# Eric Heidel, PhD PStat - Statistician For Hire

> Eric Heidel, PhD PStat is an Accredited Professional Statistician that you can hire to perform statistics for your research study.

## Pages
- [A Data Dictionary Contains Information on Variables Entered Into a Database](https://www.scalestatistics.com/data-dictionary): A data dictionary contains important information about variables and data entered into a research database. Download a free data dictionary for your database.
- [A One-Sided Hypothesis States A Directional Treatment Effect](https://www.scalestatistics.com/one-sided-hypothesis): A one-sided hypothesis is considered when researchers believe that treatment effects will go in a certain direction as a result of treatment.
- [A Pilot Study is Conducted to Establish an Effect Size](https://www.scalestatistics.com/pilot-study): A pilot study is conducted to establish an effect size for sample size calculations, as well as for hypothesis generation and studying rare outcomes.
- [A Treatment Effect can be Efficacious or Detrimental](https://www.scalestatistics.com/treatment-effect): Treatment effects can be either efficacious or detrimental. The treatment effect determines the use of absolute risk reduction or absolute risk increase.
- [A Two-Sided Hypothesis States The Treatment Effect is Not Directional](https://www.scalestatistics.com/two-sided-hypothesis): A two-sided hypothesis is considered when researchers believe that treatment effects can go in either direction. This hypothesis is considered more rigorous.
- [A Type I Error Occurs When The Null Hypothesis is Rejected When It Should Not be Rejected](https://www.scalestatistics.com/type-i-error): A type I error occurs when the null hypothesis is rejected, when it should not be rejected. A type I error is also known a false positive in hypothesis testing.
- [A Type II Error Occurs When The Null Hypothesis Is Not Rejected and It Should be Rejected](https://www.scalestatistics.com/type-ii-error): A type II error is when the null hypothesis is not rejected, when it should be rejected. A type II error is also known as a false negative in hypothesis testing.
- [A Type III Error Occurs When Statistics Are Used to Answer The Wrong Question](https://www.scalestatistics.com/type-iii-error): A Type III error occurs when a statistical test is used to answer the wrong research question. Type III errors often occur as a result of miscommunication.
- [Absolute Risk Increase is The Magnitude of a Detrimental Treatment Effect](https://www.scalestatistics.com/absolute-risk-increase): Absolute Risk Increase (ARI) is used for detrimental treatments and is the absolute difference between the experimental event rate and the control event rate.
- [Absolute Risk Reduction is The Magnitude of an Efficacious Treatment Effect](https://www.scalestatistics.com/absolute-risk-reduction): Absolute Risk Reduction (ARR) is used for efficacious treatments and is the absolute difference between the control event rate and experimental event rate.
- [Access Decision Trees for Research, Statistics, and Psychometrics](https://www.scalestatistics.com/research): The Research page provides access to decision trees for research, statistics, evidence-based medicine, databases, surveys, and psychometrics.
- [Accuracy in Measurement Means The Utility And Generalizability of Outcomes](https://www.scalestatistics.com/accuracy-in-measurement): Accuracy in measurement relates to the validity, utility, interpretability, and generalizability of predictor, confounding, and outcome variables in research.
- [Acquiring Clinical Evidence with Search Queries to Access High Level Evidence](https://www.scalestatistics.com/acquiring-clinical-evidence): When acquiring clinical evidence, use the highest level of clinical evidence available with current resources. There are six levels of clinical evidence.
- [Adjusting for Multiple Comparisons: Bonferroni, Tukey's HSD, and Scheffe's Test](https://www.scalestatistics.com/adjusting-for-multiple-comparisons): Adjusting for multiple comparisons in applied statistics is done to reduce experimentwise error rates. Bonferroni, Tukey’s HSD, and Scheffe’s test can be used.
- [All Forms of Validity Evidence Fall Under Construct Validity](https://www.scalestatistics.com/construct-validity): Construct validity is validity evidence that shows a construct is being measured for in a precise and accurate fashion within the population of interest.
- [Alternate/Parallel Forms Reliability is Not Feasible](https://www.scalestatistics.com/alternate-parallel-forms-reliability): Alternate/parallel forms reliability is a form of reliability measurement that is not feasible for most researchers. Two forms of a survey are correlated.
- [Applying Clinical Decision Analysis Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-clinical-decision-analysis-evidence): When applying clinical decision analysis evidence, look for probabilities that apply to the current patient. The patient should be able to state their utilities.
- [Applying Clinical Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-clinical-evidence): Choose the type of clinical evidence that is being applied into clinical practice. Applying clinical evidence is a critical decision made when treating patients.
- [Applying Clinical Practice Guideline Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-clinical-practice-guideline-evidence): When applying clinical practice guideline evidence, consider the current burden of illness, costs, beliefs, values, barriers, and probabilities for patients.
