8.4
Choosing Between z and t Distribution
The z and the Student t distribution estimate the population mean using the sample mean and standard deviation. However, to decide which distribution…
The z and t distributions can estimate the mean of a population using a sample statistic. But how does one choose the appropriate distribution for a given dataset?
The z distribution is preferred for populations with a known standard deviation that is normally distributed or populations that have a sample size greater than 30.
However, the Student t distribution is preferred if the population standard deviation is unknown for a normally distributed population or if the population has a sample size greater than 30.
Symmetrically distributed datasets with very large sample sizes show less variability. For such datasets, the population mean estimated by both the z and t distributions are similar.
The z and t distributions are limited to random samples drawn from normally distributed populations. Thus, they cannot estimate the population mean for samples drawn from voluntary sample responses, convenience sampling, or skewed or unknown population distributions.
Therefore, nonparametric statistics or computer bootstrapping methods are used for populations that are not normally distributed and those populations that have a sample size less than or equal to 30.
View the full transcript and gain access to JoVE Core videos
Q1: When should you use the z distribution instead of the t distribution?
Use the z distribution when the population standard deviation is known and the population is normally distributed, or when the sample size exceeds 30. The z distribution is preferred for these conditions because it provides reliable estimates of the population mean. Both distributions estimate population parameter values, but z is optimal when population parameters are known or sample sizes are large.
Q2: What conditions make the Student t distribution the better choice?
The Student t distribution is preferred when the population standard deviation is unknown and the population is normally distributed, or when the sample size exceeds 30. This distribution accounts for uncertainty in estimating the population standard deviation from sample data. It provides more conservative estimates than the z distribution when population parameters are unavailable.
Q3: Why do z and t distributions produce similar results for large samples?
Symmetrically distributed datasets with very large sample sizes show less variability, causing both distributions to converge. As sample size increases, the t distribution approaches the z distribution because the sample standard deviation becomes a more reliable estimate of the population standard deviation. This similarity reflects reduced uncertainty in parameter estimation with larger samples.
Q4: What sampling methods prevent using z or t distributions?
Voluntary response sampling, convenience sampling, and samples from skewed or unknown population distributions cannot be analyzed using z or t distributions. These sampling methods violate the assumption that data come from random samples of normally distributed populations. For such data, nonparametric statistics or computer bootstrapping methods provide more appropriate alternatives.
Q5: What should you do when sample size is less than 30 and distribution is unknown?
When sample size is less than 30 and the population distribution is unknown or skewed, neither z nor t distributions can accurately estimate the population mean. Instead, use nonparametric statistical methods such as bootstrapping for categorical data or small samples. These methods do not assume normality and provide valid estimates without relying on distribution assumptions.
Q6: How does sample size affect the choice between z and t distributions?
Both z and t distributions can be used when sample size exceeds 30, regardless of whether the population standard deviation is known. For samples smaller than 30, the choice depends on whether the population standard deviation is known and the population is normally distributed. Sample size is a critical decision factor because larger samples reduce variability and improve the reliability of both distribution estimates.
Q7: What are the key requirements for using z and t distributions?
Both z and t distributions require random samples drawn from normally distributed populations to estimate the population mean accurately. They cannot be applied to data from voluntary responses, convenience samples, or non-normal distributions. Understanding these limitations ensures appropriate statistical method selection for your data and research question.