Quiz 1

Question 1

Structured vs. Unstructured data

Answer 1

structured = quantitative/properties that vary in type of attribute/variable

unstructured = qualitative (cannot be categorized), properties that differ in amount

Question 2

CCHS

Answer 2

Canadian community health survey

cross-sectional survey for health surveillance, health care utilization and health determinants

goal: to provide single source data for health researchers

every 2 years, 2x24hr food recalls, largest

Question 3

CHMS

Answer 3

Canadian Health Measure Survey

questionnaire data (household interview) & physical measurements for baseline NCD and exposure to infectious diseases and environmental contaminants, and biobank every 2 years (for future research)

Exclusions include military, kids <12, on reserve or institutionalized

Question 4

NHANES

Answer 4

National Health and Nutrition Examination Survey

Assess health and nutrition status of adults and children

questionnaire to determine prevalence of major disease and risk factors

physical measurements 2x24hr recall + biomarkers, every year

Question 5

Syntax

Answer 5

Syntax = coding language used to perform data analysis operations

Question 6

Variable

Answer 6

Variable = factor or attribute which can be assigned 2 or more values

Question 7

Discrete vs Continuous variables

Answer 7

Discrete variables = have no intermediate value

Continuous variables = intermediate values between adjacent scale values can exist

Question 8

Numeric vs string variables

Answer 8

Numeric = number based

String = character based (can include numbers)

Question 9

Variable types:

Answer 9

Continuous, categorical, ordinal

Question 10

Continuous variables

Answer 10

AKA scale

Cannot be string variables, must be numeric

Ratio (true zero with even separation scale) or interval (arbitrary zero)

Question 11

Categorical

Answer 11

AKA nominal

Can be string or numeric

Question 12

Ordinal

Answer 12

Categorial with implicit order

Can be string or numeric

Question 13

Central limit theorem

Answer 13

central limit theorem = random average variables become normally distributed when observations are sufficiently large

Question 14

Gaussian curve

Answer 14

Standard normal curve
Mean = median = mode = 0

Std dev =1

Question 15

Parametric vs non parametric tests

Answer 15

Parametric tests - assume normal distribution of data and easier to interpret

Nonparametric tests - no normal distribution assumption

Question 16

Distribution definition:

Distribution types:

Answer 16

Definition: function showing all possible data values and how often they occur

Positively skewed - tail extends to the right (right skewed)

Negatively skewed - tail extends to the left (left skewed)

Bimodal - 2 peaks, can maybe be split into 2 sets or there is an underlying factor

Uniform - same value for all variables

Question 17

Descriptive statistics

Answer 17

Descriptive statistics = a set of methods used to summarize and describe the main features of a dataset, such as its central tendency, variability, and distribution.

can be used to calculate mode (not mean or median) - calculate proportions

Question 18

Central tendencies

Answer 18

Mode = most frequent value (peak)

Mean = average value of all data values
easily skewed by outliers and asymmetrical data

Median = literal measure of central tendency, central datum
best for asymmetrical data
- central number or 2 numbers averaged

Question 19

Population vs sample mean

Answer 19

Population = µ = xi/N

Sample = x̄ = xi/n

Question 20

Variability and variance

Answer 20

Variability = dispersion, spread or scatter

Variance = average of squared differences from the mean
Population: σ^2 = Σ(x -µ)^2 /N
Sample: s^2 = Σ(x - x̄)^2 /(n-1)

Question 21

Standard deviation

Answer 21

Standard deviation = square root of variance (degree individual values vary from mean)
√σ^2 or √s^2

Question 22

Bar chart vs histogram

Answer 22

Bar chart: summary statistics for continuous variables categorized
Can add clusters on x axis
Compares means/medians across groups

Histogram: distribution for continuous data
Frequency on y-axis
Good for evaluating the distribution shape of data
Can stack variables on x-axis

Question 23

Boxplot vs scatterplot

Answer 23

Boxplot: min/max, quartiles and interquartile range (IQR)
Can also display outliers
Box is IQR, quartiles are outer edges of IQR, median is line through IQR, whiskers are min/max

Scatterplot: displays values for 2 variables (independent x-axis and dependent y-axis)
Helpful for correlations and linear relationships
Can have more than one independent continuous variable using colors
Line drawn through to show correlation or regression

Question 24

Standard error of the mean

Answer 24

Standard error of the mean = estimate of how far sample mean (x̄) is from population (µ) mean
σ = s/(√n) standard deviation of sample/root n

Measures accuracy of sample reflecting population - deviation of sample mean from population mean
Variable data requires large sample to be accurate → large SEM = inaccurate estimate