Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
194 Cards in this Set
- Front
- Back
define measurement
|
process of putting #'s to characteristics according to defined rule
|
|
What is the hierachy of measurement scales
high to low |
RION
Ratio Interval Ordinal Nominal |
|
Nominal Scale
define examples |
classifies objects into categories
least precise no logical order mutually exclusive ex-religion,gender, eye color |
|
data on nominal scale are called?
|
qualitative observations
categorical observations |
|
nominal is described in terms of?
|
percentages
proportions |
|
nominal is most commonly displayed in?
|
contingency tables
bar graphs |
|
Ordinal Scale
define |
gives logical order to classification
categories are mutually exclusive ex-letter grading A.B.C.D |
|
Ordinal
describe observations |
some observations have MORE or are greater than other observations
|
|
Ordinal is desribed in terms of?
commonly displayed in? |
perctages%%%%%%
proportions contingency tables bar graphs |
|
Interval Scale
define example |
scale where any part of the scale reflect EQUAL differences in character measured
0 is just another pt on scale ex-tempature |
|
ratio scale
|
highest level in hierachy
most PRECISE a known/true 0 reflects absence of characteristic |
|
Raw data
define |
facts gathered from population or sample BEFORE anything has been done to the data
|
|
Variables: qualitative
-measured on? |
measured on nominal or ordinal scales
consists of unordered or ordered(ranked) discrete categories |
|
variables: quantitative
measured on? |
measured on interval or ratio scales
|
|
types of variables
3 types |
continuous, discrete, dummy
|
|
continuous variable
define examples |
variable falling in a certain interval, whcih no theorretical restrictions are placed. measured along a SCALE
|
|
discrete variable
define |
have a restriction placed on them
-NO continuity |
|
Dummy variable
define |
assumes a value of 1 if a creiterion is met, a value of 0 otherwise
|
|
examples of discrete variables
|
gender/sex
day of week eye color blood type basically something there isn't a .5 of |
|
descriptive stats
used to? |
classify and summarize numerical data
|
|
types of statisical analysis
|
univariate-1 variable involved
bivariate-relationship btwn 2 variables multivariate- >2 variables |
|
frequency distribution
|
refers to summary of frequency individual values or ranges of values for a variable
tabulation of how many times a # occurs |
|
developing class intervals for freq distributions
min? max? |
no magic formula for ideal widths of class intervals
min of 5 intervals and max of 8 intervals may be <5-gender when poss, width should be odd number |
|
more rules for frequency distributions
|
mutually exlcusive
all inclusive-considers all poss outcomes open ended(more opp) vs close ended(less opp) |
|
absolute frequency
define |
# of OBSERVATIONS in given statistical category
|
|
relative frequency
define |
ratio of absolute frequency to total number of data points
ex- 3's frequency is 6 out of 38 numbers....6/38=15.8% |
|
simple frequency
|
how MANY numbers are in each class
|
|
cumulative frequency distribution
|
add the frequency of any class interval to all intervals below it on scale of measurement
|
|
what does cumulative frequency tell you?
|
how many values arein that interval and all intervals LESS than it
|
|
shapes of frequency distributions
|
shape depends on how scores are distributed
|
|
uniform or rectangular distribution
-happens when? |
when score are evenly distributed throughout a distribution
looks like rectangle |
|
skewness
define |
refers to departures of a distribution from symmetry
|
|
negatively skewed distribution
define |
tail of a distribution points toward LOW SCORES
|
|
positively skewed
define |
distributions with a tail pointing toward HIGH values of a variable
|
|
symmetric
define most familiar example |
two halves of graph coincide if folded in center
DTD!!! bellshaped distribution-known as normal distribution |
|
bell shaped distribution AKA
|
normal distribution
|
|
kurtosis
define |
degree of peakness
|
|
what are 3 degrees of kurtosis
|
1-leptokurtic
2-platykurtic 3-mesokurtic |
|
platykurtic
|
elongated and FLAT
|
|
leptokurtic
|
appears taller and narrow
|
|
mesokurtic
|
bell-shaped like normal curve
|
|
mesokurtic appears...
