How to Calculate the Sample Mean in SPSS

What is the sample mean

The sample mean, usually displayed as Mean, is the sum of the observed values divided by the number of valid observations. It is a descriptive statistic that gives you a quick view of the typical level of a variable, such as average satisfaction, average income, or agreement with a Likert statement.

For a sample with values (x_1, x_2, ..., x_n), the formula is:

x̄ = (x₁ + x₂ + ... + xₙ) / n = Σxᵢ / n

Here, x̄ is the sample mean, Σxᵢ is the sum of the observed values, and n is the number of valid observations. If some values are missing, n must be the number of rows actually used in the calculation. It is not automatically the original number of respondents.

For example, five scores for one variable are 3, 4, 4, 5, and 2. The total is 18 and the number of observations is 5, so the mean is 18 / 5 = 3.6. If the data are entered in SPSS, you can let the software calculate the result instead of adding the values manually, but you still need to understand the formula so you can check the output.

If you are reviewing several descriptive statistics, you can also see the Statistics topic to distinguish Mean, Median, Mode, Standard Deviation, and range.

What the sample mean means in quantitative research

In a quantitative thesis, Mean is usually reported as part of the descriptive statistics. You can use it to describe the sample, comment on response patterns for each item, or compare ratings across groups of variables. If a convenience scale has a higher Mean than a cost scale, the descriptive results show that respondents rated convenience more positively in the sample being studied.

Mean does not prove a research hypothesis by itself. A high mean does not mean that an independent variable has a statistically significant effect on a dependent variable. To test relationships or effects, you need correlation, regression, ANOVA, a t-test, or a PLS-SEM model, depending on your research design.

Mean can also help you notice unusual patterns in the data. If every item has a Mean close to the highest scale point, check whether respondents selected the same level for most statements. This may indicate patterned responding, but Mean alone is not enough to conclude that the data are incorrect.

For a Likert scale, Mean is usually reported together with Standard Deviation, the number of valid observations, and sometimes the minimum and maximum values. A Mean of 3.5 may come from answers concentrated around 3 and 4. It may also come from one group choosing 1 and another group choosing 5. You need to examine the Standard Deviation to distinguish these situations.

When checking the data, you should also examine outliers in the data. A few extreme values can pull the Mean upward or downward, especially for variables measured across a wide range, such as income, revenue, or the number of service uses.

What sample mean is acceptable

There is no single acceptable Mean threshold for every topic. The appropriate value depends on the scale, the coding scheme, and the research question. For a 5-point Likert scale, a Mean of 3.8 is often interpreted as a relatively positive level of agreement, but you need to state the classification rule clearly and should not present it as evidence from a hypothesis test.

The first check is whether Mean falls within the valid range of the variable. If a Likert variable is coded from 1 to 5, a valid Mean must fall between 1 and 5. A Mean of 7.2 for this variable usually suggests a data-entry error, a coding error, or that you are reading the wrong column.

The table below gives practical checking points. These are criteria for checking and interpreting data, not mandatory statistical thresholds that every study must meet.

If you divide a 5-point Likert scale into interpretation ranges, one common approach is to calculate the range as 5 - 1 = 4 and divide it by 5 levels, giving an interval width of 0.8. You can describe 1.00 to 1.80 as very low, above 1.80 to 2.60 as low, above 2.60 to 3.40 as moderate, above 3.40 to 4.20 as high, and above 4.20 to 5.00 as very high. This is a descriptive convention selected by the researcher, not a statistical law.

If the scale has 7 levels, do not reuse the intervals from a 5-point scale. Recalculate the interval width using the value range and the number of levels you used. Before writing the results, state the lower and upper points of the scale and whether Mean was calculated for a single variable or for the average score of the whole scale.

How to read the sample mean in the output

In SPSS, the quickest way to calculate Mean is to open Analyze > Descriptive Statistics > Descriptives. Move the variables you want to examine into the Variables box, select Options, tick Mean, Std. deviation, Minimum, and Maximum, then select Continue and OK. SPSS will create the Descriptive Statistics table.

