What Is the Mean? How to Read and Report It in a Quantitative Thesis

What Is the Mean

In a data table or SPSS output, the mean usually appears under the name Mean. It is a representative value calculated by adding the observations and dividing the total by the number of valid observations. For a variable with values (x_1, x_2, ..., x_n), the formula is:

[ Mean = \frac{x_1 + x_2 + ... + x_n}{n} ]

For example, five respondents rate their level of agreement as 2, 3, 3, 4, and 5. The mean is (2 + 3 + 3 + 4 + 5) / 5 = 3,4. The value 3,4 shows the arithmetic centre of these scores, but it does not mean that a particular respondent selected exactly 3,4.

When you use a 5-point Likert scale, each item is usually coded from 1 to 5. Level 1 might mean “Strongly disagree”, while level 5 means “Strongly agree”. The coding and interpretation must match your questionnaire. The mean is a descriptive statistic. It does not by itself prove that a hypothesis is correct or that a factor affects the dependent variable.

What the Mean Tells You in Quantitative Research

The mean gives you a quick view of the response pattern in your sample. If the Mean of an item is 4,20 on a 5-point scale, the responses generally lean toward agreement. If the Mean is 2,10, the overall pattern leans toward disagreement. You should also check the standard deviation, minimum, maximum, and number of valid observations to understand the spread behind that value.

In the results chapter, the mean is commonly used for three purposes. First, it describes observed variables or research constructs. Second, it provides a preliminary comparison of ratings across variables. Third, it helps you check whether the data fall within the expected coding range. For example, if a scale only runs from 1 to 5 but the Mean is 8,00, check the data entry, calculation formula, or variable that you are reading.

You can also calculate a mean for a group of items after deciding how to combine the scale. For example, the latent variable “Service quality” has five items, CLDV1 to CLDV5. You may create an average score from the five items for the next analysis, but you must state how missing data are handled. If a respondent answered only three of the five items, whether you allow a composite score must be specified in advance rather than changed after seeing the result.

You should distinguish a descriptive mean from a coefficient in a regression model or PLS-SEM. Mean answers the question, “At what level does the sample tend to rate this item?” A path coefficient, regression coefficient, and p-value answer questions about relationships or effects between variables. A high Mean therefore does not mean that the variable has a strong effect.

With extreme values, the mean can be pulled toward the larger or smaller observations. You can read more about outliers before deciding whether to retain or remove an observation. Removing a data row simply to make the Mean look “better” is difficult to defend unless you have a clear criterion.

What Mean Value Is Acceptable

There is no single threshold that works for every study to determine whether a Mean is “acceptable”. You must read the Mean against the measurement scale, coding scheme, research objective, and convention stated in your methods chapter. With a 5-point Likert scale, some theses divide the interpretation range by calculating the interval width:

[ \text{Interval width} = \frac{5 - 1}{5} = 0,8 ]

Using that calculation, you can define 1,00 to 1,80 as very low, 1,81 to 2,60 as low, 2,61 to 3,40 as moderate, 3,41 to 4,20 as high, and 4,21 to 5,00 as very high. This is a descriptive interpretation selected by the researcher. It is not a standard for removing an item or accepting a model. If you use it, state the convention clearly in your thesis.

The table below illustrates how to read Mean on a 5-point Likert scale. The source in the table concerns the Likert scale, while the interpretation ranges are calculated from the scale width and are not presented as testing thresholds from that source.

Do not write “a Mean of 3,0 or above is acceptable” until you explain what acceptable means. If you are assessing agreement, a Mean of 3,0 may indicate a neutral attitude. If you are describing perceived level, 3,0 may be called moderate. When you assess scale reliability or validity, use Cronbach's Alpha, Corrected Item-Total Correlation, factor loading, AVE, or another appropriate index instead of using the Mean as a substitute.