- [Applying Diagnosis Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-diagnosis-evidence): When applying diagnosis evidence, look for the accuracy and probability of diagnosis, the costs and consequences for the test, and willingness of patients.
- [Applying Economic Analysis Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-economic-analysis-evidence): When applying economic analysis evidence, clinicians look for the costs, feasibility, and effectiveness of treatment in the current clinical context.
- [Applying Harm Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-harm-evidence): When applying harm evidence, assess similarities between the patient and the sample, integrate patient expectations, and check for all alternative treatments.
- [Applying Prognosis Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-prognosis-evidence): When applying prognosis evidence, look for similarities between the sample and your patient as well as the accuracy of prognosis in the current clinical context.
- [Applying Qualitative Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-qualitative-evidence): When applying qualitative evidence, look for feasible phenomena that apply to the current clinical contexts and patient harms, benefits, values, and expectations.
- [Applying Randomized Controlled Trial Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-randomized-controlled-trial-evidence): When applying randomized controlled trial evidence, look for patient similarities, feasible treatments, benefits and harms, values and expectations, and variables.
- [Applying Screening Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-screening-evidence): When applying screening evidence, assess the benefits and harms of early diagnosis of disease states and treatment effects on the patient population.
- [Applying Systematic Review Evidence Into Clinical Practice](https://www.scalestatistics.com/applying-systematic-review-evidence): When applying systematic review evidence, clinicians look for patient similarities, feasible treatments, and the potential benefits and harms from treatment.
- [Appraising Clinical Decision Analysis Evidence for Therapeutic Alternatives, Probabilities, And Util...](https://www.scalestatistics.com/appraising-clinical-decision-analysis-evidence): When appraising clinical decision analysis evidence, therapeutic alternatives and outcomes should be presented along with valid probabilities and utilities.
- [Appraising Clinical Evidence for Therapy, Diagnosis, Screening, Prognosis, and Harm](https://www.scalestatistics.com/appraising-clinical-evidence): Choose the type of clinical evidence you are appraising: Therapy, diagnosis, screening, prognosis, or harm. Appraise it for quality, interpretability, and utility.
- [Appraising Clinical Practice Guideline Evidence for Timely and Specific Evidence](https://www.scalestatistics.com/appraising-clinical-practice-guideline-evidence): When appraising clinical practice guideline evidence, look for a comprehensive and reproducible literature review in the past 12 months and relevant citations.
- [Appraising Diagnosis Evidence for Youden Index, Sensitivity, and Specificity](https://www.scalestatistics.com/appraising-diagnosis-evidence): When appraising diagnosis evidence, look for patient similarities, comparison to the gold standard, Youden index, and high sensitivity and specificity.
- [Appraising Economic Analysis Evidence for Treatments, Costs, and Consequences](https://www.scalestatistics.com/appraising-economic-analysis-evidence): When appraising economic analysis evidence, there should be valid courses of action, a specified viewpoint, evidence citations, and accurate measures.
- [Appraising Harm Evidence for Patient Similarities, Methods, Follow-Up, and Precision](https://www.scalestatistics.com/appraising-harm-evidence): When appraising harm evidence, there should be similar groups, similar treatments, sufficient follow-up, dose-response gradient, and consistent findings.
- [Appraising Prognosis Evidence for Patient Similarities, Follow-Up, Temporality, and Precision](https://www.scalestatistics.com/appraising-prognosis-evidence): When appraising prognosis evidence, there should be a defined sample of patients, sufficient follow-up, adjustment for variables, and validation of findings.
- [Appraising Qualitative Evidence Related to Patient Experiences](https://www.scalestatistics.com/appraising-qualitative-evidence): When appraising qualitative evidence, the selection of participants and methods for data collection must be described and the results must be impressive.
- [Appraising Randomized Controlled Trial Evidence for Empirical Benchmarks](https://www.scalestatistics.com/appraising-randomized-controlled-trial-evidence): When appraising random controlled trial evidence, there are certain benchmarks that have to be met: Random selection, random assignment, and intention-to-treat.
- [Appraising Screening Evidence for Prognosis and Patient Treatment](https://www.scalestatistics.com/appraising-screening-evidence): When appraising screening evidence, evidence should show early diagnosis improves survival and quality of life. Patients should be willing partners in treatment.
- [Appraising Systematic Review Evidence For Rigor, Publication Bias, and Pooled Effects](https://www.scalestatistics.com/appraising-systematic-review-evidence): When appraising systematic review evidence, there should only be RCT-level evidence included. Interpret pooled effects, heterogeneity, and the funnel plot.
- [Appraising Therapy Evidence for Several Types of Evidence](https://www.scalestatistics.com/appraising-therapy-evidence): Choose a type of therapy evidence to appraise: RCT, systematic review, qualitative, clinical decision analysis, economic analysis, or clinical practice guideline.