Illustrative output

The table below is illustrative output, not the result of a real study. The column names are kept as they appear in SPSS so you can compare them with your own file.

For the HL1 row, N equals 198, which means that SPSS used 198 valid values. Minimum equals 1 and Maximum equals 5, showing that the observed data fall within the range of a 5-point Likert scale. Mean equals 3.72, which is the average level for HL1. Std. Deviation equals 0.846 and describes how dispersed the answers are around the Mean.

HL3 has N equal to 196, while HL1 and HL2 have 198. HL3 may have two missing values. The Valid N (listwise) value of 196 means that, under listwise analysis, only 196 cases have complete data across all variables included in the table. Distinguish the N in each row from Valid N (listwise), because they answer two different questions.

When you need to calculate Mean for groups of variables, you can use Analyze > Compare Means > Means. Move the dependent variable into Dependent List and the grouping variable into Independent List if you want to compare groups. If you only need descriptive statistics for each variable, Descriptives is usually easier to check.

You can also use Analyze > Descriptive Statistics > Frequencies, select Statistics, and tick Mean, Median, Mode, and Std. deviation. Frequencies is useful when you want to see the frequency table for each response level as well. For a continuous variable with many different values, the frequency table can become long and difficult to read.

To calculate Mean for one construct, such as HL with HL1, HL2, and HL3, use Transform > Compute Variable. Enter a name for the new variable, such as HL_MEAN, and enter the formula MEAN(HL1, HL2, HL3). The MEAN function works with the valid values in each row, so you must check your study's missing-value rule before using it. If respondents must have valid data for all three variables, check the number of valid values before creating the composite score.

The Mean of an individual item and the Mean of a composite scale are not the same quantity. The Mean of HL1 describes one statement, while HL_MEAN describes the average score of the items belonging to the HL construct. When writing Chapter 4, state clearly which one you are reporting.

What to do when the sample mean looks wrong

Start by checking the value range. Go to Analyze > Descriptive Statistics > Frequencies and review Minimum, Maximum, and the frequency table. If a Likert variable coded from 1 to 5 contains 0, 6, or 99, check the codebook and the missing-value rule. Code 99 is sometimes used to represent no response, but SPSS still treats it as a number unless you have declared it as missing.

Next, check the direction of the variable. A reverse-worded variable on a 1 to 5 scale needs to be recoded using the formula 6 - old value. You can use Transform > Recode into Different Variables to preserve the original variable, then check the Mean of the new variable. Do not reverse-code a variable only because its Mean is low. Base the decision on the statement wording and the questionnaire design.

Then review missing values and the number of valid observations. If one variable has a noticeably lower N, find the blank rows in Data View and check how SPSS is handling missing values. Do not fill every blank cell with the Mean just to make N look better. That can reduce variation and affect later analyses.

Finally, check outliers and the distribution. Mean is sensitive to extreme values, so you can inspect a Histogram, Boxplot, or Explore through Analyze > Descriptive Statistics > Explore. Removing a case simply because it changes the Mean is difficult to defend. Remove a case only when you have a clear basis, such as a data-entry error, an invalid response under a pre-established criterion, or a value outside the possible range.

If Mean is low but all values are valid, do not alter the data to reach a preferred level. A low result may be a real finding in the sample. Report the result accurately, discuss possible reasons, and examine the appropriate tests. Deleting an item or respondent only to increase Mean creates a serious risk when your supervisor asks about the data-processing history.

Distinguishing Mean, Median, and Mode

Mean is the sum of the values divided by the number of observations. Median is the middle value after the data are sorted. Mode is the most frequently occurring value. These three statistics can be similar in a balanced distribution, but they may differ substantially when the data are skewed or contain outliers.

For example, income data are 6, 7, 7, 8, and 50. Mean equals 15.6, while Median equals 7. The value 50 pulls the Mean upward, whereas Median is closer to the typical level of the four lower values. For income, spending, or time variables, examine both Mean and Median instead of reading Mean alone.