A 7-point Likert scale requires its own interval calculation. You cannot apply the ranges from a 5-point scale directly to a scale from 1 to 7. Similarly, if an item has been reverse-coded, calculate the Mean using the reversed coding. If you miss this step, an item with the opposite meaning can pull the average score of the whole construct downward.

How to Read the Mean in SPSS Output

In SPSS, you can run descriptive statistics through Analyze > Descriptive Statistics > Descriptives. Move the variables you want to inspect into the Variables box, click Options, select Mean, Std. deviation, Minimum, and Maximum, then click Continue and OK.

If you want to compare means across groups, such as satisfaction by gender, use Analyze > Compare Means > Means. For a more detailed descriptive table, Analyze > Descriptive Statistics > Explore lets you examine distributions and charts. Menu names may vary slightly across SPSS versions, but the main columns have the same meaning.

Descriptive Statistics Table

Illustrative output, not the result of an actual study.

The N column is the number of valid observations that SPSS used for each variable. In the illustrative output, CLDV3 has an N of 197 while the other variables have an N of 198. This often indicates one missing value for CLDV3, but open Data View or check the missing-value rule to confirm it.

The Minimum and Maximum columns show the smallest and largest values included in the calculation. The Mean column shows the average. Std. Deviation is the standard deviation, which reflects how widely the observations are dispersed around the Mean. A standard deviation of 0,781 suggests that responses to CLDV3 are relatively more concentrated than responses to CLDV2, which has a standard deviation of 0,915. You should not turn this comparison into a conclusion about scale quality.

In the illustrative output, the Mean of CLDV3 is 3,91. If you stated the interpretation convention in your methods section, this value may be described as high. The Mean of HAILONG is 3,72, which also falls in the high range under the convention above. This wording only describes the data. It does not say that service quality causes satisfaction.

How to write this in your thesis

You can adapt the following paragraph to your actual output: “Descriptive statistics show that the variable [{VARIABLE CODE}] has a mean of [{MEAN}], a standard deviation of [{SD}], and minimum and maximum values of [{MIN}] and [{MAX}], respectively. Based on the interpretation convention for the [{NUMBER OF LEVELS}]-level scale presented in the methods chapter, respondents’ assessment of this variable is considered [{INTERPRETATION LEVEL}].”

When writing, replace every item in square brackets with your actual values. Do not paste the illustrative table into Chapter 4. If several items measure the same construct, present a summary table with the item code, Mean, and standard deviation, then comment on the items with the highest and lowest Means. Keep the comment tied to the table and avoid adding causes that the descriptive data have not tested.

What to Do When the Mean Is Not Acceptable

First, identify what “not acceptable” refers to. A low Mean is not a technical error. It may show that respondents genuinely disagreed with the statement. You only need to intervene when the Mean falls outside the coding range, the number of valid observations is unusual, an item was reverse-coded incorrectly, or the calculation does not match the research design.

Check the coding and entered values

Open Variable View and check Type, Values, and Missing. Then use Analyze > Descriptive Statistics > Frequencies to view the frequency table. If the scale runs only from 1 to 5 but values such as 0, 6, or 99 appear, determine whether they represent missing codes, entry errors, or valid values according to the questionnaire.

Check reverse-coded items

A negatively worded statement must be reverse-coded before you calculate an overall score. For a scale from 1 to 5, the new code is often calculated with 6 - old code, but check this against the design of your questionnaire. Keep a record of the original and reverse-coded variables so that you can explain the step if your supervisor or the committee asks.

Check the composite-score calculation

If the Mean of a construct is calculated from several items, check the formula in Transform > Compute Variable. Confirm that the correct items were included and that they all point in the same direction. When missing data are present, the decision to retain or remove a row must be consistent with the method you described. Do not delete a respondent simply because the Mean is low.

Inspect the distribution and outliers

If the Mean and median differ substantially, the data may be skewed or contain extreme values. In that situation, inspect the histogram, boxplot, and descriptive statistics before choosing a treatment. If you are building a model with a moderator, distinguish a descriptive Mean from the creation of an interaction term in moderating analysis. Standardization or mean-centering serves a different purpose and does not automatically make the questionnaire Mean “acceptable”.