a perfect mesokurtic is also called? |
appears symmetrical
AKA normal curve |
|
tables
adv vs disadv |
Adv-summarize detail
disadv-reader only sees #'s, may need to compare to other tables |
|
principles of table construction
really??? |
make sure table and txt refer to each other
NOT everything on table needs to be mentioned in txt choose a criterion to group data -size large to small chronology first to last |
|
charts
examples |
histogram
frequency polygon line graph box plots ogive |
|
bar graph
|
illustrate relationship btwn 2 variables when scale of independent variable is normal
|
|
why are columns is bar graph separated
|
separated by small distance because measurements are discrecte not continuous
|
|
bar graphs are used for what type of variables
|
qualitative
|
|
Pie chart
used for ? |
dumb definition
use for qualitative data |
|
histogram
define? |
type of bar graph shows frequencies of individ. scores or scores in class intervals
x axis-reange of scores y axis-frequency of scores |
|
histogram
describe the bars |
intervals are on X-axis
bars are CONTINUOUS |
|
frequency polygon
|
graphical display of frequency table
-scores of each interval is the height of a point located above the middle of interval |
|
scatter plot/scattergram
|
shows relationship btwn 2 quantitavie variables, measured on an interval or ratio scale
|
|
what does a scatter plot show
|
how much 1 variable is affected by another
|
|
scatter plot
define correlation |
relationship btwn 2 variables
|
|
line graph
|
summarizes how 2 pieces of info are related and how they VARY depending on another
|
|
line graph
define scale |
numbers along a side of the line graph
|
|
scatter plot/scattergram
|
shows relationship btwn 2 quantitavie variables, measured on an interval or ratio scale
|
|
what does a scatter plot sow
|
how much 1 variable is affected by another
|
|
scatter plot
define correlation |
relationship btwn 2 variables
|
|
Box Plot
(Box and Whisker Diagrams) what does it show and when is it useful |
shows median, range and inter-quartile range, skewness and potential outliers
useful when comparing 2 or more sets of data |
|
Ogive
define whats the use? |
graph of the cumulative frequencies against the class UPPER boundaries
use to determine the various percentile points in a distribution of scores |
|
what is central tendency
|
refers to an estimate of the CENTER of a distribution of values
|
|
what are the 3 measures of location or central tendency
|
mean, median, mode
|
|
when should mean NOT be used
|
NOT to be used when there are EXTREME values(outliers) in the data set
|
|
median
|
DTD!
middle number in a series or fifieth percentile of set of values NOT affected by outliers |
|
mode
|
the most frequently occuring number
in normal distr. it coincides with mean and median some sets may not have a mode, or may have multiple |
|
dispersion
|
the spread of the values around the central tendancy
|
|
what arethe measures of dispersion or spread
|
Range, Sum of squares, variance, standard deviation, standard error(of mean)
|
|
what is the range?
how to calc? |
the highest value minus the lowest value of data set
50 22 10 90 90-10=80 |
|
sum of squares
define |
measures the observations sum of squares of the deviations from the mean
ALWAYS positive |
|
what does the sum of squares look like
|
WILL BE ON EXAM FORMULA, but still
_ ss=(xi- x)^2 and add them all |
|
variance
define |
its spread of scores around MEAN
its sum of squared deviations divided by the # of observations -1 AKA the mean square |
|
variance
equation |
ON EXAM FORMULA
S^2= SS/ n-1 |
|
standard deviation
definition |
spread or variability of scores around their average
mathematically the square root of variance, so S |
|
standard error of the mean
what the equation look like? |
its the standard deviation of the sampling distribution
se= s/ (square root of n) or square root of s squared/n |
|
Normal distribution
AKA detailed info about it |
aka-Gaussian
its bell curved symmetrical unimodal asymptolic-goes towards, never touches continuous area is 1 or 100% |
|
in the normal distribution the mean median and mode are?
