Introduction
Many nursing students collect ordinal, ranked, skewed, Likert-scale, or non-normal data and then wonder whether Pearson correlation is still appropriate. You may have perceived stress ratings and sleep quality ratings, pain severity ranks and satisfaction ratings, symptom severity scores and quality-of-life scores, medication adherence categories and self-care scores, or ordinal patient satisfaction responses and communication ratings.
These variables often appear in nursing dissertations, DNP projects, capstone projects, public health studies, quantitative research assignments, and Chapter 4 results sections. However, they do not always meet the conditions needed for Pearson correlation.
Spearman correlation analysis in nursing research is useful when a student wants to examine the strength and direction of a monotonic relationship between two variables. A monotonic relationship means the variables tend to move together in a consistent direction, even when the pattern is not perfectly linear. Spearman correlation is especially useful when the data are ordinal, ranked, skewed, affected by outliers, or not normally distributed.
Spearman correlation is also commonly considered when nursing students analyze Likert-type items, symptom ratings, satisfaction ratings, pain categories, adherence categories, or other ordered responses. However, Spearman is not a shortcut for weak analysis. It still requires a clear research question, meaningful ordering of categories, paired observations, independent observations, careful scoring, missing-data review, and responsible interpretation.
Like Pearson correlation, Spearman correlation does not prove cause and effect. It can show that two variables are associated, but it cannot prove that stress caused poor sleep, pain severity caused low satisfaction, or symptom severity caused reduced quality of life. Correlation coefficients must be selected and interpreted carefully to avoid misleading conclusions in medical and nursing research (Mukaka, 2012; Schober et al., 2018).
This guide explains what Spearman correlation is, when nursing students should use it, how it differs from Pearson correlation, how to handle Likert-scale and ordinal data, what assumptions and data conditions matter, how to interpret Spearman’s rho, how to report results in APA 7th edition, and how to explain Spearman correlation results in Chapter 4.
For the broader topic cluster, use the main pillar guide on correlation analysis in nursing research. For the parametric alternative, review Pearson correlation analysis in nursing research. For broader test selection, see statistical tests in nursing research and inferential data analysis in nursing research.
What Is Spearman Correlation Analysis?
Spearman correlation analysis is a nonparametric correlation method used to examine whether two variables tend to move together in a consistent ranked or monotonic pattern. It is also called Spearman rank correlation, Spearman’s rho, or the Spearman correlation coefficient.
Unlike Pearson correlation, which measures linear association using the original values, Spearman correlation works with the rank order of the data. Each value is ranked, and the correlation is calculated using those ranks. This makes Spearman useful when the original values are ordinal, skewed, ranked, or not suitable for Pearson correlation.
Spearman’s rho usually ranges from −1 to +1. A positive rho means higher ranks on one variable tend to go with higher ranks on another variable. A negative rho means higher ranks on one variable tend to go with lower ranks on another variable. A rho close to zero suggests little or no monotonic association.
For example, a nursing student may examine the relationship between perceived stress rating and sleep quality rating. If students with higher stress ratings tend to report poorer sleep quality ratings, the result may show a negative Spearman correlation. The pattern does not need to form a perfect straight line, but it should move in a consistent direction.
A complete nursing interpretation should include the direction, strength, p-value, sample size, confidence interval where available, clinical meaning, and noncausal wording.
What Spearman’s Rho Means
Spearman’s rho is the statistic that summarizes the strength and direction of a monotonic relationship between two variables. It may be reported as rho, ρ, rs, or rₛ depending on software, journal style, or university guidance. In dissertation writing, the most important point is to name the test clearly and use consistent notation.
Spearman’s rho ranges from −1 to +1.
A rho of +1 means a perfect positive monotonic relationship. Participants with higher ranks on one variable also have higher ranks on the other variable.
A rho of −1 means a perfect negative monotonic relationship. Participants with higher ranks on one variable have lower ranks on the other variable.
A rho close to 0 means little or no monotonic relationship.
The sign shows direction, not importance. A rho of −.60 is stronger than a rho of .25 because .60 is farther from zero. The negative sign only means the variables move in opposite directions.
A statistically significant rho may still be weak. For example, rho = .18, p = .02 shows a statistically significant association, but the relationship is weak. A non-significant rho does not always prove that no relationship exists, especially in a small sample. Students should interpret rho together with sample size, p-value, confidence interval where available, measurement quality, and nursing context.
The American Statistical Association warns that p-values do not measure effect size or practical importance and should not be used alone to make scientific conclusions (Wasserstein & Lazar, 2016).
Table 1. General Guide for Interpreting Spearman Correlation Coefficients
| Spearman’s rho | General strength | Direction | Nursing interpretation caution |
|---|---|---|---|
| 0 to ±.09 | Very weak or negligible | Positive or negative | May have limited practical meaning unless the clinical outcome is highly important |
| ±.10 to ±.29 | Weak | Positive or negative | Can become statistically significant in large samples but should not be exaggerated |
| ±.30 to ±.49 | Moderate | Positive or negative | May be meaningful depending on the clinical, educational, or public health context |
| ±.50 to ±.69 | Strong | Positive or negative | Check whether variables measure related constructs or share similar scoring |
| ±.70 to ±1.00 | Very strong | Positive or negative | May suggest overlapping variables, restricted categories, ceiling effects, or scoring patterns that need review |
These thresholds are general guides, not absolute rules. A weak Spearman correlation may matter if the outcome involves falls, medication safety, treatment adherence, patient satisfaction, or readmission risk. A strong Spearman correlation may need careful review if the two variables measure nearly the same construct.
Spearman Correlation Analysis in Nursing Research: When Should Students Use It?
Nursing students should use Spearman correlation when the research question asks about a relationship, association, or correlation and the variables are ordinal, ranked, skewed, non-normal, or better understood through rank order.
Spearman correlation is appropriate when the variables are measured on the same participants, the categories have a meaningful order, and the relationship is monotonic. It is also useful when Pearson assumptions are not met because the data are not normally distributed, contain outliers, or are ordinal rather than continuous.
Spearman correlation is commonly used in cross-sectional nursing surveys, DNP project evaluations, patient satisfaction studies, symptom-rating studies, public health questionnaires, and observational healthcare datasets. It is not the correct test when the student is comparing groups, testing pretest-posttest change, or predicting an outcome using several variables.
Strong nursing research questions for Spearman correlation include:
What is the relationship between perceived stress level and sleep quality rating among nursing students?
Is pain severity associated with patient satisfaction rating among postoperative patients?