For a Likert scale, Mean is often used to describe the rating level because quantitative studies commonly calculate average scores for items or scales. Median is still useful when you want to describe the middle response, while Mode shows the most common response level.

Do not confuse Mean with Standard Deviation. Mean answers the question of how high or low the average level is, while Standard Deviation answers how dispersed the data are around Mean. Also do not confuse Mean with a factor score, regression score, or path coefficient in SmartPLS. Those measures are created for different analytical purposes.

If your study includes a moderating variable or a mediator, the Mean of each variable is still only a descriptive statistic. You can read more about moderating or mediator variables to distinguish their roles in a model from their average values.

Common errors

Dividing by the total number of questionnaires instead of valid observations

You distributed 220 questionnaires, but one variable has only 214 valid responses. The Mean for that variable must be based on the valid values used by SPSS. Read the N column before interpreting the result.

Using Mean to conclude a hypothesis

A high Mean describes the rating level in the sample. It does not show a statistically significant effect, a causal relationship, or whether a hypothesis is accepted. A hypothesis conclusion must be based on the appropriate test.

Ignoring reverse-coded variables

If a statement is worded in the opposite direction, determine the coding method before calculating the scale score. If you skip this step, the scale Mean may be low and Cronbach's Alpha or later analyses may be affected.

Combining Means from scales with different numbers of levels

Do not directly average a variable coded from 1 to 5 with a variable coded from 1 to 7 without a clear methodological reason. Before aggregation, check the content, measurement direction, and value range of each variable.

Reporting Mean without N

A Mean of 3.72 does not tell the reader whether it was calculated from 30 or 300 respondents. In the results table, report the valid N together with Mean and Std. Deviation so the reader can assess data completeness.

How to write this in your thesis

You can adapt the following paragraph to your actual output: “The descriptive statistics show that variable [VARIABLE CODE] has a Mean of [MEAN] and a Standard Deviation of [SD], based on [N] valid observations. The minimum and maximum values are [MIN] and [MAX], respectively, showing that the responses fall within the coding range of the [NUMBER OF LEVELS]-point scale.”

If you are reporting the whole scale, you can write: “The Mean score of the [CONSTRUCT NAME] scale is [SCALE MEAN], indicating that the surveyed sample tends to rate it at the [INTERPRETATION LEVEL] level according to the classification rule [STATE THE RULE].” Replace every placeholder with your actual results and state clearly that this is a descriptive interpretation, not a conclusion from a hypothesis test.

Frequently asked questions

How do you calculate the sample mean?

Add all valid observed values and divide the total by the number of valid observations. The formula is x̄ = Σxᵢ / n. For data in SPSS, use Analyze > Descriptive Statistics > Descriptives so the software calculates Mean and N.

What sample mean is acceptable?

There is no single acceptable Mean. Check the measurement scale, the value range, and the purpose of the study. For a 5-point Likert scale, use the interpretation rule stated in your methods chapter and check whether Mean falls between 1 and 5.

What should you do when Mean and Median differ?

This is often a sign of a skewed distribution or extreme values. It does not automatically mean that the data are wrong. Review the Histogram, Boxplot, Minimum, Maximum, and possible outlier cases before deciding how to report the result.

Why is the Mean in SPSS different from a manually calculated result?

Common causes include dividing by the total number of rows instead of the valid N, SPSS handling missing values differently, or values such as 99 not being declared as missing. Check the N column, missing-value codes, and the calculation formula in Transform > Compute Variable.

Should you use Mean for Likert data?

Many quantitative theses use Mean to describe individual items and scale scores. You should still report Standard Deviation, state the coding method, and consider Median if the data are strongly skewed or contain extreme values.

Open your SPSS file, run Analyze > Descriptive Statistics > Descriptives, compare N, Mean, Minimum, Maximum, and Std. Deviation, then save the output table together with the processing syntax. If you need to run the statistics on your own .sav or .csv file and check each step, you can use M4 data analysis from DoThesis.