|
mean, median, and mode are EQUAL
|
|
z-Distribution
AKA fun facts |
standard normal curve
or standard normal distribution normal distr. with a mean=0, stand dev and variance =1 curve is symmetric and asympotic |
|
what is the range of z-distribution
|
minus infinity to plus infinity
99.7% of area is within 3 standard deviations of mean z=-3 and z=3 |
|
how to find areas between z scores in standard distribution
|
she had to make a rule cuz people r so DUMBBBB
if desired proportion falls on both sides of mean add two areas if on same side of mean, subtract |
|
z transformation
|
used when mean of normal distrib. is not 0 and stand devis not 1 in order to use standard table
|
|
transformed value also known as
|
z score, normal deviate, standard score, or critical ratio
|
|
whats the formula to transform your raw values into z-scores
|
_
z= (x - x )/s x=raw score, s=stand dev. NEED TO KNOW THIS** |
|
percentiles
define |
point in distribution at or below which a given percentage of scores is found
|
|
students t Distribution
|
family of symmetrical bell shaped that change as sample size changes
specific t distrib for every sample of given size |
|
t distribution
fun facts: mean? |
mean=0
symmetrical variance is greater than 1 range -inf to +inf |
|
shape of t distribution
|
less peaked in center and higher tails than normal distribution
|
|
each t distribution is associated with?
|
a unique # of degrees of freedom
as # of freedom increases, the diff btwn t distribution and normal distr. decreases |
|
degrees of freedom (df)
|
number of observations that are free to vary after sample mean has been calculated
|
|
t table
describe |
each row represents a diff t distribution
all ascores below the mean are negative |
|
what are the 2 types of research for research objectives
|
exploratory
descriptive |
|
definition of research objective
|
clear, concise, declarative statement
states which characteristics of a variable are to be identified and described |
|
definition of research quesitons
|
a concise, interrogative statement
describes variables and identifying relationships among variables |
|
what are types of research in research questions
|
exploratory
descriptive CORRELATIONAL |
|
define hypothesis
|
a conjecture about some phenomenon or set of facts
translates research prob into clear explantion of expected results |
|
uses of hypothesises
|
when the relationship/results of study can be anticipated
|
|
types of research that use hypotheses
|
correlational
quasi-experimental experimental |
|
sources of hypothesis
|
life experiences
literature review medical diagnosis theories |
|
what is inferential statistics
|
procedure for making generalizations
|
|
Inferential stats
what is a population or universe |
ALL members of a specified group
ex-all P3s |
|
inferential stats
sample |
a subset of a population
ex-50 students from p3 group |
|
characteristics of a population are called?
represented by? |
parameters
greek |
|
characteristics of a sample are called?
represented by? |
estimates
roman |
|
parameter examples
mean? variance? standard deviation? |
mean-mu
vairance- o with a long tail squared( O-)^2 (sigma?) stand dev- O- |
|
estimate examples
mean? variance? stand dev? |
. _
mean- x variance- s squared stand dev- s |
|
hypothesis testing invovlves making?
|
inferences about population baed on observations on sample
|
|
what are 2 types of hypothesis testing
|
paramertic stats-involving parameters
nonparametric-yup u guessed it |
|
null hypothesis
noted by |
H0
shows hypothesis to be tested no relationship |
|
alternative hypothesis
symbol? describe |
H1 or HA
is all possible outcomes NOT covered by H0 indicates the direction of rejection |
|
what are the steps in hypothesis testing
|
1-state the hypothesis
2-describe test 3-select probability 4-CCompute test statistic 5-decide about H0 rules not set in stone, can be 8, 10 rules |
|
how to present hypothesis: possible ways H0 can be
|
=
> or = <or= |
|
all possible way H1 can be
|
not equal
< > always opposite of the Ho |
|
converting terms into H0s
no more than- at least- is |
no more than- < or equal
at least- > or equal is, durr = |
|
step 2-determine the test
|
one tail or 2 tailed
id the meaning of the rejection use z or t test |
|
use t when?
use z test when? |
t- < or = to 30 ppl
z > or = to 30 ppl |
|
errors in hypothesis testing
|
error 1-when we reject a true H0
denoted by alpha |
|
alpha
|
type 1 error
also known as rejection region, level of significance, alpha level -alpha= 1- probability value |
|
type 2 error
|
when we fail to reject a false H0
-falsely concluded no significant difference exists when a true diff exists denoted by a beta |
|
if H0 is true and we reject H0....