What is the relationship between medication adherence ranking and self-care behavior score?
Is symptom severity related to quality-of-life score among patients with chronic illness?
Is communication rating associated with overall patient satisfaction rating?
Is burnout score related to job satisfaction score when both variables are skewed?
Spearman correlation fits these examples because the student is examining whether two variables move together in an ordered or monotonic pattern. If the student wants to compare group means, examine pretest-posttest change, or predict an outcome from multiple variables, another test may be more appropriate.
A student who is unsure whether Spearman fits the research question should review types of data analysis in quantitative research before finalizing Chapter 3.
Variables Suitable for Spearman Correlation
Spearman correlation works well with variables that can be ranked or meaningfully ordered. These may include ordinal variables, ranked variables, Likert-type items, skewed scale scores, non-normal clinical scores, symptom severity scores, satisfaction ratings, adherence categories, pain ratings, quality-of-life scores, and questionnaire scores that violate Pearson assumptions.
The key requirement is that the values must have a meaningful order. Pain severity rated as mild, moderate, and severe has an order. Satisfaction rated as very dissatisfied, dissatisfied, neutral, satisfied, and very satisfied has an order. A variable such as blood type does not have a meaningful order, so Spearman correlation would not be appropriate.
Spearman may also be used for continuous variables when Pearson assumptions are not met. For example, if two questionnaire total scores are strongly skewed or affected by outliers, Spearman may be safer than Pearson because it analyzes rank order rather than the original values.
Table 2. Nursing Variables Suitable for Spearman Correlation
| Variable 1 | Variable 2 | Why Spearman may fit | Caution before analysis |
|---|---|---|---|
| Pain severity rating | Satisfaction rating | Both are ordinal or rating-based | Confirm that rating categories have a meaningful order |
| Stress rating | Sleep quality rating | Both may be Likert-type or ordinal | Check whether higher scores mean better or worse sleep |
| Adherence category | Self-care behavior score | One variable may be ordinal and the other rankable | Confirm category coding and score direction |
| Symptom severity score | Quality-of-life score | Scores may be skewed or non-normal | Check scoring rules and missing items |
| Communication rating | Patient satisfaction rating | Both may be ordinal survey responses | Consider ties, ceiling effects, and restricted range |
| Burnout score | Job satisfaction score | Spearman may fit if scale scores are skewed | Check whether Pearson assumptions are violated |
| Patient activation level | Medication self-management score | Activation level may be ordinal | Confirm that category order reflects increasing activation |
| Functional limitation category | Quality-of-life score | Functional limitation may be ordered | Check whether categories are clinically meaningful |
For basic preparation of variable distributions, frequency tables, and summary statistics, students may review descriptive data analysis in nursing research.
Spearman Correlation and Likert Scale Data
Likert-scale data create one of the most common test-selection problems in nursing research. Many nursing students collect questionnaire responses using 1-to-5 or 1-to-7 response options, then ask whether they should use Pearson or Spearman correlation.
The first step is to distinguish between a single Likert item and a Likert scale total score.
A single Likert item asks one question, such as “Rate your satisfaction from 1 to 5.” This type of response is usually ordinal because the categories have order, but the distance between categories may not be equal. The difference between “neutral” and “satisfied” may not be the same as the difference between “satisfied” and “very satisfied.”
A Likert scale total score is different. It combines multiple related items into a total or average score, often from a validated instrument. For example, a patient satisfaction questionnaire may include several items about communication, respect, responsiveness, discharge teaching, and overall care. The total score may sometimes be treated as approximately continuous if the instrument is designed that way, the score has enough range, the distribution is acceptable, and the committee approves the approach.
Sullivan and Artino (2013) explain that individual Likert-type items are ordinal, while Likert scale data require careful interpretation depending on how the items are combined and analyzed (Sullivan & Artino, 2013). Jamieson (2004) also discusses the long-standing debate about treating Likert responses as interval-level data and cautions against careless use of parametric methods for ordinal scales (Jamieson, 2004).
Spearman correlation may be safer when the student is analyzing single Likert items, ordinal ratings, strongly skewed variables, responses that cluster at the top or bottom of the scale, data with many ties, or data where the committee expects a nonparametric approach.
Pearson correlation may be defensible for summed or averaged validated scale scores when the scale behaves approximately continuously and assumptions are acceptable. However, students should not assume that all Likert data are the same. Chapter 3 should justify the test choice based on measurement level, scoring method, distribution, sample size, and approved analysis plan.
Single Item vs Scale Score Example
A single item would be:
“Rate your satisfaction with discharge teaching from 1 = very dissatisfied to 5 = very satisfied.”
A scale score would be:
A total patient education satisfaction score created by summing 10 validated items.
A single item is usually treated as ordinal. A total score may sometimes be treated as approximately continuous, but the decision depends on instrument scoring, distribution, sample size, committee expectations, and assumption checks.
Why Likert Data Often Create Ties
Ties occur when multiple participants choose the same response option. Likert items often create ties because there are only a few response categories. For example, if 60 out of 100 patients select “satisfied,” many participants receive the same rank.
Ties do not automatically invalidate Spearman correlation. Most statistical software adjusts ranks when tied values occur. However, many ties reduce the amount of rank information in the data. If almost all participants choose the same category, the variable has little variation, and the correlation may be weak, unstable, or difficult to interpret.
For example, if nearly every patient rates nurse communication as 5 out of 5, Spearman correlation may show little association with satisfaction because communication ratings have a ceiling effect. This does not necessarily prove that communication is unrelated to satisfaction. It may mean the item did not capture enough variation.
Practical Rule for Nursing Students
Use Spearman correlation more confidently when Likert responses are ordinal, meaningfully ordered, variable enough to rank, and aligned with a relationship-based research question. Be more cautious when responses are heavily tied, almost all responses fall in one category, the item is poorly worded, or the instrument scoring guide requires a different analytic approach.
Spearman Correlation Assumptions and Data Conditions
Spearman correlation is less restrictive than Pearson correlation, but it still has assumptions and data conditions. It should not be used automatically just because a student is uncertain.
The main data conditions include:
The variables should be ordinal, continuous, or rankable.
Both variables should be measured on the same participants.
Observations should be independent.
The relationship should be monotonic.
Categories should have a meaningful order.
Scores should be coded correctly.
Missing data should be handled appropriately.
Impossible values and data-entry errors should be checked.
Sample size should be adequate for the planned analysis.