- [Asking Clinical Questions with The PICO Framework](https://www.scalestatistics.com/asking-clinical-questions): When asking clinical questions, use the PICO framework to ask foreground questions. Clinical questions often come from patient care and medical treatment.
- [Assess Evidence-Based Clinical Practice](https://www.scalestatistics.com/assessing-clinical-practice): When assessing clinical practice, look for barriers and facilitators for change, strategies for implementing change, checks of performance, and sustainability.
- [Assess Homogeneity of Variance When Using ANOVA in SPSS](https://www.scalestatistics.com/homogeneity-of-variance-and-anova): The assumption of homogeneity of variance is assessed when conducting ANOVA. SPSS can be used to conduct the Levene’s Test for Equality of Variances.
- [Assess Homogeneity of Variance When Using Independent Samples t-test in SPSS](https://www.scalestatistics.com/homogeneity-of-variance-and-independent-samples-t-test): The assumption of homogeneity of variance must be met to conduct independent samples t-test. SPSS can be used to conduct Levene's Test of Equality of Variances.
- [Assess Normality When Using ANOVA in SPSS](https://www.scalestatistics.com/normality-and-anova): The assumption of normality is assessed when conducting ANOVA. Normality is assessed using skewness and kurtosis statistics in SPSS. Values should be below 2.0.
- [Assess Normality When Using Independent Samples t-test in SPSS](https://www.scalestatistics.com/normality-and-independent-samples-t-test): The assumption of normality is assessed when conducting independent samples t-test. Normality is assessed using skewness and kurtosis statistics in SPSS.
- [Assess Normality When Using Repeated-Measures ANOVA in SPSS](https://www.scalestatistics.com/normality-and-repeated-measures-anova): The assumption of normality is assessed when conducting repeated-measures ANOVA. Normality is assessed using skewness and kurtosis statistics in SPSS.
- [Assess Normality When Using Repeated-Measures t-test in SPSS](https://www.scalestatistics.com/normality-and-repeated-measures-t-test): The assumption of normality is assessed when conducting repeated-measures t-test. Normality is assessed using skewness and kurtosis statistics in SPSS.
- [Assess Sphericity When Using Repeated-Measures ANOVA in SPSS](https://www.scalestatistics.com/sphericity-and-repeated-measures-anova): The assumption of sphericity is assessed when conducting repeated-measures ANOVA. Sphericity for repeated-measures ANOVA is tested using Mauchly’s test in SPSS.
- [Assess The Acquiring of Clinical Evidence in Clinical Practice](https://www.scalestatistics.com/acquiring-clinical-evidence-assessment): To conduct an assessment of acquiring clinical evidence, look for the best sources of evidence, easy access to sources, and more efficient searching.
- [Assess The Applying of Clinical Evidence in Clinical Practice](https://www.scalestatistics.com/applying-clinical-evidence-assessment): To conduct an assessment of applying clinical evidence, look for integration after appraisal, accurate and efficient adjusting of care, and management decisions.
- [Assess The Appraising of Clinical Evidence in Clinical Practice](https://www.scalestatistics.com/appraising-clinical-evidence-assessment): To conduct an assessment of appraising clinical evidence, look for ease of appraisal, becoming more accurate and efficient in applying, and keeping a summary.
- [Assess The Asking of Clinical Questions in Clinical Practice](https://www.scalestatistics.com/asking-clinical-questions-assessment): To conduct an assessment of asking clinical questions, look for asking any questions, focused questions, filling knowledge gaps, and using good resources.
- [Assessing Evidence-Based Practice Principles In Clinical Practice](https://www.scalestatistics.com/assessing-evidence-based-practice): Assessing evidence-based practices is a form of meta-assessment where clinicians audit evidence-based interventions and their respective clinical outcomes.
- [Bayes Theorem is Used to Calculate The Probability of an Outcome](https://www.scalestatistics.com/bayes-theorem): Bayes' Theorem allows for clinicians to calculate the probability of an outcome based on pretest probabilities and the effectiveness of diagnostic tests.
- [Between-Subjects Designs Decrease Statistical Power and Increase The Needed Sample Size](https://www.scalestatistics.com/statistical-power-and-between-subjects-designs): Between-subjects designs decrease statistical power and increase the needed sample size. This is due to needing more people to compare independent groups.
- [Between-Subjects Statistics Are Used to Compare Independent Groups on Outcomes](https://www.scalestatistics.com/between-subjects-statistics): Between-subjects statistics are used to compare one, two, or three or more independent groups on categorical, ordinal, and continuous outcome variables.
- [Between-Subjects Statistics for One Group: Categorical, Ordinal, and Continuous Outcomes](https://www.scalestatistics.com/between-subjects-statistics-for-one-group): There are three between-subjects statistical tests for one group. Choose from Chi-square Goodness-of-fit, one-sample medial test, and one-sample t-test.
