What Is STD? How to Read Standard Deviation in SPSS

You open the Descriptive Statistics table in SPSS and see the columns Mean, Std. Deviation, and N. The Std. Deviation column is STD, short for Standard Deviation. It shows how far the values in your data are spread from the mean.

STD does not tell you automatically whether your data are good or bad. Read it together with the scale, the mean, the valid value range, and the purpose of your analysis. An STD of 2 may be very large when a variable is measured from 1 to 5, but it may not be concerning when the variable ranges from 0 to 100.

If you are reviewing the statistics in your output, you can also see the Statistics section to distinguish descriptive statistics from tests and inferential models.

What Is STD

STD measures how widely observations are dispersed around the mean. When responses are concentrated close to the Mean, STD is small. When responses spread across several levels, STD is large. At the population level, the formula is commonly described as the square root of variance, while SPSS uses the sample standard deviation for a sample, with N - 1 in the denominator.

You can picture it this way: one group of students has scores of 7, 7, 8, 8, and 7. Its Mean is close to 7.4 and its STD is small. Another group has scores of 4, 5, 8, 9, and 9. It may have a similar Mean, but its STD is larger because the scores lie farther from the Mean.

In SPSS, STD usually appears under the name Std. Deviation. It is neither a hypothesis test nor a p-value. It is a descriptive statistic used to summarize data variation before you run Cronbach's Alpha, EFA, regression, or another model.

STD uses the same unit as the original variable. If the variable is measured with Likert points, STD is also expressed in Likert-point units. If the variable is income measured in millions of currency units, STD has the same unit. Looking at a number without checking the measurement range can therefore lead you to the wrong conclusion.

The Meaning of STD in Quantitative Research

STD helps you answer three practical questions. First, do respondents tend to choose similar options, or are their responses spread across the scale? Second, does the variable have enough variation for further analysis, or are most responses concentrated at one level? Third, might an observation be far from most of the data?

For example, the variable HL1 measures satisfaction on a 1 to 5 Likert scale, with a Mean of 4.20 and an STD of 0.45. The responses are relatively concentrated at a high level. If the Mean is 3.10 and the STD is 1.35, respondents are more dispersed in their views. A large STD in this situation does not by itself show that the questionnaire has a problem. It may reflect genuine differences between respondents.

In descriptive statistics, you usually read STD together with Minimum, Maximum, Mean, and N. If a variable has a Minimum of 1 and a Maximum of 5 but an STD of only 0.10, almost all observations may be concentrated around one value. The scale may then have limited ability to distinguish between respondents.

STD is also relevant when you identify unusual data, but do not use STD as the only criterion for deleting a row. If you suspect that one respondent answered very differently, check the response pattern, completion time, answer-selection rules, and related variables. You can also read the guide to outliers to separate unusual observations from the overall distribution of the variable.

For regression, the STD of the dependent and independent variables helps you understand the measurement scale before interpreting coefficients. STD does not determine whether a hypothesis is supported. That decision depends on the coefficient, p-value, confidence interval, and checks appropriate for the model.

When a variable belongs to a multi-item scale, you also need to understand how it is coded and what the scale measures. STD is calculated from the data you collected, while a scale describes how a construct is measured.

What Level of STD Is Acceptable

There is no single STD threshold for every research topic. The question of whether a standard deviation is acceptable must be answered based on the variable's measurement range. For a 1 to 5 Likert scale, STD cannot exceed the level allowed by the scale. For income, age, or revenue, STD may be much larger because the unit and data range are different.

The table below is a practical way to read STD, not a rule for automatically deleting a variable. Interpret statements about Likert data according to the questionnaire's measurement range. Sample-size criteria and distribution checks should be based on the relevant methodological source.

Do not write in your thesis that “STD below 1 is acceptable” for every variable. The number 1 has meaning only when you are reading a scale with a specific value range, such as a 1 to 5 Likert scale. Even when STD is above 1, the data may still be usable if there are no data-entry errors, the variable is meaningful, and the conditions for the next analysis are satisfied.