Spearman does not require normally distributed variables in the same way Pearson does. That is one reason it is useful for ordinal, skewed, and non-normal nursing data. However, Spearman still requires a sensible research question and valid measurement. A nonparametric test does not fix poor coding, weak measurement, missing data, or a research question that calls for a different analysis.
Monotonic Relationship
A monotonic relationship means that as one variable increases, the other tends to increase or decrease consistently. The pattern does not need to be a perfect straight line.
For example, as stress ratings increase, sleep quality ratings may tend to worsen. The relationship may not be perfectly linear because sleep quality may decline slowly at first and more sharply at higher stress levels. Spearman correlation can still be appropriate if the overall direction is consistently downward.
A relationship is not monotonic if it changes direction. For example, if moderate exercise is associated with better sleep but very high exercise is associated with worse sleep, the pattern may be curved rather than monotonic. Spearman may not summarize that relationship well.
Ranked or Ordinal Data
Spearman works by ranking values. This makes it useful for ordinal or non-normal data. However, the categories must still have a meaningful order.
Pain severity categories such as mild, moderate, and severe are ordered. Satisfaction categories such as very dissatisfied to very satisfied are ordered. Blood type, hospital department name, and marital status categories do not naturally follow a numeric order for correlation purposes.
Ties in the Data
Ties occur when many participants have the same score or rating. Ties are common in Likert-scale data, ordinal symptom categories, satisfaction ratings, and short clinical rating scales.
Most statistical software adjusts for ties when calculating Spearman correlation. However, ties affect interpretation because Spearman depends on rank ordering. If many participants share the same score, the analysis has less information about who ranks above or below whom.
For example, a 5-point satisfaction item may have many repeated responses. If 80% of participants choose “satisfied” or “very satisfied,” the distribution may show a ceiling effect. Spearman correlation may still run, but the result should be interpreted cautiously because limited variation can weaken or obscure relationships.
Students should report the result honestly and avoid overstating conclusions when tied responses, ceiling effects, or floor effects limit the analysis.
Outliers
Spearman correlation is less sensitive to extreme raw values than Pearson because it uses ranks rather than original values. However, students should still check for impossible values, data-entry errors, and unusual response patterns.
For example, a satisfaction score coded as 55 on a 1-to-5 scale is not an outlier to interpret; it is likely an error to correct. A symptom severity score outside the instrument’s valid range should also be checked before analysis.
Independent Observations
Each participant’s response should be independent. If the same participant contributes repeated measurements, or if patients are nested within units, hospitals, classrooms, or clinical groups, the independence assumption may be weakened. Spearman correlation may not be enough for clustered or repeated-measures data.
For example, if a student collects pain ratings from the same patients every day for seven days, those observations should not be treated as seven independent participants. The correct analysis depends on the design and research question.
How to Check Whether Spearman Correlation Is Appropriate
Students should check whether Spearman correlation matches the research question, variables, and data pattern before interpreting results.
Practical checks include reviewing the measurement level, confirming that categories have a meaningful order, inspecting frequency tables, reviewing descriptive statistics, using histograms or bar charts, inspecting scatterplots or ranked scatterplots, checking missing data, checking impossible values, verifying questionnaire scoring, confirming that both variables are measured on the same cases, and assessing whether the relationship appears monotonic.
Spearman Correlation Checklist for Nursing Students
Use this checklist before selecting Spearman correlation:
| Question | Why it matters |
|---|---|
| Am I examining the relationship between two variables? | Spearman is for association, not group comparison |
| Are the variables ordinal, ranked, skewed, non-normal, or rankable? | Spearman is useful when rank-order analysis fits |
| Are both variables measured on the same participants? | Correlation requires paired values |
| Do the categories have a meaningful order? | Spearman cannot rank unordered categories meaningfully |
| Does the relationship appear monotonic? | Spearman summarizes monotonic association |
| Have missing values been checked? | Missing values reduce valid paired cases |
| Have impossible values or scoring errors been checked? | Bad coding can distort ranks |
| Are reverse-coded items handled correctly? | Wrong scoring can reverse or weaken the correlation |
| Is the sample size adequate? | Small samples can produce unstable estimates |
| Would Pearson be better? | Pearson may fit continuous, normal, linear data |
| Can the result be reported without causal language? | Correlation does not prove causation |
If you cannot answer these questions confidently, Nursing Dissertation Help can support you with test selection, Likert-scale decisions, SPSS output interpretation, and Chapter 4 reporting.
Scatterplots and Monotonic Relationships
Scatterplots are still useful for Spearman correlation. Even though Spearman uses ranks, students should still inspect whether the relationship makes sense visually and clinically.
A scatterplot can show the direction of the relationship, monotonic pattern, clusters, outliers, flat patterns, irregular patterns, ceiling effects, floor effects, and whether unusual values drive the result.
How to Check Monotonicity in Practice
A monotonic pattern does not need to be perfectly straight. The question is whether the overall pattern moves consistently upward or consistently downward.
A positive monotonic pattern means that as one variable increases, the other generally increases. For example, higher communication ratings may generally go with higher satisfaction ratings.
A negative monotonic pattern means that as one variable increases, the other generally decreases. For example, higher stress ratings may generally go with poorer sleep quality ratings.
A non-monotonic pattern changes direction. For example, mild stress may have little relationship with sleep, moderate stress may reduce sleep quality, and extremely high stress may create a different pattern because of medication use, fatigue, or clinical intervention. In that case, one Spearman rho may oversimplify the relationship.
What Students Should Look for in a Scatterplot
Look for an overall upward or downward pattern.
Look for many tied values or stacked points.
Look for ceiling or floor effects.
Look for clusters that may represent subgroups.
Look for impossible values or coding errors.
Look for patterns that curve or change direction.
Look for one subgroup driving the result.
A statistically significant rho should not be interpreted blindly. A non-significant rho may reflect small sample size, restricted range, poor measurement, too many tied responses, or a relationship that is not monotonic.
Spearman vs Pearson Correlation
Spearman and Pearson both examine relationships between variables, but they answer different statistical questions.
Pearson correlation measures the strength and direction of a linear relationship between two continuous variables. It works best when the variables are approximately normally distributed, linearly related, and free from serious outliers.
Spearman correlation measures the strength and direction of a monotonic rank-order relationship. It is useful when the data are ordinal, ranked, skewed, non-normal, or affected by outliers. Spearman may be safer for many nursing datasets that use ordinal responses, symptom categories, satisfaction ratings, or skewed questionnaire scores.