- [Between-Subjects Statistics for Three or More Groups: Categorical, Ordinal, and Continuous Outcomes](https://www.scalestatistics.com/between-subjects-statistics-for-three-or-more-groups): There are three between-subjects statistical tests for three or more groups. These include unadjusted odds ratios, Kruskal-Wallis, and one-way ANOVA.
- [Between-Subjects Statistics for Two Groups: Categorical, Ordinal, and Continuous Outcomes](https://www.scalestatistics.com/between-subjects-statistics-for-two-groups): There are four between-subjects statistical tests for two groups. These include Fisher's Exact Test, chi-square, Mann-Whitney U, and independent samples t-test.
- [Blinding in Randomized Controlled Trials Reduces Bias](https://www.scalestatistics.com/blinding): Blinding is used in randomized controlled trials to reduce observational biases associated with study participants, researchers, staff, and clinicians.
- [Blocked Randomization - Experimental Research Designs and Randomized Controlled Trials](https://www.scalestatistics.com/blocked-randomization): Blocked randomization is a method of random assignment in experimental research designs and randomized controlled trials where study participants are randomized in small blocks of four or six.
- [Calculate Odds Ratio with 95% Confidence Intervals](https://www.scalestatistics.com/odds-ratio): The odds ratio with 95% confidence interval is used for chi-square and case-control designs. The width of the confidence interval is the primary inference.
- [Calculate Relative Risk with 95% Confidence Intervals](https://www.scalestatistics.com/relative-risk): Relative risk with 95% confidence intervals is calculated in cohort and experimental studies. The width of the confidence interval is the primary inference.
- [Calculate The Sample Size for ANOVA](https://www.scalestatistics.com/sample-size-for-anova): The steps for calculating the sample size for ANOVA in G*Power are presented. The effect size is the difference in means and standard deviations between groups.
- [Calculate The Sample Size for Chi-Square](https://www.scalestatistics.com/sample-size-for-chi-square): The steps for calculating the sample size for a Chi-square test in G*Power are presented. The effect size is the difference in proportions between treatment groups.
- [Calculate The Sample Size for Independent Samples t-test](https://www.scalestatistics.com/sample-size-for-independent-samples-t-test): The steps for calculating the sample size for an independent samples t-test in G*Power are presented. The effect size is the difference in means between groups.
- [Calculate The Sample Size for Mann-Whitney U](https://www.scalestatistics.com/sample-size-for-mann-whitney-u): The steps for calculating the sample size for a Mann-Whitney U test in G*Power are presented. The effect size is the difference in means between groups.
- [Calculate The Sample Size for McNemar's Test](https://www.scalestatistics.com/sample-size-for-mcnemars-test): The steps for calculating the sample size for McNemar’s Test in G*Power are presented. The effect size is the odds ratio for change in the outcome variable.
- [Calculate The Sample Size for Pearson's r](https://www.scalestatistics.com/sample-size-for-pearsons-r): The steps for calculating the sample size for a Pearson’s r correlation in G*Power are presented. The effect size is the hypothesized coefficient of determination.
- [Calculate The Sample Size for Point Biserial](https://www.scalestatistics.com/sample-size-for-point-biserial): The steps for calculating the sample size for a point biserial in G*Power are presented. The effect size is the hypothesized coefficient of determination.
- [Calculate The Sample Size for Repeated-Measures t-test](https://www.scalestatistics.com/sample-size-for-repeated-measures-t-test): The steps for calculating the sample size for a repeated-measures t-test in G*Power are presented. The effect size is the difference in observation means.
- [Calculate The Sample Size for Wilcoxon Test](https://www.scalestatistics.com/sample-size-for-wilcoxon-test): The steps for calculating the sample size for a Wilcoxon test in G*Power are presented. The effect size is the difference in means between observations.
- [Calculate The Sample Size Needed for a Research Study](https://www.scalestatistics.com/sample-size): Calculate the needed sample size for ten different statistical tests using G*Power. Choose inclusion criteria, exclusion criteria, and sampling methods.
- [Calculate The Sample Size ror Repeated-Measures ANOVA](https://www.scalestatistics.com/sample-size-for-repeated-measures-anova): The steps for calculating the sample size for a repeated-measures ANOVA in G*Power are presented. The effect size is the hypothesized partial eta-squared.
- [Case Series Designs Are Used to Generate Hypotheses and Establish Effect Sizes](https://www.scalestatistics.com/case-series): The case series design is a type of observational research design used to generate hypotheses and establish effect sizes for future research studies.
- [Case-Control Designs Are Used to Study Rare Outcomes and Generate Hypotheses](https://www.scalestatistics.com/case-control): The case-control design is an observational research design that is used to generate hypotheses and study rare outcomes. Choosing controls is important.
