TALENT AQC.- Ch. 8

measurement

translating concepts into numbers

variables

concept being measured usually columns in dataset

data

observed values

predictor variables

-independent variable

-x

-used to predict outcomes

-ex. hiring test scores

criterion vairbales

-dependent variable

-y

-outcome variable

-what is being predicted by predictor variables

-ex. job performance

what are the more levels of measurement

-nominal

-ordinal

-interval

-ratio

-gets more sophisticated as you go down the list

nominal

categories, no numerical value

example of nominal

student names/student IDs

ordinal

ordered categories

example of ordinal

being a junior vs. a senior bc of credit hour differences

interval

equal intervals b/w values

example of interval values

temperature differences like 80/79 is the same as 31/32 degrees

ratio

interval level characteristics plus a meaningful zero

descriptive statistics

summarize a data set with a picture or graphing numerically

what are the measures of central tendency

-mean

-median

-mode

mean

average

median

mid point (1/2 above, 1/2 below)

mode

most frequent

what are the measures of variability

-variance

-standard deviation

range

largest observation minus smallest observation

variance

square deviations from the mean

standard deviation

square root of variance

nominal distribution

frequence distribution of scores that is symmetrical, with the bulk of frequency in the middle

raw score

original form of data

standard scores

converted from raw data and is the number of standard deviation unions from the mean

standard deviation equation

raw score-mean/standard deviation

positive z

above mean

negative z

below mean

percentile

% of sample observations that fall below a given raw score

correlation coefficents

describe relationship b/w 2 variables

R

-tells direction and strength

-range from -1.0 to 1.0

a negative/positive correlation:

-tells the direction it is going

-can still be a strong correlation while being negative

sampling error

random variation across samples

statistical significance

degree to which observed relationship is not likely due to sampling error

If P-value is <.05:

statistically significant result

regression

uses statistical significance testing to predict one variable (y1 criterion) using data on other variable(s) (x predictors)

Y

predicted performance