Table 3. Spearman vs Pearson Correlation in Nursing Research
| Feature | Spearman correlation | Pearson correlation | Student decision point |
|---|---|---|---|
| Main purpose | Measures monotonic rank-order association | Measures linear association | Look at the research question and data pattern |
| Best for | Ordinal, ranked, skewed, or non-normal data | Continuous variables with linear patterns | Check measurement level and assumptions |
| Data used | Ranks of the values | Original values | Spearman reduces sensitivity to raw-value extremes |
| Relationship type | Monotonic | Linear | A monotonic pattern does not need to be perfectly straight |
| Outlier sensitivity | Less sensitive to extreme raw values | More sensitive to outliers | Always check unusual values |
| Likert data | Often safer for single ordinal items | Sometimes used for scale totals if assumptions fit | Distinguish item-level data from total scores |
| Example | Stress rating and sleep quality rating | Medication adherence score and systolic blood pressure | Match test to variable type and pattern |
For a deeper explanation of Pearson correlation, use the published guide on Pearson correlation analysis in nursing research.
Spearman Correlation vs Other Statistical Tests
Spearman correlation is not the right test for every nursing research question. It is used for association between two ordinal, ranked, skewed, non-normal, or rankable variables. Other goals require different tests.
Table 4. When Spearman Correlation Is Not the Right Test
| Research goal | Example question | Better test | Why |
|---|---|---|---|
| Compare two group means | Do intervention and control groups differ in anxiety scores? | Independent-samples t-test or Mann–Whitney U | The question compares groups |
| Compare three or more group means | Do satisfaction scores differ across three units? | ANOVA or Kruskal–Wallis | The question compares more than two groups |
| Test pretest-posttest change | Did knowledge scores improve after education? | Paired t-test or Wilcoxon signed-rank | The question tests change over time |
| Examine nominal categorical association | Is readmission status associated with discharge destination? | Chi-square or Fisher’s exact test | Both variables are nominal categorical |
| Predict an outcome | Do stress, age, and workload predict sleep quality? | Regression | The question involves prediction or adjustment |
| Examine two continuous normal variables | Are adherence score and blood pressure linearly related? | Pearson correlation | Pearson may fit if assumptions are met |
This section should stay brief in a Spearman article. If your real question is about group comparisons, prediction, or categorical association, use a broader test-selection guide rather than forcing Spearman correlation.
Sample Size and Statistical Power in Spearman Correlation
Sample size affects Spearman correlation in important ways. Small samples can produce unstable correlation estimates. A rho calculated from 15 or 20 participants can change noticeably if only a few rankings change. Small samples may also fail to detect a meaningful monotonic relationship because statistical power is limited.
Large samples create a different issue. A weak Spearman correlation can become statistically significant when the sample is large. For example, rho = .16 may be statistically significant in a large sample but still have limited clinical meaning. A significant result should not be described as strong unless the coefficient supports that interpretation.
A moderate rho may fail to reach significance in a small sample. For example, rho = .42, p = .07 may be non-significant, but the coefficient may still suggest a possible pattern worth discussing cautiously as a limitation or direction for future research.
Students should interpret Spearman correlation using rho, sample size, p-value, confidence interval where available, measurement quality, and nursing context. Brydges (2019) emphasizes that effect-size interpretation, statistical power, and sample size should be understood together rather than treated as isolated rules (Brydges, 2019).
P-Values in Spearman Correlation
The null hypothesis in Spearman correlation usually states that there is no monotonic association between the two variables in the population.
A p-value helps the student evaluate whether the observed rho is statistically significant under the null hypothesis. In many nursing dissertations, p < .05 is treated as statistically significant. However, the p-value does not show the strength of the relationship, does not prove causation, does not prove clinical importance, and is affected by sample size.
For example, rho = .20, p = .01 may be statistically significant but weak. A result of rho = .45, p = .07 may be non-significant in a small sample but still worth discussing cautiously as a possible pattern or limitation.
Students should interpret rho, p-value, sample size, and clinical context together. The American Statistical Association’s statement on p-values warns that p-values do not measure the size or importance of an effect and should not be used as the only basis for interpretation (Wasserstein & Lazar, 2016).
Confidence Intervals for Spearman’s Rho
A confidence interval for Spearman’s rho gives a range of plausible values for the population monotonic association. It helps students understand uncertainty around the sample result.
A narrow confidence interval suggests a more precise estimate. A wide confidence interval suggests more uncertainty, often due to small sample size, variable data, many ties, restricted response range, or unstable estimates.
For example:
A Spearman correlation showed a statistically significant relationship between perceived stress and sleep quality rating, rho = −.38, p = .004, 95% CI [−.57, −.15].
This result suggests a negative monotonic association. The confidence interval suggests that the population association may plausibly range from weak-to-moderate negative to moderate negative.
Confidence intervals are useful because two studies can have the same rho but different precision. A study with rho = −.38 and 95% CI [−.57, −.15] gives a clearer estimate than a study with rho = −.38 and 95% CI [−.72, .05]. The second result is more uncertain because the interval includes values close to zero.
Bonett and Wright (2000) discuss sample size requirements for estimating Pearson, Kendall, and Spearman correlations, emphasizing the importance of precision when estimating correlation coefficients (Bonett & Wright, 2000). IBM’s SPSS Bivariate Correlations documentation also identifies confidence interval options for Pearson and Spearman correlations, although students may need the correct software version, settings, extensions, or alternative tools depending on their workflow (IBM, n.d.-a).
Not all basic outputs provide confidence intervals automatically. Some software packages or procedures report rho and p-value but not a confidence interval. Students may need SPSS settings, Jamovi, JASP, R, or professional statistical support to obtain and interpret confidence intervals correctly.
How to Explain a Spearman Confidence Interval in Chapter 4
A strong Chapter 4 explanation should connect the confidence interval to uncertainty:
The Spearman correlation between perceived stress rating and sleep quality rating was statistically significant, rho = −.38, p = .004, 95% CI [−.57, −.15]. The confidence interval suggests that the population association was likely negative, ranging from weak-to-moderate to moderate in strength.
A weak explanation says only:
The correlation was significant.
The stronger version is better because it reports the estimate, significance, uncertainty, direction, and strength.
Missing Data in Spearman Correlation
Missing data can reduce the number of valid cases in Spearman correlation. If a participant is missing one of the two variables, that participant cannot be included in that specific correlation.
For example, if some participants completed the stress rating but skipped the sleep quality rating, they cannot be included in the Spearman correlation between stress and sleep quality. The analysis requires paired values for both variables.
