Seasonally adjusted vs not seasonally adjusted data is a common confusion point in statistics, economics, and business homework. The short version is this: seasonally adjusted data remove predictable calendar patterns, while not seasonally adjusted data show the values as originally observed before that statistical adjustment.
This matters whenever students read CPI, PPI, jobs reports, unemployment rates, retail sales, or PCE inflation. A monthly change often uses seasonally adjusted data, while a 12-month comparison or contract adjustment may use not seasonally adjusted data. Choosing the wrong series can make the answer look inconsistent even when the data are correct.
Sources: BLS Handbook of Methods seasonal adjustment topic, BLS CPI guidance on seasonally adjusted and unadjusted data, BLS general FAQ on seasonal adjustment, BLS CPS concepts and definitions, and BEA Personal Income and Outlays additional information.
Quick answer: Seasonally adjusted data are changed statistically to remove recurring seasonal patterns such as holidays, weather, school calendars, production cycles, and regular sales periods. Not seasonally adjusted data keep the observed values. Use adjusted data for short-term month-to-month trends and unadjusted data when the actual observed value or year-over-year comparison is the focus.
Table of Contents
What Does Seasonally Adjusted Mean?
Seasonally adjusted data are data that have been statistically adjusted to reduce the effect of patterns that usually happen around the same time each year. Common seasonal influences include weather, holidays, school openings and closings, travel patterns, production cycles, model changeovers, and scheduled sales.
The purpose is not to hide real data. The purpose is to make short-term comparisons easier. If a series always rises every December because of holiday shopping, a seasonally adjusted version tries to remove that normal December pattern so readers can focus on what is unusual for that month.
What Does Not Seasonally Adjusted Mean?
Not seasonally adjusted data, often shortened to NSA or unadjusted data, show the values before seasonal adjustment. These numbers still follow the agency’s normal survey, collection, estimation, and publication process, but they do not remove recurring seasonal patterns.
Unadjusted data are useful when the actual observed level matters. For CPI, BLS notes that unadjusted data are important for consumers concerned with prices actually paid and for escalation uses such as some contracts and benefit adjustments.
Seasonally Adjusted vs Not Seasonally Adjusted
The easiest way to compare seasonally adjusted vs not seasonally adjusted data is to match each series to the question being asked.
| Question | Better Series | Reason |
|---|---|---|
| What changed from last month to this month? | Seasonally adjusted | Month-to-month comparisons can be distorted by normal calendar patterns. |
| What did consumers actually pay in a specific month? | Not seasonally adjusted | The unadjusted value keeps the observed price level before seasonal smoothing. |
| How much did the index change compared with the same month last year? | Often not seasonally adjusted | Same-month year-over-year comparisons already compare the same season. |
| What is the recent short-term trend? | Seasonally adjusted | Removing recurring patterns makes recent monthly movement easier to see. |
| Should a contract or payment be adjusted? | Usually not seasonally adjusted | Some adjusted series can be revised, which can create problems for fixed agreements. |
Neither series is automatically “better.” The right choice depends on the purpose of the analysis.
CPI Example: Monthly vs 12-Month Inflation
CPI is one of the clearest examples because BLS publishes both seasonally adjusted and not seasonally adjusted CPI series. If a headline says CPI rose from June to July, it often uses the seasonally adjusted monthly change. If it says inflation rose over the last 12 months, it often uses a not seasonally adjusted year-over-year comparison.
12-month percent change = ((current month NSA index – same month last year NSA index) / same month last year NSA index) x 100
This is why two CPI numbers can appear in the same article without contradicting each other. They may use different series because they answer different questions.
For the current CPI article, see Statskan’s CPI report explained guide. For the broader inflation comparison, use CPI vs PPI vs PCE explained.
Jobs Report Example: Employment and Unemployment
Seasonal adjustment also matters in jobs reports. Employment, unemployment, labor force participation, and payroll jobs can move in regular seasonal ways because of school calendars, holidays, weather, retail hiring, tourism, agriculture, and other repeating patterns.
BLS explains that seasonal adjustment makes it easier to observe underlying labor-market movement instead of normal seasonal fluctuation. That is why many monthly jobs-report headlines use seasonally adjusted values.
For a live labor-market example, read Statskan’s August jobs report statistics explained guide.
Why Seasonally Adjusted Data Can Be Revised
Seasonally adjusted data can be revised because seasonal factors are estimated from patterns in the time series. When new months or new years of data are added, the estimate of the seasonal pattern may change.
For CPI, BLS explains that seasonal adjustment factors are updated each February and that the new factors are used to revise the previous five years of seasonally adjusted data. This is one reason BLS advises against using seasonally adjusted CPI data in escalation agreements.
| Issue | What It Means | Student-Friendly Explanation |
|---|---|---|
| Seasonal factors | Estimated patterns used to adjust the original series. | The model estimates the normal seasonal effect and removes it. |
| Revision | Earlier adjusted values can change when seasonal factors are recalculated. | A revised value is not necessarily an error; it may reflect better information. |
| Current-year adjustment | Recent data can be more likely to change than older final data. | Be careful when quoting the newest adjusted values as final. |
Student Rules for Choosing SA or NSA
When a homework prompt gives both adjusted and unadjusted data, do not guess. Use the purpose of the question to choose the series.
- For month-to-month changes, start with seasonally adjusted data unless the prompt says otherwise.
- For same-month year-over-year changes, look for not seasonally adjusted data unless the source specifically reports adjusted 12-month changes.
- For contracts, benefit adjustments, or real-world payment escalation, do not assume seasonally adjusted data are appropriate.
- For jobs-report headlines, check whether the reported payroll, unemployment, or participation value is seasonally adjusted.
- For PCE and BEA releases, remember that many levels are shown at seasonally adjusted annual rates, so read the table note carefully.
If the assignment asks for a calculation, write the exact series name in your answer before the formula. That one line can prevent many wrong interpretations.
Common Mistakes
- Mixing SA and NSA values in one calculation: Do not subtract a seasonally adjusted value from a not seasonally adjusted value unless the assignment specifically tells you to.
- Calling unadjusted data “raw data”: Unadjusted published data may still be estimated, edited, weighted, or indexed. It simply has not been seasonally adjusted.
- Thinking adjusted means fake: Seasonal adjustment is a statistical method used to answer a specific question about underlying movement.
- Ignoring revisions: Seasonally adjusted series can be revised when agencies update seasonal factors.
- Using the wrong denominator: For rates and percent changes, always identify the old value, new value, and period being compared.
For calculator support, Statskan’s p-value calculator, normal distribution calculator, and statistics calculators hub can help with related statistics homework.
Need Help Choosing the Right Data Series?
Send your prompt, data table, source link, software output, and deadline. Statskan can help you identify the correct series and explain the result clearly.
Ask for Statistics Help Work With an Online Statistics TutorFrequently Asked Questions
Seasonally adjusted data are data changed statistically to remove recurring seasonal patterns so month-to-month movement is easier to interpret.
Not seasonally adjusted data keep the observed values before seasonal adjustment. They are also called unadjusted or NSA data.
Use seasonally adjusted CPI for short-term monthly comparisons. For same-month year-over-year CPI comparisons or some escalation uses, not seasonally adjusted CPI is often more appropriate.
Seasonally adjusted data can be revised because agencies recalculate seasonal factors when new data become available. The revision can improve the estimate of the seasonal pattern.
Leave a Reply