For a Likert scale, the safer reporting approach is to state the Mean and STD for each variable or group of variables, then describe the response pattern. You do not need to turn a self-imposed STD threshold into a deletion rule when the methodology for your study does not require it.

How to Read STD in the Output

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

The table below is illustrative output, not the result of a real study. The column names follow the way SPSS displays them so you can compare them with the file currently open.

The N column shows the number of observations used for each variable. In the example, HL4 has an N of 177, lower than the other variables. Check missing values before deciding whether HL4 has a high or low STD. The Valid N (listwise) row shows how many rows remain when you keep only respondents with no missing data across all variables included in the table.

The Minimum and Maximum columns show the smallest and largest observed values. With a 1 to 5 Likert scale, a Maximum of 7 requires you to check the coding or data entry. If Minimum and Maximum both fall between 1 and 5, you still need to check the distribution, but there is no obvious range violation.

For HL1, the Mean is 3.84 and the STD is 0.76. You can describe HL1 as having a relatively high average rating and a dispersion of 0.76 on the scale used. This wording follows the data instead of mechanically labeling the STD as acceptable.

You can also view STD through Analyze > Descriptive Statistics > Frequencies. In the Frequencies dialog, click Statistics, then select Mean, Std. deviation, Minimum, and Maximum. If you need a chart, also select Charts, but a chart does not replace checking the original data table.

How to write this in your thesis

You can adapt the following sentence to your own results: “The descriptive statistics show that [VARIABLE CODE] has a mean of [MEAN] and a standard deviation of [STD], with a minimum value of [MIN] and a maximum value of [MAX] on a scale ranging from [LOWEST LEVEL] to [HIGHEST LEVEL]. This indicates that the responses have [appropriate description] dispersion, while the data remain within the questionnaire's coding range.”

When writing about a group of variables, place the descriptive statistics table in Chapter 4 and explain the variables that stand out. You do not need to repeat the same sentence for 20 variables when the table already provides all the values.

What to Do When STD Seems Unacceptable

First, define what “unacceptable” means in your analysis. If you only see an STD above 1 on a 1 to 5 Likert scale, that is not enough reason to delete the variable. If STD is close to 0 and most respondents selected the same option, the more relevant issue is whether the variable can distinguish between respondents.

Check the Value Range and Coding

Open Analyze > Descriptive Statistics > Frequencies to view Minimum, Maximum, and the frequency of each level. Check Value Labels, missing codes, and reverse-coded variables. A variable that should range from 1 to 5 but contains 0, 6, or 99 may produce an incorrect STD.

If you use 99 for “no opinion,” declare 99 as missing in Variable View, under the Missing column. Then run the statistics again. Do not replace 99 with the Mean just to make STD look better unless you have a clear methodological reason.

Check Data Entry and Unusual Respondents

Compare the original file with the .sav or .csv file. Check rows shifted into the wrong columns, decimal-separator errors, reverse-coded variables, and cases where every answer was entered with the same code. If a few rows are very different, determine whether they are genuine outliers or data-entry errors.

Deleting an outlier requires a criterion and a recorded decision. Keep the original dataset, the processed dataset, and a list of deleted rows. The committee may ask why a respondent was removed, while nobody will ask why STD was rounded to make the table look better.

Recheck the Context of the Variable

A high STD sometimes reflects real differences between respondents. For example, income may vary substantially between people who have just started working and people with many years of experience. In that situation, report STD and also consider the median or groups that fit your research question.

Delete a Variable Only With Overall Evidence

Delete an item only when you have identified a problem with its content, coding, data quality, or scale-test results. Cronbach's Alpha, Corrected Item-Total Correlation, and EFA each have a separate purpose. STD cannot replace those tests.

If you are analyzing a moderating effect, do not confuse a high STD with a moderator variable. A moderator concerns whether one variable changes the strength or direction of a relationship, not merely how dispersed a variable is. See moderator and moderating when you are reading a model with an interaction.

