PBSI STATS 301 EXAM 1

3 goals on science

Description, prediction, explanation

Descriptive Statistics

used to organize/describe data

Inferential Statistics

used to make inferences about a larger group from a smaller group

Sample

group we’re getting data from

Population

group we want to draw conclusions about

Variable

smth that changed/varies for diff. individuals

Data

Info we collect from sample on variables we’re interested in

Continuous Data

data measured on a continuum Ex:) age, iq

Categorical Data

data that sorts people into a category Ex:) college major, eye color

3 measures of central tendency:

mean, median, mode

Mean

Average, sensitive to outliers, most common central tendency measure

Median

midpoint of scores, point where half scores are higher and half are lower, rearrange #’s lowest to highest and find middle, better represents central tendency

Mode

value that occurs most often in data set, most used in categorical data

Variability

tells us how different scores are from each other, helps us understand nature of samply

3 ways to calculate variability:

range, standard deviation, variance

Range

how far apart lowest score is from highest, r=h-l, ignores middle values and only considers extreme scores

Standard Deviation

most common, represents how far away scores are from mean,

Outlier cutoff

2 SD or more

Windsorizing

replacing extreme scores w next highest score + 1

Transformations

mathematical constant applied to all data to pull in ends of distribution

How to find outlier:

Variance

(standard deviation)²

Histogram

represents continuous data

Bar Graph

categorical data

Skewness

lack of symmetry, positive skew= tail to the right, negative skew= tail to the left

Kurtosis

how peaked/flat a distribution is

Platykurtic

LOW kurtosis, flat, platy=playtpus=flat

Journal Article Sections

Introduction, method, results, discussion

Misleading graphs

Scaling Y-axis

Correlation

How do changes in one variable relate to another change?

Scatter plot

Direction of correlation

Positive=variables moving in same direction, Negative=opposite

Strength of correlation

Magnitude of coefficient, r=0=none, r=-1 or 1=perfect, closer to 1= more strong

Limitations of correlation coefficient

Only detects linear relationships, restriction of range, outliers

Correlation coefficient formula

Reporting a correlation:

What we’re reporting ,actual result, conclusion

Coefficient of Determination

r², how much two variables have in common

Coefficient of Alienation

Remaining variance after determination

Measurement

Act of assigning numbers to phenomena according to a rule

Ex:) inches, pounds, GPA

Independent Variable

Creates conditions in experiment, we can change these

ex:) low dose, high dose, no dose

Dependent Variable

Outcome we are investigating

ex:) no change, high change, low change

Nominal Scale

Splits ppl into categories, must be mutually exclusive, nominal = nameable

Ex:) hair color, major, political party

Ordinal Scale

Number is a ranking, ordinal=ordering, not clear how much “distance” separates points on scale- how far is A-tier from S-tier?

Ex:) rank, tier lists

Interval Scale

Ordered events w equal spacing, most commonly used in psychological research, 0 has no meaning

Ex:) IQ test

Ratio Scale

Similar to interval, but 0 has meaning, uncommon (most precise)

Reliability

Consistency or reproducibility of measure/method

Validity

Accuracy of a measure

Observed score

actual number/value on a test

True Score

Exact score w zero error

Error Score

Discrepancy between true + observed score

Ex:) test anxiety, distractions, tired

Test-retest reliability

correlation between 2 assessments

Inter-Item Reliability

How similar are a person’s answers to different items meant to measure the same thing?

Cronbach’s Alpha, 0-1, closer to 1= better

Inter-Rater reliability

Observations made by 2 ppl

3 Types of Validity:

Content, Criterion, Construct

Content

Does the measure sample the entire universe of items that could be used to assess the construct

Criterion

Does our measure reflect/relate to outcomes it “should” w right now or in the future?

Concurrent Validity

Correlation right now

Predictive Validity

Predict outcomes in the future

Construct

Is the measure related other constructs it should be related or and not to ones it shouldn’t?

Convergent Validity

Is measure related to things it should

ex:) self esteem + depression

Discriminant Validity

Is measure NOT related to things it shouldn’t

ex:) self esteem + political party