- [Categorical Outcomes Decrease Statistical Power and Increase The Needed Sample Size](https://www.scalestatistics.com/statistical-power-and-categorical-outcomes): Categorical outcomes decrease statistical power and increase the needed sample size. This is because categorical outcomes lack precision and accuracy.
- [Categorical Variables Are Used to Quantify Group Membership and Events](https://www.scalestatistics.com/categorical-variables): Categorical variables are used to categorize or quantify phenomena, groups, and/or events into numerical categories that can be measured and analyzed.
- [Choose The Correct Statistics to Answer Research Questions](https://www.scalestatistics.com/statistics): The Statistics decision tree will help you choose the correct statistical test based on your research question and the meeting of statistical assumptions.
- [Choose The Correct Variables to Answer a Research Question](https://www.scalestatistics.com/variables): Choose the variables that will answer your research question. There are different types of variables and several scales of measurement for variables.
- [Clustered Random Sampling is Used to Randomly Sample from Naturally Occurring Groups or Areas](https://www.scalestatistics.com/clustered-random-sampling): Clustered random sampling is a probability sampling method where natural clusters in a population are targeted for representation using random selection.
- [Concurrent Validity Evidence Shows Correlations With Other Measures](https://www.scalestatistics.com/concurrent-validity): Concurrent validity is validity evidence that shows a survey instrument correlates with validated measures. Correlations are used for concurrent validity.
- [Conduct a Pilot Study for The Survey](https://www.scalestatistics.com/survey-pilot-study): The pilot study of a survey means administering it to a sample of between 150-300 participants from the population of interest and conducting psychometrics.
- [Conduct a Survey Pretest Using a Focus Group](https://www.scalestatistics.com/survey-pretest): Conducting a survey pretest means administering a survey to 5-10 people and running a focus group. Then, changes and revisions are made to the survey.
- [Conduct a Validation Study for The Survey](https://www.scalestatistics.com/survey-validation-study): The validation study of a survey means administering it to a sample of 300-1,000 participants and then testing for validity evidence. CFA is also performed
- [Conduct and Interpret Over 50 Statistical Tests in SPSS](https://www.scalestatistics.com/spss): Access the methods for conducting and interpreting over 50 different statistical tests in SPSS. Choose which statistical test you want to run in SPSS.
- [Confirmatory Factor Analysis (CFA) Validates Theoretical Frameworks](https://www.scalestatistics.com/confirmatory-factor-analysis): Confirmatory factor analysis is a statistical technique used to confirm or validate the internal structure of a given survey instrument or construct.
- [Confounding Variables Adjust The Relationships Between Variables](https://www.scalestatistics.com/confounding-variables): Confounding variables adjust the association between predictor and outcome variables. Confounding variables are taken into account in multivariate models.
- [Content Validity is How Well a Construct Represents The Empirical Literature](https://www.scalestatistics.com/content-validity): Content validity is validity evidence that shows the content in the existing empirical literature is being properly represented in a survey instrument.
- [Continuous Outcomes Increase Statistical Power and Decrease The Needed Sample Size](https://www.scalestatistics.com/statistical-power-and-continuous-outcomes): Continuous outcomes increase statistical power and decrease the needed sample size. This is because continuous outcomes possess the most precision and accuracy.
- [Continuous Variables Lead to Precision and Accuracy in Measurement](https://www.scalestatistics.com/continuous-variables): Continuous variables are actual numerical values that allow for measures of distance and magnitude. Continuous variables provide more precision and accuracy.
- [Control Event Rate Is The Proportion Of Control Participants That Have An Outcoma](https://www.scalestatistics.com/control-event-rate): Control event rate (CER) is an epidemiological calculation used with experimental event rate to calculate absolute risk reduction or absolute risk increase.
- [Control Variables Are Used in Multivariate Models](https://www.scalestatistics.com/control-variables): Control variables are entered into multivariate models to account for how they may change the association between independent and dependent variables.
- [Convenience Sampling means Sampling From Populations That Are Accessible and Available](https://www.scalestatistics.com/convenience-sampling): Convenience sampling is a non-probability sampling method where researchers recruit study participants into a study that are convenient to find and analyze.
- [Convergent Validity is The Ability of a Measure to Correlated With Similar Measures](https://www.scalestatistics.com/convergent-validity): Convergent validity is validity evidence that shows a survey instrument correlates with similar constructs. Correlations are used for convergent validity.
- [Count Variables Have Skewed or Kurtotic Distributions](https://www.scalestatistics.com/count-variables): Count variables are naturally skewed or kurtotic distributions. Poisson regression and negative binomial regression are used to analyze count variables.
- [Counterbalanced Designs Can Test Multiple Inverventions](https://www.scalestatistics.com/counterbalanced-design): The counterbalanced design is a type of quasi-experimental design. Counterbalanced designs allow for the testing of multiple interventions concurrently.