|
type 1 error
|
|
if H0 is true and we fail to reject H0...
|
correct decision
|
|
if H0 is false and we reject H0..
|
correct decision
|
|
if H0 is false and we fail to reject H0
|
type 2 error
|
|
p value
|
refers to the probability of obtaining the observed diff in a study, if there is no real diff in larger pop
its the prob of being wrong when asserting that a true diff exists or just chance alone |
|
p-value
what does it indicate/ doesnt it |
doesnt indicate the clinical importance of result
dont tell u about size of direction of effect may be statistically significant but clinically unimportant |
|
if p value is < or = to alpha then....
|
reject H0
|
|
if p value is > alpha....
|
fail to reject H0
|
|
relationship between alpha and beta
|
inverse
decreae alpha=increase beta |
|
what is most common way to decrease type 1 and 2 error
|
increase the sample size
|
|
step 4 compute the test statics
|
compare the calculated value with the table value(critical value)
|
|
statistical significance
definition |
probability that a result is not due to chance alone
|
|
clinical significance
definition |
extent to which the results of a study will be impt useful and APPLICABLE in treating pts
|
|
the goal is to achieve...
|
both statistical and clinical significance
statistically significant results may not be deemed clinically significant, and visa vers |
|
one sample case
|
comparing observed value(the mean) with the hypothesized value of population
|
|
one sample formula for z
|
_
z= x - mu/ (s/ square root of n) x is mean sample, mu is mean population s is stand deviation |
|
1 case formula for t
|
same as z
|
|
critical value
define |
values that represent beginning of rejection regions
value that is compared to determine wether or not H0 is rejected |
|
purpose of surveys
|
to obtain info necesary to fullfil the purpose of study that cant be done any other way
|
|
roles of research in medical research
|
assess burden, risk factors, or behaviors to diseases
needs assesssments measures access to health care QOL |
|
more roles of surveys in medical research
|
evaluate programs or interventions
eval access, dissemination and use of info |
|
what are the types of survey study designs
|
1-cross sectional
2-cross sectional surveillance 3-cross sectional intervention 4-cross-sectional comparison 5-cohort study 6-panel study |
|
cross sectional surveys
|
describes pop characteristics at 1 given point in time
data is collected ONCE sample-from large group |
|
cross sectional surveillance
|
for planning or observing changes over time as consequence of normal events
data collected at multiple times ex-same group evaluting president every year |
|
cross sectional intervention
|
used for evaluating intervention
data collected multiple times assumes no change in population |
|
group comparisons
|
used to test an association
data collected at same point for 2 groups retrospective-ONCE |
|
cohort study
|
used to observe changes over time in comparative groups or to test an association
data is collected at baseline and reasseed after time prospective-MULTIPLE times |
|
panel study
|
used to make indepth descriptive observations
data collected multiple times used over long periods of time-years to decades |
|
etical considerations in survey research
|
respect
informed consent |
|
take steps to minimize and avoid negative effect by:
|
maintaing anonymity
keeping access to data only to essential personnel IRB documentation |
|
more ethics in survey research
|
collect data in timely manner
avoid duplication publication of results |
|
examples of survey designs
mail surveys |
survey is mailed
designed to accommodate respondent requires-clear instructions, simple language, logical order of questions |
|
mail surveys
adv vs disadv |
adv-cheaper, avoids interview BIAS, used as screening tool
disadv-responses accepted as final, low response rate-need 50-80% for publication |
|
examples of survey designs
web surveys |
like mail surveys
easy access for respondnets |
|
web surveys
disadv vs adv |
adv-cheap, simple
disadv-unsure of respondant computer access and abilities limits this |
|
examples of survey designs
interview surveys |
we read question, and record answer
good for open ended ?'s allows for in depth responses ppl will answer with only socially acceptable responses |
|
interview surveys includes?