Distinguishing the Mean from the Median

The mean and median both describe central location, but they are calculated differently. The mean uses every value in the addition, so it is more easily affected by a few very large or very small observations. The median is the middle value after the data are sorted, or the average of the two middle values when the number of observations is even.

For example, assume income data are 6, 7, 7, 8, and 50. The Mean is 15,6, while the median is 7. The Mean is pulled upward by the value 50, while the median more closely reflects the typical level of the first four observations. With Likert data, the Mean is often reported in descriptive statistics according to the convention of the study, but you can still inspect the median and distribution to identify skewness.

You should also distinguish the mean from the mode, which is the value that occurs most often. If most respondents select level 4 but many responses are still at level 1, the Mean may be lower than 4. Each statistic answers a different question, so choose the Mean or median based on the data type and descriptive objective.

Finally, do not confuse the Mean of an item with a grade-point average, group average, or average score for a subsample. Read the variable name, value labels, and N in the output before interpreting the result. If you are reading about concepts such as what is a scale, identify whether you are looking at an individual item or a composite score for the entire scale.

Common Errors

Calling the Mean an effect. The Mean only describes the response pattern. To conclude that one variable affects another, you need an analysis that matches your model, such as regression or PLS-SEM, and then you must read the coefficient and significance level.

Using one Mean threshold for every scale. A scale from 1 to 5, a scale from 1 to 7, and a scale from 0 to 10 have different value ranges. The interpretation convention must match the scale you used.

Ignoring the number of valid observations. Two variables may have the same Mean, but one may be based on 198 observations while the other is based on only 150. They should not be discussed as if they came from the same data base.

Rounding too early. Keep the decimal places in the analysis file and round only when presenting the results. If every Mean is rounded to 4, readers will have difficulty checking the differences between variables.

Deleting data to raise the Mean. This is not a defensible cleaning method. You need a data-cleaning rule defined in advance, a record of the removed row, and a reason for its removal. If the issue concerns scale quality, check reliability and validity instead of altering the descriptive result.

Read more: outliers, moderating, moderator, what is a scale, mediator.

Frequently asked questions

What is the mean?

The mean is the sum of the observed values divided by the number of valid observations. In SPSS output, it usually appears in the Mean column. It shows the arithmetic centre of the data and does not represent a specific respondent’s answer.

What Mean value is acceptable?

There is no single threshold for deciding that every Mean is acceptable. Base your interpretation on the measurement scale and the convention stated in your methods section. With a 5-point Likert scale, you can divide 1,00 to 5,00 into five intervals, but this is a descriptive convention rather than a scale-quality test.

What is the difference between the mean and the median?

The Mean adds all values and divides the result by the number of observations, while the median is the middle position after sorting the values. The Mean is more sensitive to outliers. When the data are strongly skewed, inspect both statistics and the distribution chart before interpreting them.

Does a low Mean mean that I should remove the item?

A low Mean is not automatically a reason to remove an item. It may accurately reflect the attitude of the sample. Consider changing or removing data only when you find a coding error, incorrect reverse coding, a value outside the permitted range, or a problem covered by your stated data-cleaning and scale-testing criteria.

Where can I read the Mean in SPSS?

Run Analyze > Descriptive Statistics > Descriptives and select Mean, Std. deviation, Minimum, and Maximum. In the Descriptive Statistics table, read the N column first to identify the number of valid observations, then read the Mean and standard deviation. Check the variable name against the questionnaire before writing your interpretation.

Open your SPSS file, run Analyze > Descriptive Statistics > Descriptives, check each N, Mean, and Std. Deviation column, then write a paragraph using your actual values and the template above. If you need the full analysis run on your .sav or .csv file, use M4 data analysis by DoThesis.