Skip to main content

Section 6.2 Errors in Sampling

Definition 6.2.1.

  • Sampling error: refers to differences between a sample and the population that exist only because of the observations selected for the sample (i.e. errors due to chance).
    In other words, the difference between the statistic and the parameter are due to the fact that the sample is not a perfect representation of the population.
  • Nonsampling error: refers to differences between a sample and the population due to mistakes made in data acquisition or improper sample selection.

Exercise 6.2.2.

The Excel file below includes data about the total points scored by the women’s basketball team the Seattle Storm in each game during the 2023 regular season. external/sheets/SeattleStormSample2023.xlsx

(a)

Find the average number of points per game for the Seattle Storm during the 2023 regular season.
Answer.
78.825

(b)

Calculate the sampling error using the first 5 games in the Excel file as your sample.
Answer.
\begin{equation*} \overline{x}=76.8 \end{equation*}
\begin{equation*} \overline{x}-\mu=-2.025 \end{equation*}

(c)

Calculate the sampling error using the first 10 games in the Excel file as your sample.
Answer.
\begin{equation*} \overline{x}=77 \end{equation*}
\begin{equation*} \overline{x}-\mu= -1.825 \end{equation*}

(d)

Calculate the sampling error using the first 20 games in the Excel file as your sample.
Answer.
\begin{equation*} \overline{x}=79 \end{equation*}
\begin{equation*} \overline{x}-\mu=0.175 \end{equation*}

(e)

Calculate the sampling error using the first 30 games in the Excel file as your sample.
Answer.
\begin{equation*} \overline{x}\approx 78.2667 \end{equation*}
\begin{equation*} \overline{x}-\mu\approx -0.5583 \end{equation*}

(f)

Calculate the sampling error using the first 35 games in the Excel file as your sample.
Answer.
\begin{equation*} \overline{x}\approx 78.7429 \end{equation*}
\begin{equation*} \overline{x}-\mu\approx -0.0821 \end{equation*}

(g)

What happens to the sampling error as the sample size gets bigger?
Answer.
The sampling error decreases as the sample size increases.

(h)

Using a sample size of 4, what is the largest sampling error that can be observed from this population?
Answer.
largest \(\overline{x}=\frac{109+97+97+97}{4}=100\)
\begin{equation*} \overline{x}-\mu=21.175 \end{equation*}

Exercise 6.2.3.

(Donnelly 7.6)
The data table found below contains daily revenues for the past 350 business days earned from cell phone accessories (Bluetooth headsets, memory cards, and so on) sold at a local Verizon retail store in Delaware. external/sheets/CellPhoneAccessoryRevenue.xlsx

(a)

What is the population mean, \(\mu\text{?}\)
Answer.
\(\mu=1839\)

(b)

Use Excel to draw a systematic sample consisting of 14 days, and then calculate the sampling error for the sample.
Answer.
\(\overline{x}=1796.04\text{,}\) error\(=-42.69\)

(c)

Use Excel to draw a systematic sample consisting of $35$ days, and then calculate the sampling error for the sample.
Answer.
\(\overline{x}=1988.9\text{,}\) error\(=150.17\)

(d)

Use Excel to draw a systematic sample consisting of $50$ days, and then calculate the sampling error for the sample.
Answer.
\(\overline{x}=1900.43\text{,}\) error\(=61.7\)

(e)

Compare the sampling error for parts a,b, and c, and explain the reason for the differences.
Answer.
As the sample size increases, the sampling error tends to decrease.

(f)

What problems might be encountered with the sample obtained in part c?
Answer.
Selecting every 7th day is the same day of every week. (This potentially biases the data.)