Abstract

This lecture explores the concepts of standardization and the standard error of the mean (SEM) in psychological statistics. Standardization enables comparability across variables measured on different scales, while SEM quantifies the uncertainty of sample means in relation to the population mean. We derive these concepts mathematically and implement them using base R to reinforce statistical reasoning in psychological research.

Introduction

Statistical standardization plays a crucial role in analyzing data across different scales, enabling meaningful comparisons between variables. This lecture extends the discussion to the concept of standard error of the mean (SEM), a measure that quantifies sampling variability. These concepts form the foundation for inferential statistics, such as hypothesis testing and t-tests.

Standardization and Z-Scores

Standardization transforms raw data into a common scale with a mean of zero and a standard deviation of one. The formula for standardizing a score \(X\) into a \(Z\)-score is:

\[Z = \frac{X - \mu}{\sigma}\]

In R, the \(Z\)-score can be computed manually as follows:

data <- c(45, 50, 55, 60, 65)
mean_val <- mean(data)
sd_val <- sd(data)
z_scores <- (data - mean_val) / sd_val
print(z_scores)

Central Limit Theorem and Standard Error of the Mean

The Central Limit Theorem (CLT) states that the sampling distribution of the sample mean approaches a normal distribution as the sample size increases. The SEM quantifies the variability of sample means around the population mean and is given by:

\[SEM = \frac{\sigma}{\sqrt{n}}\]

The SEM can be computed in R as follows:

n <- length(data)
sem <- sd_val / sqrt(n)
print(sem)

Application: Comparing Sample Means

To illustrate the significance of SEM, we analyze changes in screen time usage among psychology students across different years.

stats_2019 <- subset(stats_data, Semester == 201910)$ScreenTime
stats_2025 <- subset(stats_data, Semester == 202510)$ScreenTime

mean_2019 <- mean(stats_2019)
mean_2025 <- mean(stats_2025)
sem_2019 <- sd(stats_2019) / sqrt(length(stats_2019))
sem_2025 <- sd(stats_2025) / sqrt(length(stats_2025))

print(c(mean_2019, sem_2019))
print(c(mean_2025, sem_2025))

Visualization: Bar Plot with Error Bars

A bar plot with error bars visually represents the mean and the associated SEM.

bars <- barplot(c(mean_2019, mean_2025), names.arg=c("2019", "2025"),
                ylim=c(0, max(mean_2025 + sem_2025) * 1.2),
                ylab="Screen Time (hours per week)", col=c("blue", "red"))
arrows(bars, mean_2019 - sem_2019, bars, mean_2019 + sem_2019, angle=90, code=3, length=0.1)
arrows(bars[2], mean_2025 - sem_2025, bars[2], mean_2025 + sem_2025, angle=90, code=3, length=0.1)

Discussion and Conclusion

This lecture emphasized the importance of standardization in comparing variables and the role of SEM in understanding the uncertainty of sample estimates. By applying these concepts using base R, students can gain a deeper understanding of how variability influences inferential statistics.