Software may use pairwise or listwise deletion depending on the procedure and settings. Pairwise deletion uses all available cases for each variable pair, which means different correlations in a matrix can have different sample sizes. Listwise deletion uses only cases with complete data across all selected variables, which can reduce the sample size sharply.
Students should check the valid n for each Spearman correlation. Missing-data handling should be reported or explained when it affects the results. In observational research, transparent reporting of missing data is part of responsible methods reporting. The STROBE guidance encourages clear reporting of study methods, participants, variables, and missing-data issues in observational studies (STROBE, n.d.; von Elm et al., 2007).
One-Tailed vs Two-Tailed Significance in Spearman Correlation
Spearman correlation can be tested using a one-tailed or two-tailed significance test. Nursing students should be cautious with this choice.
A two-tailed test is usually the safer default because it tests whether an association exists in either direction. A one-tailed test should only be used when the direction was specified before analysis and the opposite direction would not be interpreted as support for the hypothesis.
Students should not choose a one-tailed test after seeing the results just to make a finding significant. Chapter 4 should match the approved Chapter 3 analysis plan, proposal, committee guidance, or institutional expectations.
For example, if the proposal says, “There is a relationship between perceived stress and sleep quality,” a two-tailed test is usually appropriate. If the proposal specifically predicts that “higher perceived stress is associated with poorer sleep quality,” a one-tailed test may be defensible only if it was justified before analysis.
IBM’s SPSS documentation states that one-tailed probabilities may be selected when the direction of association is known in advance; otherwise, two-tailed probabilities should be selected (IBM, n.d.-a).
Spearman Correlation in Chapter 3, Chapter 4, and Chapter 5
Spearman correlation should connect logically across the dissertation. It should not appear suddenly in Chapter 4 without a clear Chapter 3 justification.
Chapter 3: Methodology
In Chapter 3, students should explain why Spearman correlation was selected. The explanation should name the research question, identify the variables, describe the measurement level, and explain why Spearman is more appropriate than Pearson.
A strong Chapter 3 statement may look like this:
Spearman rank correlation will be used to examine the relationship between perceived stress rating and sleep quality rating because both variables are ordinal and the research question asks whether the variables are monotonically associated. Data conditions will be assessed by reviewing score coding, frequency distributions, missing values, and scatterplots.
Chapter 4: Results
In Chapter 4, students should report the actual findings. This includes valid n, Spearman’s rho, p-value, confidence interval where available, direction, strength, and nursing interpretation.
Chapter 4 should not use causal language. It should say that variables were associated, related, or correlated.
Chapter 5: Discussion
In Chapter 5, students should explain what the result means in relation to the research question, nursing practice, prior literature, limitations, and future research. If the study is cross-sectional or observational, causal interpretation should remain cautious.
How to Run Spearman Correlation in SPSS
SPSS is commonly used for nursing dissertation statistics. This section provides a practical overview, not a screenshot-level tutorial.
A typical SPSS workflow includes preparing and cleaning the dataset, checking variable coding, checking questionnaire scoring, reviewing missing values, reviewing descriptive statistics, inspecting the pattern of the relationship, going to Analyze > Correlate > Bivariate, selecting the variables, selecting Spearman, choosing two-tailed significance unless there is a justified directional hypothesis, and interpreting the Spearman correlation coefficient, significance value, and sample size.
IBM’s official Bivariate Correlations documentation states that SPSS computes Pearson’s correlation coefficient, Spearman’s rho, and Kendall’s tau-b with significance levels and identifies the menu path as Analyze > Correlate > Bivariate (IBM, n.d.-a).
Students should justify why Spearman was selected. The justification may involve ordinal measurement, ranked data, skewed variables, non-normal distributions, outliers, or a monotonic rather than linear relationship.
How to Run Spearman Correlation in Excel
Excel does not provide Spearman correlation as directly as SPSS. Students may need to rank the variables first and then correlate the ranks. Microsoft’s RANK.EQ function returns the rank of a number within a list, while the CORREL function returns the correlation coefficient of two cell ranges (Microsoft, n.d.-a; Microsoft, n.d.-b).
In simple terms, Excel-based Spearman analysis usually requires two stages: first rank each variable, then calculate the correlation between the ranked variables. This is useful to know, but dissertation students should avoid treating Excel output as automatically APA-ready.
Excel can support basic exploration, ranking, simple correlation checks, and preliminary scatterplots. However, it may not provide APA-ready interpretation, confidence intervals, assumption guidance, missing-data diagnostics, or dissertation-level reporting support.
For dissertation-level Spearman correlation, SPSS, Jamovi, JASP, R, or professional statistical support may be more appropriate. For broader Excel guidance, review Using Excel for Data Analysis. A separate article on How to Do Correlation Analysis in Excel for Nursing Research covers the full Excel procedure without overloading this Spearman-focused guide.
How to Interpret Spearman Correlation Results
A strong Spearman interpretation should include more than “there was a significant relationship.” Students should explain the direction, strength, p-value, sample size, confidence interval where available, practical or clinical meaning, whether the finding answers the research question, and whether causation can be inferred.
Positive Significant Spearman Correlation
Example:
rho = .46, p < .001
Interpretation:
There was a statistically significant moderate positive Spearman correlation between communication rating and patient satisfaction rating. Higher communication ratings were associated with higher patient satisfaction ratings.
Negative Significant Spearman Correlation
Example:
rho = −.38, p = .004
Interpretation:
There was a statistically significant moderate negative Spearman correlation between perceived stress rating and sleep quality rating. Higher stress ratings were associated with poorer sleep quality ratings.
Weak but Significant Spearman Correlation
Example:
rho = .20, p = .010
Interpretation:
There was a statistically significant but weak positive Spearman correlation between medication adherence ranking and self-care behavior score. Although the result was statistically significant, the relationship was weak and should be interpreted cautiously.
Moderate Non-Significant Spearman Correlation in a Small Sample
Example:
rho = −.42, p = .070
Interpretation:
The Spearman correlation was not statistically significant at the .05 level, although the coefficient suggested a moderate negative pattern. The small sample size may have limited statistical power.
Non-Significant Spearman Correlation
Example:
rho = .11, p = .382
Interpretation:
There was a weak, non-significant Spearman correlation between pain severity rating and satisfaction rating. The data did not show a statistically significant monotonic association between the variables.
Spearman Correlation With a Wide Confidence Interval
Example:
rho = −.33, p = .061, 95% CI [−.61, .02]
Interpretation:
The Spearman correlation was not statistically significant, and the wide confidence interval suggests uncertainty about the population association. The true association may range from negligible to moderate negative.