Distinguishing STD From Variance and SE

STD, variance, and Standard Error all relate to variation, but they answer different questions.

STD shows how dispersed observations are around the Mean and uses the same unit as the original variable. If the variable is measured in Likert points, STD is also expressed in those points. This is the statistic you commonly see in the Descriptive Statistics table.

Variance is the square of STD. Because its unit is squared, variance is usually harder to interpret directly. SPSS may display variance in some procedures, but students commonly report STD because readers can understand the level of dispersion more easily.

Standard Error of the Mean, often written as Std. Error, reflects the uncertainty of the sample Mean when it is used to estimate the population Mean. Standard Error decreases as sample size increases, while STD describes differences between individuals in the sample. These two columns should not be used interchangeably.

You also need to distinguish STD from p-value. A large STD does not mean that a result lacks statistical significance, and a small STD does not guarantee a small p-value. For a mediator, read the indirect effect and confidence interval rather than using STD to reach a conclusion. The concept is explained further at mediator.

Common Mistakes

Calling STD a Regression Coefficient

Regression coefficients appear in the Coefficients table, usually with the columns B, Std. Error, Beta, t, and Sig.. Std. Error is not Std. Deviation. Check the table name and column name before interpreting the result.

Applying One STD Threshold to Every Variable

Likert items, age, income, and test scores have different measurement ranges. Comparing STD across variables with different units can be misleading. If you need to compare relative dispersion, state clearly how you standardized the variables and why the comparison is relevant.

Deleting a Variable Only Because STD Is High

A high STD may be a genuine feature of the sample. Check coding, missing values, outliers, and the research purpose first. Deleting a variable simply to make the table look cleaner can remove information and weaken your explanation in Chapter 4.

Failing to Check N

When variables have different Ns, their STDs are calculated from different numbers of observations. Check missing values and state how you handled them. In particular, do not overlook the Valid N (listwise) row before running regression or EFA.

Reporting Only the Mean and Omitting STD

Mean shows the central tendency, while STD shows dispersion. The two statistics are usually read together. A high Mean may be paired with a low or high STD, leading to different interpretations of agreement among respondents.

Frequently asked questions

What is STD in SPSS

STD in SPSS stands for Standard Deviation and is commonly displayed as Std. Deviation. It shows how widely observed values are dispersed around the Mean.

You can view STD through Analyze > Descriptive Statistics > Descriptives, then select Std. deviation under Options.

What standard deviation is acceptable

There is no single number for every variable. With a 1 to 5 Likert scale, read STD in relation to the measurement range, Mean, Minimum, Maximum, and research purpose.

An STD above 1 does not automatically mean that the variable is unacceptable. Before taking action, check coding, missing values, outliers, and the results of the relevant scale tests.

Is a smaller STD always better

A small STD shows that responses are concentrated, but that is not always desirable. If almost every respondent selects the same level, the variable may have limited ability to distinguish between cases.

Read STD together with the frequency of each response level. A moderate STD may be appropriate when it reflects genuine differences in the research sample.

Does a high STD mean that I should delete the variable

Not yet. Check the value range, missing codes, data-entry errors, and outliers first. Then check whether the variable has problems with Corrected Item-Total Correlation, Cronbach's Alpha, or EFA.

If you delete the variable, record the reason and rerun the analysis using a consistent procedure. Do not delete it only because you want STD to fall below a self-imposed number.

Are STD and Standard Error the same

They are different. STD describes dispersion among observations in the sample, while Standard Error describes uncertainty around the sample Mean. The columns have similar names but different meanings and reporting purposes.

When reading the Coefficients table, pay attention to Std. Error. When reading the Descriptive Statistics table, the column for standard deviation is Std. Deviation.

Open the Descriptive Statistics table in your own file, check N, Minimum, Maximum, Mean, and Std. Deviation, then record the variables with unusual values before running the next step. If you need to run the analysis on your .sav or .csv file, DoThesis M4 analysis supports this step.