- [Create Reliable and Valid Surveys in Eight Steps](https://www.scalestatistics.com/surveys): The Surveys decision tree will help you work through the eight steps of creating a survey instrument. Reliable and valid surveys measure for human constructs.
- [Cross-Sectional Designs Are Used to Establish Prevalence](https://www.scalestatistics.com/cross-sectional): The cross-sectional design is an observational research design that can establish the prevalence of an outcome in a population. Surveys are cross-sectional.
- [Cross-Tabulation Statistics Describe Relationships Between Categorical Variables](https://www.scalestatistics.com/cross-tabulation): Cross-tabulation tables are used to describe the association between categorical variables. Cross-tabulation tables are used with epidemiology and diagnosis.
- [Crossover Randomized Design - Experimental Research Designs and Randomized Controlled Trials](https://www.scalestatistics.com/crossover-randomized-design): Crossover randomized designs are experimental research designs where treatment groups switch interventions after a specified period of time and washout.
- [Databases Are Where Research Data is Stored and Manipulated](https://www.scalestatistics.com/databases): Databases are where researchers store and manipulate research data. Download free databases for between-subjects, within-subjects, and multivariate statistics.
- [Demographic Variables Describe Samples Taken From Populations](https://www.scalestatistics.com/demographic-variables): Demographic variables are used to describe the nature of samples derived from populations. Age, gender, and ethnicity are popular demographic variables.
- [Dependent Variables Are What is Being Measured in a Research Study](https://www.scalestatistics.com/dependent-variables): Dependent variables are what researchers measure for in a research study. Dependent variables are the measurements, variables, or outcomes being studied.
- [Descriptive Statistics Are Used to Describe Samples and Inferential Statistics Findings](https://www.scalestatistics.com/descriptive-statistics): Choose descriptive statistics to use in your research study. Descriptive statistics describe samples and give context to inferential statistical analyses.
- [Diagnostic Accuracy is The Proportion of Correct Diagnoses Found by a Diagnostic Test](https://www.scalestatistics.com/diagnostic-accuracy): Diagnostic accuracy is the number of times that a diagnostic test correctly detects disease or healthy people. These are the “correct” diagnostic findings.
- [Education](https://www.scalestatistics.com/education): Use the education page to access information for writing instructional goals and objectives, Maslow’s Hierarchy of Needs, and a research/statistics dictionary.
- [Epidemiology is The Study of Diseases in Populations](https://www.scalestatistics.com/epidemiology): Epidemiology is the study of disease states in populations. Choose calculations like prevalence, incidence, number needed to treat, and number needed to harm.
- [Equivalence Trials: Are Two Treatments Equally As Good](https://www.scalestatistics.com/equivalence-trial): Equivalence trials establish the equivalency of a treatment versus a "gold standard" treatment. Equivalence is interpreted as meaning "equally as good."
- [Equivalent Time Sample Designs Use Multiple Wash-Out Periods](https://www.scalestatistics.com/equivalent-time-sample-design): The equivalent time sample design is a type of quasi-experimental design. Several "washout" periods establish treatment stability and longitudinal effects.
- [Eric Heidel, PhD PStat - Statistician For Hire](https://www.scalestatistics.com/)
- [Evidence-Based Medicine is Practiced Using Five Steps](https://www.scalestatistics.com/evidence-based-medicine): Practice evidence-based medicine using five steps: Asking questions, acquiring evidence, appraising evidence, applying evidence, and assessing practice.
- [Exclusion Criteria Are Used to Define a Population](https://www.scalestatistics.com/exclusion-criteria): Exclusion criteria help specify study participants that will not be studied due to loss to follow-up, inability to provide data, or potential adverse events.
- [Experimental Event Rate is The Proportion of Experimental Participants with an Outcome](https://www.scalestatistics.com/experimental-event-rate): Experimental event rate (EER) is an epidemiological calculation used with control event rate to calculate absolute risk reduction or absolute risk increase.
- [Experimental Research Designs Yield Causal Effects](https://www.scalestatistics.com/experimental-research-designs): Experimental research designs must have random selection and random assignment of trial participants. Experimental designs can yield causal effects.
- [Extensive Variance in an Effect Size Decreases Statistical Power and Increases The Needed Sample Siz...](https://www.scalestatistics.com/statistical-power-and-extensive-variance-of-effect-size): Extensive variance in the effect size decreases statistical power and increases the needed sample size. This is because heterogeneity is hard to detect.
- [Face Validity is Not a Real Form of Validity Evidence](https://www.scalestatistics.com/face-validity): Face validity is validity evidence that shows a survey instrument does what it is supposed to do at “face value.” It is not a true form of validity evidence.
- [FINER and ethical research questions](https://www.scalestatistics.com/ethical-research-questions): Ethical research questions are written to protect human beings and animals from unfair or dangerous research methods. Ethical is the fourth part of FINER.
