specific and understand how different answers can come about.

How to measure construct and who going to ask to answer q. Imp. Covered next class.

The who:
Age, other demographics…

Remember, correlation does not mean causation.
Ses leads to x, but which effects which.

Question 1

Review from last week

Answer 1

Measurements, reliability,
Validity. Measuring what I think it is measuring,
Sampling, population,
External validity, how much result can be generalized.

Think about how convenience sample differs from the whole population,

Question 2

To establish causality

5 things

Answer 2

1. Need theory that a causes b
2. Must be related empirically
   Ex. Show that can have one or other, but need both for causality
3. Then must eliminate other possible causes
4. Establish temporal order
5. Demonstrate responsives, manipulating the variable.

Question 3

Correlate of disorder

Answer 3

Natural experiement or case control designs.
Rates of psychopathology in children with and without abuse,
Groups must be comparable in other ways, k

This is a case control design!
Ex. In patients, and schools.
Social interaction issue,

Seeking treatment vs not, are very different.
Don’t know if it is depression, or a mental health problem. Need to control for this in inpatient clinic.

Question 4

Case control advantages

Answer 4

Ethical, and allow to answer meaningful questions.
If careful of control groups can get good internal validity.

But!
Does not establish temporal order,

Does it establish a causes b

Question 5

What are prospective designs.

Answer 5

Is a longitduinal study. Once children have been followed over time.
Helps to establish TEMPORAL ORDER
Answering:
Who goes on to develop a disorder,
What happens to children who HAVE the disorder.

Helps to identify risk and protective factors -> disorder.
Ex. Child abusers and conduct disorder.

Follows the before and or after onset of symptoms.
Tracking change or stagility,

Challenges include:
Internal validity: have to rule out other factors.
Eg. Child psychopathology and ses in adulthood.
Negative associatn found by Dirks. But did she control for ses? She did show this!

Advantages:
Establish temporal order
Not relying on retrospective report.

Disadvantages
Takes a loooooong time.
Hard to keep track I guess. Establishing internal validity is very tricky. Age is also a confound,
Ging effects issue then.
May need to recruit an enormous sample! Esp. For rare disorders to occur over time and have a good sample.
Does not establish a changes B

Question 6

Intervention studies.

Answer 6

Evidence based treatment. 
Scientific evidence of treatment helping. 
Best PRACTISE guidelines. 
Approaches used combined:
1. Scientific approach. 
Identify treatments for scientific efficacy. 
Expert consensus approach 
  Gaps filled in. 

Well established means more than 9 single case studies.
Or at least two between group designs.

Single case experimental design.
Repeated measures, look for replication of treatment effects.
Get baseline, intervene, get another baseline, intervene again to see.
Abab reversal design.

Seat exiting behaviour eg.
Popsicle stick and stickers wow. Choice and reward there.

advantages: 
Internal validity. 
External events controlled for, as removal of intervention shows return to baseline! 
Temporal ordering, 
A changes B

But
External validity
Not so much generalize able. Need 9 of them!
Can be really hard to interpret findings.
Stable change potential problem, no return to baseline.
Ethics, that’s why for serious issues, withdrawing

Question 7

Group based designs

Answer 7

Rct's, randomized controlled trials. 
A therapy experiment, psychotherapy experiement. 
Have experimental (treatment) and control. 

Random assignment is all participants equally opportunity tt of being in either group.
Internal validity increased, due to no differences between the groups, means intervention has more power.