The standard deviation of the sampling distribution of the sample mean is given by σx̄ = σ / √n. This quantity is commonly called the standard error of the mean. It shows how much the sample mean is expected to vary from sample to sample when repeated samples of size n are taken from the same population. In Six Sigma Measure Phase analysis, this relationship is important because it explains why subgroup means are more stable than individual observations. As sample size increases, the denominator √n becomes larger, so the variation of the sample mean becomes smaller. That means larger samples produce more precise estimates of the population mean. This concept supports confidence intervals, hypothesis tests, x-bar chart limits, and sampling distribution logic used throughout Black Belt work. The incorrect formulas either increase variation with sample size or reverse the relationship entirely. Therefore, the correct answer is A, because the standard deviation of the sample mean decreases as the square root of the sample size increases.

===========

The correct answer is A. Ensuring that product tolerances are reasonable . In the CSSBB DFSS material, manufacturability is one of the key DFX characteristics that must be considered during product development. Supporting DFSS content also links design to tolerance / robust design , showing that producibility depends on making design requirements realistic for manufacturing capability.

Reasonable tolerances are essential to DFM because overly tight tolerances can increase cost, reduce yield, complicate production, and make the design difficult to produce consistently. A design may function well in theory yet still be poor from a manufacturability standpoint if it demands unnecessary precision or impractical processing conditions. The other options are less direct. KPOVs relate to key output variables, test measurement uncertainty is a metrology issue, and takt time is a Lean flow concept rather than a DFM design action. Since manufacturability is about making the design practical and efficient to produce, the best verified answer is A. Ensuring that product tolerances are reasonable .

The correct answer is A. attribute data . In the CSSBB source, attribute charts are used for characteristics such as pass/fail, good/bad, number of defects, and number of defectives, and the text explicitly states that attributes are discrete, counted data . This fits the question exactly because error counts in the final product are count-based and therefore discrete rather than continuous. The CSSBB material also identifies attribute data as one of the primary types of data used in Six Sigma, alongside variables data and locational data.

A data collection plan is a method for organizing how data will be gathered, not a type of data. Stratified data refers to data separated into meaningful subgroups, such as by shift, machine, or supplier, but that does not define whether the data are discrete or continuous. Continuous data, by contrast, involve measurements on a scale such as weight, time, or length. Since the question specifically describes discrete counts of errors, the CSSBB-consistent term is attribute data . This is why the best and verified answer is A .

The correct answer is B. Testing products returned by customers . The CSSBB source defines external failure costs as failure costs that occur after shipment of the product, or during or after furnishing a service, to the customer . The detailed external-failure checklist includes items such as returns/rework , retesting , restocking costs , recalls/repair , warranty expenses , and customer dissatisfaction , all of which clearly happen after the product has reached the customer.

Testing products that have been returned by customers fits this category exactly, because the failure has already escaped the organization and is now being handled after delivery. In contrast, measurement system analysis is a prevention or evaluation activity, destructive testing of end-product samples is an appraisal activity, and supplier-source inspection is also appraisal-related. Therefore, the only option that clearly belongs in the external failure category is testing products returned by customers . That makes B the verified CSSBB answer.

The correct answer is C. 81 . In DOE, the total number of runs starts with the number of treatment combinations in the design, then is multiplied by the number of replicates. For a 3-level design with 4 factors , a full factorial design would require 3⁴ = 81 treatment combinations. Since this question specifies a 1/3 fractional factorial , the base design is reduced to 81 ÷ 3 = 27 treatment combinations. With 3 replicates , the total required number of runs becomes 27 × 3 = 81 .

This is consistent with CSSBB DOE principles, where the number of runs depends on the number of factors, levels, fraction, and replicates. The CSSBB materials also emphasize that replicates are used to reduce experimental error and increase the power of the experiment, and that the design choice must balance information needs with budget and practicality. The source further discusses screening and multi-level designs, noting that 3-level designs are used when non-linear effects are expected. Therefore, after applying both the fractional reduction and the 3 replicates, the required total is 81 runs .