These examples show why students should not report only the p-value. Nursing interpretation should explain what the association means in relation to the research question.
Correlation Does Not Mean Causation
Spearman correlation shows association, not cause and effect. Ordinal or ranked association does not prove that one variable caused another.
This caution is especially important in cross-sectional nursing studies. Many dissertation datasets measure participants at one point in time. If stress rating and sleep quality rating are associated, the study does not prove whether stress affected sleep, poor sleep affected stress, or a third factor influenced both variables.
Confounding variables may also explain an observed association. For example, workload, shift pattern, anxiety, chronic illness, medication use, or family responsibilities may partly explain a relationship between stress and sleep quality.
Table 5. Wrong vs Correct Wording for Spearman Correlation Results
| Weak or incorrect wording | Better APA-style wording |
|---|---|
| Stress caused poor sleep quality. | Higher stress ratings were associated with poorer sleep quality ratings. |
| Pain severity affected satisfaction. | Pain severity rating showed a negative association with satisfaction rating. |
| Medication adherence influenced self-care. | Medication adherence ranking was positively associated with self-care behavior score. |
| Spearman proved that symptoms reduce quality of life. | Spearman correlation showed a negative association between symptom severity and quality-of-life score, but causality cannot be inferred. |
| The significant result means stress has an effect on sleep. | The significant result suggests a monotonic association between stress rating and sleep quality rating. |
Noncausal wording protects the accuracy and credibility of Chapter 4 and Chapter 5.
Statistical Significance vs Clinical Significance
Statistical significance and clinical significance are not the same. A Spearman correlation can be statistically significant but clinically weak. It can also be clinically interesting but not statistically significant in a small sample.
Statistical significance depends partly on sample size. Large samples can make weak correlations statistically significant. Small samples may fail to detect moderate associations.
Clinical significance asks whether the relationship matters for nursing practice, patient outcomes, education, safety, or healthcare improvement. A weak association between symptom burden and medication adherence may still deserve attention if the outcome affects patient safety or chronic disease management.
Students should consider sample size, effect size, confidence intervals, clinical relevance, patient care implications, measurement quality, and study design. Weak correlations should not be exaggerated. Non-significant results should not be dismissed without considering sample size, confidence intervals, and limitations.
Spearman Correlation and Correlation Matrix
A Spearman correlation matrix is a table showing Spearman correlations among several ordinal, ranked, skewed, or non-normal variables. It can be useful in Chapter 4 when a student has several relationship-based research questions.
A Spearman correlation matrix may examine relationships among stress rating, sleep quality rating, pain severity, satisfaction rating, symptom severity, quality-of-life score, adherence category, and self-care behavior score.
A matrix can be useful for preliminary analysis, questionnaire score relationships, and Chapter 4 presentation. However, students should avoid over-interpreting too many correlations. The interpretation should stay tied to the approved research questions.
Missing data can also complicate a matrix. If pairwise deletion is used, different correlations may have different sample sizes. Students should check and report valid n values where needed.
APA Style provides sample table guidance and table setup principles for presenting statistical results clearly (American Psychological Association, n.d.-a; American Psychological Association, n.d.-b).
Table 6. Sample Spearman Correlation Matrix Layout
| Variable | 1 | 2 | 3 | 4 |
|---|---|---|---|---|
| 1. Perceived stress rating | — | |||
| 2. Sleep quality rating | −.38** | — | ||
| 3. Symptom severity score | .41** | −.34* | — | |
| 4. Quality-of-life score | −.46** | .39** | −.52** | — |
Note. Values are examples only. p < .05. p < .01. Replace with actual dissertation results.
How to Report Spearman Correlation in APA 7th Edition
APA reporting is one of the most important parts of Chapter 4. A strong Spearman correlation report should be clear, complete, and noncausal.
Students should report the statistical test, variables being correlated, Spearman’s rho, p-value, sample size where helpful, confidence interval where available or required, direction of the relationship, strength of the relationship, plain-language interpretation, missing-data handling if it affects valid n, and noncausal wording.
APA Style’s statistics guidance states that statistical symbols such as p, r, and N should be italicized, and exact p-values should generally be reported where possible (American Psychological Association, 2024).
APA Reporting Rules for Spearman Correlation
Use correct statistical notation, such as Spearman’s rho, ρ, or rs, depending on your program guidance.
Use italic statistical symbols when appropriate.
Report exact p-values when possible.
Use p < .001 when the value is very small.
Do not write p = .000.
Report confidence intervals when available or required.
Report valid n if missing data reduced the sample.
Use association language.
Avoid causal language.
Wrong vs Correct APA Reporting Example
Weak report:
The Spearman correlation was significant, so stress affected sleep quality.
Better report:
A Spearman rank correlation was conducted to examine the relationship between perceived stress rating and sleep quality rating. The results showed a statistically significant negative association, ρ = −.38, p = .004, indicating that higher perceived stress ratings were associated with poorer sleep quality ratings. Causality cannot be inferred from the correlational design.
The stronger version is better because it names the test, identifies the variables, reports rho and p, explains direction and meaning, and avoids causal language.
APA Write-Up Examples for Spearman Correlation
1. Significant Positive Spearman Correlation
A Spearman rank correlation was conducted to examine the relationship between nurse communication rating and patient satisfaction rating. The results showed a statistically significant positive association, ρ = .46, p < .001. This finding indicates that higher communication ratings were associated with higher patient satisfaction ratings.
2. Significant Negative Spearman Correlation
A Spearman rank correlation was conducted to examine the relationship between perceived stress rating and sleep quality rating among nursing students. The results showed a statistically significant negative association, ρ = −.38, p = .004. This finding indicates that higher stress ratings were associated with poorer sleep quality ratings.
3. Non-Significant Spearman Correlation
A Spearman rank correlation was conducted to examine the relationship between pain severity rating and satisfaction rating. The results showed a weak, non-significant positive association, ρ = .11, p = .382. This finding indicates that pain severity rating was not significantly associated with satisfaction rating in the sample.
4. Weak but Significant Spearman Correlation
A Spearman rank correlation was conducted to examine the relationship between medication adherence ranking and self-care behavior score. The results showed a statistically significant but weak positive association, ρ = .20, p = .010. Although the result was statistically significant, the weak coefficient suggests that the association should be interpreted cautiously.
5. Spearman Correlation With Confidence Interval
A Spearman correlation was conducted to examine the relationship between symptom severity score and quality-of-life score. The results showed a statistically significant negative association, ρ = −.44, p < .001, 95% CI [−.61, −.24]. The confidence interval suggests that the population association may range from weak-to-moderate negative to strong negative.
