there will be a statistically significant difference in

indicates where the difference will be found(ie the direction)

* not commonly used in research, except in epidemiology b/c assumed exposure to diseases or agents usually doesn't result in better health!

this variable is the characteristic that the researcher controls or manipulates according to the purpose of the study

used to classify or categorize subjects or groups

a measure of the effect of the independent variable

used to measure outcomes or compare differences

may be measured before and after the conditions are manipulated

these are variables that interfere in a study "bugs", must be eliminated(controlled)

characteristics not of primary interest that affect the dependent variable

Nominal

ordinal

ratio

interval

mutually exclusive categories, but have no order or meaning

male, female,

No mathmatics**

mutually exclusive categories based on some defined characteristics and the relative amt of that characteristic with a logical order to the categories

ex agree, disagree,

never use math on this level of data**

NO math on

Nominal

Ordinal

classifies objects into mutually exclusive categories based on defined characteristics

logical order

and equal units of differences

example= units

can use math

mutually exclusive categories, based on some defined characteristic  and the relative amt. of that characteristic with a logical order, equal units of difference and a meaningful zero point, representing absence of the characteristic

example= # of cigarettes

WBC count

focuses on the consistency of measurement by a particular instrument- what if the instrument is faulty?

the key with instrument reliability is calibration

consistency with which an individual takes measurements

is it done the same by the person every time?

poor technique?

consistenly following technique?

consistency in measurements between or among more than 1 individuals taking the measurements.

must be an adequate assurances that changes are true rather than b/c of error-

*does everyone follow the same protocol for measuring?

0 to 1.00

0.9 or above -excellen reliability

.75-0.89 good to verygood

below 0.75 mod to poor

the lower the worse

1. rateres and ratings do not agree

2. lack of sufficient variability of subjects' scores(range)

does the test or measure actually do what it is intended to do?

are we measuring what we think we are measuring?

what type of validity?

does the mesurement appear appropriate?

what type of validity

is the measurement based on theory?

waht type of validity

is the test broad enough to address the scope of the content?

(example: anatomy, only covers sm. part of the body, not the whole thing)

what type of validity?

how does the test perform?

is it useful?

what type of validity

does the test predict a specific phenomenon or outcome?

what type of validity?

does the test perform as well as an accepted test?

does a smaller test do as good as a bigger, or the gold standard?

what type of validity?

does the test predict well enough to determine an intervention?

history, maturation, and attrition

history-passage of time,

maturation-age(developmental changes, like that of a toddler)

attrition-drop out, ex. if happy with treatment drop out of study, the only ones left are those who are not being helped

testing-the act of testing itself, once testing is done learning is done about test, can't repeat

instrumentation- what if there is a problem with instrument chosen to test (ex in class testing writing with "pencil-like object" but not really pencil, this learning affects handwriting also)

statistical regression to the mean-tendancy for outliers to move toward the mean as test is repeated, due to laws of probability

sampling-bias

1. accessibility-the subject may not represent the entire population, which may include individuals wha have comprimised access to medical care and treatment

2.  subject treatment interaction-everybody reacts differently-we are all genetically different.

1. description of the variables-not given enough data to exactly replicate and get same results

2.  multiple treatments (test order)-potentiation of therapies-fatigue, physiology of test-what was done first

hawthorne effect

individuals anticipate what will be tested, this is an external validity threat

R=

M=

O=

X=

R=randomized

M=subjects are matched

O= means measurement

X=treatment

what type of design is

characterized by

• no control group
• sesitivity to temporal threats to internal validity
• poor generalization

the investigator measures temp. of a sample of children on day before and one day after innoculation

plot and this is an example of what?

O   X   O 

pre-experimental design

"before-after"

no control grom is described

what type of design

• is the gold standard
• randomized-likelihood is balanced
• there is a control group-comparison

experimental designs

randomization means what?

not?

randomization= equivilant

NOT=identical

ex. women with RA were randomly assigned to one of 3 tx groups: heat, cold, contrast baths, edema was measured using an immersion technique one day before tx and later after they received 6 tx's (before tx and after tx measurements)

plot and describe design

this is a pretest-post test control group design using 3 groups. because pts are used a strict control (withholding tx) cannot be used

R  O X₁ O

R  O X₂ O

R  O X₃ O

ex. in order to compare the effects of tPA and no tPA , pts presenting within 4 hrs of a CVA were randomly selected to receiveor not receive tPA. 6 days after admission to hosp. pts. wree compared

do diagram and explain this test

R  X_tPA  O

R X_notPA O

tis is experimental design, but no pretest is possible b/c cannot prediagnose CVA

 do diagram and explain type of test

in order to test the effect of pos. imagery on psycomotor skills 2 classes form the same school were used.  the subjects in 1 class received instruction on visual imagery  while the other class did not.  Subjects' times to complete the psycomotor test battery were measured before and after one clase received visual imagery training

this is a quasi-experiment b/c 2 whole classes were compared and one does not know it is the equivalent

O  X₁  O

-----------

O X₂  O

what type of sample?

any subject that qualifies is included

what type of sample?

all potentioal subjects from a specific source

all subjects are chosen form a protocol, for example, every fifth M and M

is what type of sample?

this type of sample -subjects are chosen so that the composition lof the sample represents the population on specified characteristics

(ex. if asking nurses and 85% are female-make sure 85% of your sample is also female)

want it to look similar

what type of sampling-after a subject is chosen at random, the subject is returned to the population, usually not used in clinical research

odds don't change-slot machine!

50 in 1 is same as 1 in 50

what type of sampling- after a subject is chose at random the subject is not returned to the population

this method is typically used in clinical researc

p= girl/boy, girl

1/2=50%

replacement vs not

drawing the king of hearts in 2 draws

replacement 1/52 AND 1/52

no replacement 1/52 and 1/51

• describes the population sample(descriptive statistics)
• useful as a shorthand
• describes the most typical subject

the most common score

all levels of measurement

may be more than 1 mode

the value occuring at the midpoint when the data is arranged in order of magnitude

useful for skewed distributions

ordinal, interval or ratio data

the sum of the scores divided by the number of subjects (mathematical average)

interval or ratio level data

parameters describe _____

statistics describes_____

parameters- populations

statistics -samples

greek letters are used to denote_____

latin letters are used to denote_____

greek-paramenters

latin-statistics

__

X

if original distribution is trimmed by 10% what does this mean?

in normal distribution -1 to +1 is____%

-2 to +2 is____%

-3to +3 is ____%

65%

95+%

>99%