Intro to Business Statistics: Time Series Analysis - Seasonal Indices, Multiplicative Model Deseasonalizing Data

What This Is

Seasonal indices are used to identify and remove seasonal patterns from time series data. This is crucial in business decisions, such as forecasting sales, inventory management, and quality control. For example, a retail chain wants to know if average daily sales exceed $10,000 during the holiday season. By applying seasonal indices, the chain can identify the seasonal pattern and make informed decisions about inventory and staffing.

Key Formulas & Symbols

• Seasonal Index (SI) = (Xt / T) / (X / N) where Xt = seasonal total, T = time period, X = total, N = number of time periods.
• Multiplicative Model: Xt = (ST) * (T) * (I) where Xt = seasonal total, ST = seasonal index, T = trend, I = irregular.
• Deseasonalized Data: Xt = Xt / SI where Xt = deseasonalized total, SI = seasonal index.
• Trend (T) = (Xt / N) / (X / N) where Xt = seasonal total, N = number of time periods, X = total.
• Irregular (I) = Xt / (ST * T) where Xt = seasonal total, ST = seasonal index, T = trend.
• Degrees of Freedom (df) = n - 2 where n = number of observations.
• Critical Value (CV) = t(df, ?/2) where df = degrees of freedom,-= significance level, t = t-distribution.

Step-by-Step Procedure

1. State hypotheses: H?: no seasonal pattern, H?: seasonal pattern exists.
2. Choose test: Use the t-test for small samples or the F-test for large samples.
3. Compute test statistic: Calculate the t-statistic or F-statistic using the formula t = (X? - ?) / (s / ?n) or F = (MSB / MSW) where X? = sample mean,-= population mean, s = sample standard deviation, n = sample size, MSB = between-group mean square, MSW = within-group mean square.
4. Find p-value or critical value: Use a t-distribution table or calculator to find the p-value or critical value.
5. Compare to ?: Compare the p-value or critical value to the significance level (? = 0.05).
6. Conclude: If p-value <-or critical value > t(df, ?/2), reject H? and conclude that a seasonal pattern exists.

Common Mistakes

• Mistake: Using the Z-test when the sample size is small.
• Correction: Use the t-test instead, as it is more robust for small samples.
• Mistake: Misinterpreting the p-value as the probability that H? is true.
• Correction: The p-value is the probability of observing the data (or more extreme) if H? is true.
• Mistake: Failing to check for normality of residuals.
• Correction: Use a normality test, such as the Shapiro-Wilk test, to ensure that the residuals are normally distributed.

Quick Practice Problems

1. A company wants to know if there is a seasonal pattern in its sales data. The sample mean is $10,000, the sample standard deviation is $5,000, and the sample size is 12. What is the t-statistic? Answer: t = (10,000 - 0) / (5,000 / ?12) = 2.65. The calculation involves calculating the t-statistic using the given values.
2. A retail chain wants to know if there is a significant difference in sales between the holiday season and the rest of the year. The F-statistic is 4.2, and the degrees of freedom are 2 and 18. What is the p-value? Answer: p-value = 0.03. The calculation involves using an F-distribution table or calculator to find the p-value.
3. A company wants to deseasonalize its sales data. The seasonal index is 1.2, and the total sales are $100,000. What is the deseasonalized data? Answer: $83,333. The calculation involves dividing the total sales by the seasonal index.

Last-Minute Cram Sheet

• p-value is NOT the probability that H? is true – it’s the probability of observing the data (or more extreme) if H? is true.
• Use the t-test for small samples and the F-test for large samples.
• Degrees of freedom (df) = n - 2.
• Critical value (CV) = t(df, ?/2).
• Seasonal index (SI) = (Xt / T) / (X / N).
• Multiplicative model: Xt = (ST) * (T) * (I).
• Deseasonalized data: Xt = Xt / SI.
• Trend (T) = (Xt / N) / (X / N).
• Irregular (I) = Xt / (ST * T).
• F-statistic = (MSB / MSW).
• t-statistic = (X? - ?) / (s / ?n).