- [FINER and feasible research questions](https://www.scalestatistics.com/feasible-research-questions): Feasible research questions focus on the time, scope, resources, expertise, and funding needed to conduct research. Feasible is the first part of FINER.
- [FINER and interesting research questions](https://www.scalestatistics.com/interesting-research-questions): Interesting research questions are personally and professionally rewarding, interesting to others, and worth funding. Interesting is the second part of FINER.
- [FINER and novel research questions](https://www.scalestatistics.com/novel-research-questions): Novel research questions produce new clinical evidence, fill knowledge gaps in the literature, and validate published studies. Novel is the third part of FINER.
- [FINER and relevant research questions](https://www.scalestatistics.com/relevant-research-questions): Relevant research questions impact clinical practice, improve outcomes, change protocols, and guide future research. Relevant is the last part of FINER.
- [FINER Stands for Feasible, Interesting, Novel, Ethical, and Relevant](https://www.scalestatistics.com/finer): The FINER framework stands for feasible, interesting, novel, ethical, and relevant. FINER is used to write valid and answerable research questions.
- [Frequency is The Number of Times That Something Occurs in a Categorical Distribution](https://www.scalestatistics.com/frequency): Frequency statistics are used to describe the number of times that something occurs. Frequency is used to describe categorical variables in applied statistics.
- [Get an Expert Review of the Construct Specification](https://www.scalestatistics.com/expert-review): Expert review of the construct specification by a panel of experts is important so that expert opinions, suggestions, and revisions can be integrated.
- [Hierarchical Regression is Used to Test Theory](https://www.scalestatistics.com/hierarchical-regression): Hierarchical regression is used to predict for continuous outcomes when testing a theoretical framework. Hierarchical regression can be conducted in SPSS.
- [Homogeneity of Variance Means That Independent Groups Must Have Equal Variances](https://www.scalestatistics.com/homogeneity-of-variance): The assumption of homogeneity of variance states that independent groups must have equal variances. Levene's Test of Equality of Variances is used to test it.
- [Hypothesis Testing is Used with Inferential Statistics](https://www.scalestatistics.com/hypothesis-testing): Hypothesis testing is central to conducting inferential statistics. Researchers should state the null and alternative hypotheses before conducting statistics.
- [Incidence is The Number of New Cases in a Population](https://www.scalestatistics.com/incidence): Incidence is the number of new cases in a given population. Incidence in a given population is calculated using prospective and experimental studies.
- [Inclusion Criteria Are Used to Define a Population](https://www.scalestatistics.com/inclusion-criteria): Inclusion criteria are used to choose relevant study participants that possess specific characteristics. Inclusion criteria must be explicitly defined.
- [Incremental Validity is The Ability of a Measure to Add New and Unique Variance](https://www.scalestatistics.com/incremental-validity): Incremental validity is validity evidence that shows a survey instrument can predict for new variance. Stepwise regression is used for incremental validity.
- [Independence of Observations Means Each Study Participant is Independent of All Other Observations](https://www.scalestatistics.com/independence-of-observations): The assumption of independence of observations stipulates that each observation in a sample is independent. Participants or observations must be independent.
- [Independent Variables Are Manipulated By Researchers](https://www.scalestatistics.com/independent-variables): Independent variables are manipulated by researchers to better understand their association with dependent variables. Interventions are independent variables.
- [Intention-to-Treat in Randomized Controlled Trials](https://www.scalestatistics.com/intention-to-treat): "Intention-to-Treat" analysis in randomized controlled trials (RCTs) means that all study participants are analyzed in the same group they were randomly assigned to at the beginning of a study.
- [Inter-Rater Reliability: Kappa and Intraclass Correlation Coefficient](https://www.scalestatistics.com/inter-rater-reliability): Inter-rater reliability is a form of reliability that assesses the level of agreement between raters. Use Kappa and Intraclass Correlation Coefficients in SPSS.
- [Internal Consistency Reliability Looks At Associations Between Survey Items](https://www.scalestatistics.com/internal-consistency-reliability): Internal consistency reliability assesses the inter-correlations between survey items. Use and interpret Cronbach’s alpha, Split-half, and KR-20 in SPSS.
- [Interquartile Range is Used to Describe Variance in Skewed Distributions](https://www.scalestatistics.com/interquartile-range): Interquartile range is a statistical measure of dispersal used in skewed distributions of continuous values. Interquartile range is often used with medians.
- [Interrupted Time Series Designs Can Establish Longitudinal Treatment Effects](https://www.scalestatistics.com/interrupted-time-series-design): An interrupted time series design is a type of quasi-experimental design. Study participants are observed several times before and after treatment.
- [Interval Variables Provide Measures of Distance](https://www.scalestatistics.com/interval-variables): Interval variables do not possess a "true zero" but can yield measures of distance. There are equal distances between points or values with interval variables.