when to do? |
phone and face to face
when precise responses needed complex question items |
|
telephone vs face to face
|
face to face is more exp
easier to get sensitive info over phone if long-in person is better flexible-phone better |
|
what is a research question
|
a statement that defines issue, condition or phenomenon to be studied
driven by ided prob or need |
|
what is a survey objective
|
defines what the survey aims to address
if met, research questions will be answered |
|
measuring facts in survey research
|
behavior-"during past month have you drank redbull"
knowledge-"are you aware of whats in redbull" opinion-varies in times attitudes-position of given opinion |
|
how to formulate questions in survey research
|
specific
avoid vague words and adverbs avoid leading questions |
|
response styles for research
|
check boxes
pictures slide rule |
|
survey reliability
|
researcher needs to ensure person would answer question in the same way each time asked, even if asked 2 diff ways
|
|
survey validity
|
need to ensure that questions are asked measures the true staement
|
|
what are the 3 types of validity
|
face validity-questions that are in survey make sense to the responders
content validity-questions dont lead to certain responses construct validity-questions relate to content/ objective of the study content validity- construct validity- |
|
how to improve reliability and validity
|
pre-testing and pilot testing survey
use perviously tested or standardized tests |
|
after writing question ask yourself..
|
1 is the question clear?
2 any possibilities for misinterpretation |
|
pooling define
|
when divide by n-> giving greater weight to biggest group
|
|
z + t are the
|
ratio of diff of means + standard error
|
|
what does 2 group comparison do
|
its the diff btwn 2 sample averages
|
|
conficdence interval
|
rangeof values in which your confident that actual pop value is within that range
|
|
diff btwn z + t
|
t < 30 observations, NL curve NL distribution
z> 30 observations bell shaped curve....after 30 degrees of freedom z + t look same both have s=1 and mean=0 |
|
define concept
|
label for an abstract idea
set of objects that shar characteristics and common name |
|
constructs define
|
highly abstract concepts
|
|
operationalization define
|
refers to the translation of a concept or construct into a functioning and operating reality
|
|
measurement allows for
|
quantification
classification comparison |
|
attribute
|
particular features of an object along various dimensions
|
|
attitudes
|
refers to enduring patterns of belief which it is believed are predictive of behavior
|
|
scaling
|
construction of an instrument that associates qualitative units
|
|
thurstone equal appearing interval scaling
|
measuring peoples attitudes along single dimension by asking if they agree or disagree
choice of 2 possibilities |
|
people should exercise if they want to be healthy
agree____ diasgree____ example of? |
thurstone
|
|
guttman scales
other name? purpose? |
cumulative scaling
purpose-1-dimensional continum for concept we wish to measure focuses on 1 concept limited use |
|
guttman
series of what? |
of increasing intensity is tanked in order
most extreme to least then add up points ex-some kids require physical restraint(least) ... ... ... .. some kids should be beaten QD!!(EXTREME!!, not really but yea) |
|
Paired comparison scale
|
given 2 choices at once,(frequency/ speed) or (speed/price) asked to select more important one
|
|
rank order scale
|
putting a number to a bunch of given statements based upon the rank given(like rank based upon time spent)
|
|
q sort technique
|
eveloped to discriminate among a relatively large # of objects
object->piles->order of preference |
|
q sort technique
example |
SORT items according to those with which you most agree and those u most disagree
|
|
constant sum scale
|
subject assigns a # of points among alternatives
more points=more importance |
|
line marking scale
|
to what extent do you feel about the Nazis killing the jews
<------------------------------> horrible EPIC |
|
smiling faces
|
indicate how you feel based on smiley faces DTD
|
|
Likert scale
|
given a range of possible responses to positive and negative statements
give points per response |
|
ex of likert scale
|
ppl should exercise to be heatlhy
strongly agree agree neutral disagree strongly disagree |
|
semantic differential scale
|
paired polar opposite adjectives
pharmacists are: exciting_ _ _ _ _ _ _boring |
|
stapel scale
|
rxn to commerical
+3 +2 +1 exciting -1 -2 -3 similar to semantic but with only 1 onjective |