6. Spearman Correlation Matrix Summary
Spearman correlations were conducted to examine relationships among perceived stress rating, sleep quality rating, symptom severity score, and quality-of-life score. Perceived stress rating was negatively associated with sleep quality rating, ρ = −.38, p = .004, and positively associated with symptom severity score, ρ = .41, p = .002. These findings suggest that higher stress ratings were associated with poorer sleep quality ratings and higher symptom severity scores, but causality cannot be inferred from the correlational design.
Sample Spearman Correlation Table for Chapter 4
Table 7. Sample APA-Style Spearman Correlation Results Table
| Variable pair | n | Spearman’s rho | 95% CI | p-value | Direction | Interpretation |
|---|---|---|---|---|---|---|
| Stress rating and sleep quality rating | 82 | −.38 | [−.57, −.15] | .004 | Negative | Higher stress was associated with poorer sleep quality |
| Pain severity rating and satisfaction rating | 76 | −.29 | [−.48, −.06] | .014 | Negative | Higher pain severity was associated with lower satisfaction |
| Communication rating and satisfaction rating | 104 | .46 | [.29, .60] | < .001 | Positive | Higher communication ratings were associated with higher satisfaction |
| Symptom severity and quality-of-life score | 90 | −.44 | [−.61, −.24] | < .001 | Negative | Higher symptom severity was associated with lower quality of life |
| Adherence category and self-care behavior | 68 | .20 | [−.04, .42] | .096 | Positive | The association was weak and not statistically significant |
Note. n = valid paired cases; CI = confidence interval; rho = Spearman rank correlation coefficient; p = probability value. Values are examples only.
Common Mistakes Nursing Students Make With Spearman Correlation
Using Pearson When Spearman Would Be More Appropriate
Students sometimes use Pearson correlation for ordinal, skewed, or non-normal data without checking assumptions. Spearman may be safer when variables are ranked, ordinal, skewed, or monotonic rather than linear.
Using Spearman Without Explaining Why
Students should justify Spearman in Chapter 3. The reason may involve ordinal data, Likert-type items, non-normal distributions, outliers, or monotonic relationships.
Treating All Likert Data the Same Way
A single Likert item is not the same as a validated scale total. Students should distinguish item-level data from summed or averaged scale scores.
Ignoring Whether Categories Have a Meaningful Order
Spearman requires meaningful ordering. Nominal categories such as blood type or hospital department should not be forced into a rank correlation.
Ignoring Monotonicity
Spearman measures monotonic association. If the relationship is irregular, flat, curved in different directions, or not meaningful, rho may not summarize it well.
Ignoring Ties in the Data
Many tied scores can affect interpretation, especially when responses cluster heavily in one category.
Ignoring Ceiling and Floor Effects
If most participants choose the highest or lowest response category, the variable may have too little variation for meaningful correlation interpretation.
Ignoring Missing Data
Students should check valid n values and explain missing-data handling when it affects the results.
Confusing Pairwise and Listwise Deletion
Pairwise deletion can produce different sample sizes across a correlation matrix. Listwise deletion may reduce the sample size across all variables.
Claiming Causation
Spearman correlation does not prove cause and effect. Use association language.
Reporting Only P-Values
Always report Spearman’s rho, direction, strength, and p-value.
Over-Interpreting Weak Correlations
A weak but statistically significant correlation should not be described as strong or highly meaningful.
Using Spearman When Regression Is Needed
Regression is more appropriate when the goal is prediction or adjustment for covariates.
Using Spearman When a Group-Comparison Test Is Needed
Use t-tests, ANOVA, paired tests, Wilcoxon signed-rank, Mann–Whitney U, Kruskal–Wallis, or chi-square when the research question requires those tests.
Using Unscored Questionnaire Items Instead of Correct Scale Totals
Students should follow instrument scoring instructions before running any correlation analysis.
Using One-Tailed Tests Without Justification
One-tailed tests should be planned before analysis and justified in the methodology.
Failing to Explain Results in Nursing Terms
Chapter 4 should explain what the association means for the nursing topic, not only list the coefficient and p-value.
How to Know Whether Spearman Correlation Is Right for Your Nursing Dissertation
Use this checklist before choosing Spearman correlation:
Am I examining the relationship between two variables?
Are the variables ordinal, ranked, skewed, or non-normal?
Are the variables measured on the same participants?
Do the categories have a meaningful order?
Does the relationship appear monotonic?
Have I checked for missing values?
Have I checked for impossible values or scoring errors?
Have I checked for many ties, ceiling effects, or floor effects?
Is the sample size adequate for the planned analysis?
Am I comparing groups instead?
Am I testing change over time?
Am I predicting an outcome using multiple variables?
Would Pearson correlation be better because both variables are continuous, approximately normal, and linearly related?
Can I report the result without causal language?
Table 8. Spearman Correlation Decision Guide
| Student situation | Use Spearman? | Better option | Reason |
|---|---|---|---|
| Two ordinal variables with meaningful order | Yes | Spearman correlation | Spearman fits ordinal or ranked association |
| Two skewed questionnaire scores | Often yes | Spearman or transformed/alternative analysis | Spearman may be safer than Pearson |
| Two continuous variables with linear normal pattern | Usually no | Pearson correlation | Pearson may be more appropriate |
| Two nominal categorical variables | No | Chi-square or Fisher’s exact test | Spearman requires meaningful order |
| Pretest and posttest scores | No | Paired t-test or Wilcoxon signed-rank | The question tests change over time |
| Two or more groups compared on a score | No | t-test, ANOVA, Mann–Whitney U, or Kruskal–Wallis | The question compares groups |
| One outcome predicted from several variables | No | Regression | The question involves prediction or covariate adjustment |
| Relationship changes direction | Usually no | Consider another model or visual analysis | Spearman summarizes monotonic association |
| Many responses tied at one category | Use cautiously | Consider descriptive analysis or revised interpretation | Limited variation can weaken interpretation |
Students who are unsure about test selection should get help before writing Chapter 4. Choosing the wrong test can lead to incorrect interpretation, weak methodology, and avoidable committee feedback.
Getting Help With Spearman Correlation Analysis in Nursing Research
Spearman correlation analysis may look simple, but it affects several parts of a nursing dissertation or capstone project. It affects Chapter 3 methodology, data cleaning, questionnaire scoring, ordinal-data decisions, Likert-scale analysis, SPSS output, Excel analysis, APA reporting, Chapter 4 tables, and Chapter 5 interpretation.
