Research methods chap 6

Bias

A systematic tendency in data collection or analysis that leads to inaccurate, skewed, or distorted results. Can occur at the source, in data collection methods, or in analysis.

selection bias

Occurs when the sample is not representative of the population being studied.

measurement bias

Systematic error caused by faulty measurement tools or inconsistent procedures.

observer bias

When a researcher’s expectations influence how data is collected or interpreted.

recall bias

Errors that occur when participants inaccurately remember past events.

outlier

A data point that is very different from the rest, which can distort parameter estimates like the mean.

parametric statistics

Statistical methods that assume data follows a specific distribution and rely on parameters such as the mean, variance, and correlation. Examples: t-test, ANOVA, Pearson correlation, regression.

linearity

Predictors have a linear relationship with the outcome.

additivity

Effects of predictors combine additively (no interactions).

normality

Data or residuals are normally distributed

Homoscedasticity

Equal variances across groups or levels of predictors.

Independence

Observations are independent of one another.

Median & IQR (Interquartile Range)

Statistics that are more robust than the mean in non-normal or skewed distributions.

trimming data

Removing extreme scores according to clear rules to reduce bias.

Windsorizing

Replacing extreme outliers with the next highest/lowest non-outlier score.

bootstrapping

A resampling method that estimates statistics (like means) and confidence intervals by repeatedly sampling from the data.

data transformation

Applying mathematical functions (log, square root, reciprocal) to correct skew, stabilize variance, or reduce bias.

log transformation

Reduces positive skew by compressing large values.

square root transformation

Reduces positive skew and stabilizes variance.

histogram

Graphical display used to check normality.

P-P Plot (Probability-Probability Plot)

Plots cumulative probabilities of observed vs. expected normal data to test normality.

skewness

Measure of symmetry in a distribution. Negative = tail left, Positive = tail right.

kurtosis

Measure of "peakedness" in a distribution. Low = flat (platykurtic), High = sharp peak with heavy tails (leptokurtic).

Kolmogorov-Smirnov Test / Shapiro-Wilk Test

Statistical tests that compare sample data to a normal distribution. Significant p-value means data deviate from normal.

Z-score

A standardized score indicating how many standard deviations a value is from the mean. A score > |3| often indicates an outlier.

levene’s test

A statistical test for equality of variances across groups (used in ANOVA).