Introduction

Methods

Results 

Discussion

- take bits of info and make general assumption

- derive clinical hypothesis through induction

RANDOMIZATION! (R)

MANIPULATION! (X AND O)

CONTROL!

ROXO

- recruitment:representative of population of interest ex: stroke

- assignment: internal validity of study...not involved in putting (assigning) to a specific group; this can be done by a computer/3rd party

- When we "fiddle" with something/change

- independent (X): cause: thing we change; typically treatment

- dependent (O): effect: variable; changes due to manipulation

CONTROL

ROO (control group)

Why?--> see if tx worked better? For example, could have been a lapse in time and client could have just gotten better

- gives us basis of comparison

- interested in difference b/w control and experimental group

theory principle which believe is general and apply to individual instance/specific

ex: people respond well to kindness, therefore I will be kind to this individual and they in turn will be kind to me

- Opposite from induction!

- What's the probability/chance that this # is true

- CI tells you the potential spread of real answers could get

- Tells us what indications are for this particular analysis

STATISTICAL SIGNIFICANCE

- chances that a difference bw two groups will be found when there isn't a true difference

- telling you how "trustworthy" finding is

-a bout the mean difference

- lower the #, the less chance that this result could be accidental (statistically significant)

ex: p=.001=1/1000 chance--> smaller the p value= GREATER the confidence

- did NOT occur just by chance

- conclude hypothesis supported when in fact it was NOT

- FALSE POSITIVES

- you find difference where there actually is NOT one

- book: conclusion that there is a relationship bw variables when in fact there is NOT

- reject when actually support

- (book): conclusion that there is NO relationship bw the variables when in fact there IS

- analysis of other data on same subject

- gives you more of a sense what's going on

ONE TAILED

- tails: end of normal distributions

- the results are expected to go in one particular direction

- ex: male clinical therapy students are MORE likely to fail clinical assessment than females (can only go one way)

- results could either go in two directions

- ex: student clinical therapists w/ 'A'-level physics differ from students w/OUT 'A'-level physics in term of their theory exam.... could go in two directions

- agreement bw two individuals during process of making rating scales

- book: the degree of consistency in the measures taken by two or more raters

- UNTIL WE CAN OPERATIONALIZE CONCEPT (CONSTRUCT) THEN WE CAN'T PROCEED!

- has to do with recruitment

- Are working with representative of population of interest?

- stability of ratings through time

ex: someone 2 yrs post stroke...and then 2 weeks later test, do they get same score?

- exposure to stimulus makes it easier

(his example of APPLE, KEY, ORANGE)-- easier to remember the second time around

Nominal

- very basic level: "categorical data"

- numbers are just labels...can't really add together

- no common value attached to the labels

ex: male/female, hair/eye color

- rating things in terms of 0, 1's & 2's

-  hierarchically organized

- allows the researcher to rank order subjects along the dimension of interest (ex: most improved to least improved), the difference between the points are NOT equal

- techniques of statistical analysis which are said to be robust and sensitive

- they require certain conditions or parameters to be fulfilled before they can be applied, the most important of which is that the data must be interval/ratio

- #'s assigned have equal distances bw data points

ex: height, blood pressure, etc

- has NO absolute zero