Nursing Dissertation Help can support students with Spearman correlation analysis, Pearson correlation analysis, SPSS correlation output, Excel correlation analysis, assumption and data-condition checking, Likert scale analysis, ordinal data analysis, sample size and power guidance, confidence interval interpretation, missing-data handling, questionnaire scoring, correlation matrix interpretation, APA 7th reporting, Chapter 4 results writing, and dissertation statistics consultation.
If you already have a dataset, questionnaire responses, SPSS output, Excel file, or Chapter 4 draft, expert support can help you choose the correct correlation test, justify it in Chapter 3, interpret the output accurately, and report the findings in clear APA 7th format.
Conclusion
Spearman correlation analysis in nursing research is useful when students need to examine relationships involving ordinal, ranked, skewed, Likert-scale, or non-normal data. It helps students evaluate whether two variables move together in a consistent monotonic pattern, especially when Pearson correlation assumptions are not met.
However, Spearman correlation must be used responsibly. Students should justify why Spearman was selected, confirm that variables are ordered or rankable, check for monotonic patterns, review ties and restricted response ranges, handle missing data correctly, interpret rho and p-values carefully, avoid causal language, and report results clearly in APA 7th edition.
If you are unsure whether Spearman correlation is the right test, or if you need help with SPSS, Excel, Likert-scale data, ordinal-data analysis, APA reporting, confidence intervals, or Chapter 4 results writing, request expert help from Nursing Dissertation Help. Expert support can help you avoid wrong-test errors, weak Chapter 3 justification, misreported output, and unclear Chapter 4 interpretation.
Frequently Asked Questions About Spearman Correlation Analysis in Nursing Research
What is Spearman correlation in nursing research?
Spearman correlation is a nonparametric correlation test used to examine the strength and direction of a monotonic relationship between two ordinal, ranked, skewed, non-normal, or rankable variables.
When should I use Spearman correlation?
Use Spearman correlation when your research question asks about a relationship between two variables and the data are ordinal, ranked, skewed, non-normal, or better analyzed using ranks.
What type of variables are needed for Spearman correlation?
Spearman correlation can be used with ordinal variables, ranked variables, Likert-type items, skewed scale scores, non-normal continuous variables, symptom ratings, satisfaction ratings, and other ordered variables.
Is Spearman correlation good for Likert scale data?
Spearman correlation is often appropriate for single Likert-type items or ordinal ratings. For summed or averaged Likert scale scores, the decision depends on the instrument, distribution, sample size, assumptions, and committee expectations.
What is the difference between Spearman and Pearson correlation?
Spearman correlation measures monotonic rank-order association and is useful for ordinal, ranked, skewed, or non-normal data. Pearson correlation measures linear association between continuous variables and works best when Pearson assumptions are met.
Can Spearman correlation prove cause and effect?
No. Spearman correlation can show association, but it cannot prove cause and effect. Students should avoid causal wording unless the research design supports causal inference.
How do I report Spearman correlation in APA 7th edition?
Report the test, variables, Spearman’s rho, p-value, sample size where helpful, confidence interval if available, direction, strength, and noncausal interpretation.
Can I run Spearman correlation in Excel?
Yes, but Excel does not provide Spearman correlation as directly as SPSS. Students usually need to rank variables first and then correlate the ranks. Excel may not provide APA-ready interpretation, confidence intervals, or complete dissertation-level support.
What should I do if my data are not normally distributed?
If your data are non-normal, skewed, ordinal, or affected by outliers, Spearman correlation may be more appropriate than Pearson correlation. You should also inspect the relationship pattern and confirm that the variables are meaningfully ordered.
What does a confidence interval for Spearman’s rho mean?
A confidence interval gives a range of plausible values for the population monotonic association. A narrow interval suggests more precision, while a wide interval suggests more uncertainty.
References
Akoglu, H. (2018). User’s guide to correlation coefficients. Turkish Journal of Emergency Medicine, 18(3), 91–93. https://doi.org/10.1016/j.tjem.2018.08.001
American Psychological Association. (n.d.-a). Sample tables. APA Style.
American Psychological Association. (n.d.-b). Table setup. APA Style.
American Psychological Association. (2024). Number and statistics guide. APA Style.
Bonett, D. G., & Wright, T. A. (2000). Sample size requirements for estimating Pearson, Kendall and Spearman correlations. Psychometrika, 65(1), 23–28. https://doi.org/10.1007/BF02294183
Brydges, C. R. (2019). Effect size guidelines, sample size calculations, and statistical power in gerontology. Innovation in Aging, 3(4), igz036. https://doi.org/10.1093/geroni/igz036
IBM. (n.d.-a). Bivariate correlations. IBM SPSS Statistics Documentation.
Jamieson, S. (2004). Likert scales: How to (ab)use them. Medical Education, 38(12), 1217–1218. https://doi.org/10.1111/j.1365-2929.2004.02012.x
Microsoft. (n.d.-a). RANK.EQ function. Microsoft Support.
Microsoft. (n.d.-b). CORREL function. Microsoft Support.
Mukaka, M. M. (2012). Statistics corner: A guide to appropriate use of correlation coefficient in medical research. Malawi Medical Journal, 24(3), 69–71.
Schober, P., Boer, C., & Schwarte, L. A. (2018). Correlation coefficients: Appropriate use and interpretation. Anesthesia & Analgesia, 126(5), 1763–1768. https://doi.org/10.1213/ANE.0000000000002864
STROBE. (n.d.). Strengthening the Reporting of Observational Studies in Epidemiology.
Sullivan, G. M., & Artino, A. R., Jr. (2013). Analyzing and interpreting data from Likert-type scales. Journal of Graduate Medical Education, 5(4), 541–542. https://doi.org/10.4300/JGME-5-4-18
von Elm, E., Altman, D. G., Egger, M., Pocock, S. J., Gøtzsche, P. C., & Vandenbroucke, J. P. (2007). The Strengthening the Reporting of Observational Studies in Epidemiology Statement: Guidelines for reporting observational studies. Annals of Internal Medicine, 147(8), 573–577. https://doi.org/10.7326/0003-4819-147-8-200710160-00010
Wasserstein, R. L., & Lazar, N. A. (2016). The ASA statement on p-values: Context, process, and purpose. The American Statistician, 70(2), 129–133. https://doi.org/10.1080/00031305.2016.1154108
Wisniewski, S. J., & Brannan, G. D. (2024). Correlation coefficient, partial, and Spearman rank, and regression analysis. In StatPearls. StatPearls Publishing.