- [Known-Groups Validity is The Ability of a Measure to Differentiate Between Independent Groups](https://www.scalestatistics.com/known-groups-validity): Known-groups validity is validity evidence that shows a survey instrument can detect differences between independent groups using between-subjects statistics.
- [Large Effect Size Increase Statistical Power and Decrease The Needed Sample Size](https://www.scalestatistics.com/statistical-power-and-large-effect-sizes): Large effect sizes increase statistical power and decrease the needed sample size. This is because large effect sizes are easier to detect with statistics.
- [Large Sample Sizes Increase Statistical Power and Increase The Flexibility Of The Effect Size](https://www.scalestatistics.com/statistical-power-and-large-sample-sizes): Large sample sizes increase statistical power and increase the flexibility of detecting effect sizes. Large sample sizes can detect treatment effects.
- [Levene's Test is Used to Test for The Assumption of Homogeneity of Variance](https://www.scalestatistics.com/levenes-test): Levene's Test of Equality of Variances is used to assess the assumption of homogeneity of variance with independent samples t-tests and one-way ANOVA.
- [Limited Variance of Effect Size Increases Statistical Power and Decreases The Needed Sample Size](https://www.scalestatistics.com/statistical-power-and-limited-variance-of-effect-size): Limited variance in the effect size increases statistical power and decreases the needed sample size. This is because homogeneity is easier to detect.
- [Linearity Means Variables Must Have a Linear Relationship with The Outcome](https://www.scalestatistics.com/linearity): The statistical assumption of linearity assumes that there is a linear relationship between continuous predictor variables and an outcome in regression models.
- [Logarithmic Transformations Can Normalize Skewed or Kurtotic Data](https://www.scalestatistics.com/logarithmic-transformations): Logarithmic transformations allow for analysis of non-normal distributions using parametric statistics. Logarithmic transformations normalize distributions.
- [Maslow's Hierarchy of Needs](https://www.scalestatistics.com/maslows-hierarchy-of-needs): Maslow's Hierarchy of Needs is a theoretical framework for understanding the development and achievement of goals and self-actualization in one's life.
- [Mauchly's Test is Used to Test The Assumption of Sphericity](https://www.scalestatistics.com/mauchlys-test): Mauchly's test is used to assess the assumption of sphericity with repeated-measures ANOVA. Greenhouse-Geisser corrections are used with sphericity violations.
- [Measures of Central Tendency Include The Mean, Median, and Mode](https://www.scalestatistics.com/measures-of-central-tendency): Measures of central tendency are used to describe normal distributions of continuous variables. Mean, median, and mode are measures of central tendency.
- [Measures of Variability Include Variance, Standard Deviation, and Interquartile Range](https://www.scalestatistics.com/measures-of-variability): Measures of variability are used to describe the dispersal of a continuous distribution. These include variance, standard deviation, and interquartile range.
- [Multicollinearity is Where Two Predictor Variables Are Correlated in Regression](https://www.scalestatistics.com/multicollinearity): Multicollinearity occurs in regression analysis when predictor variables are highly correlated and artificially inflate or deflate multivariate relationships.
- [Multivariate Designs Decrease Statistical Power and Increase The Needed Sample Size](https://www.scalestatistics.com/statistical-power-and-multivariate-designs): Multivariate designs decrease statistical power and increase the needed sample size. This is because more observations to detect multivariate effects.
- [Negative Predictive Value and Believable Negative Test Results](https://www.scalestatistics.com/negative-predictive-value): Negative Predictive Value (NPV) is a measure of how believable a negative test result is for a diagnostic test in a given healthy patient population.
- [Nested Case-Control Designs Are Part of Prospective Cohort Studies](https://www.scalestatistics.com/nested-case-control): Nested case-control designs are embedded in prospective cohort designs where fresh participant specimens exist and valid risk factors have been collected.
- [Nominal Variables Categorize Events, Names, and Groups](https://www.scalestatistics.com/nominal-variables): Nominal variables are used to name or categorize phenomena, groups, and/or events into numerical categories so that they can be measured and analyzed.
- [Nomograms Are Predictive Tools That Show Probabilities of an Outcome](https://www.scalestatistics.com/nomograms): Nomograms are predictive tools that use regression models to give relative context and probabilities of clinical outcomes to patients in clinical practice.
- [Non-Inferiority Trial: Are Two Treatments Just As Good](https://www.scalestatistics.com/non-inferiority-trial): Non-inferiority trials establish if a treatment is "just as good" as the "gold standard" treatment. Non-inferiority trials are different from equivalence trials.
- [Non-Parametric Statistics Are Used with Categorical and Ordinal Outcomes](https://www.scalestatistics.com/non-parametric-statistics): Non-parametric statistics are used with categorical and ordinal outcomes. Non-parametric statistics are also used when the statistical assumptions are violated.
- [Non-Probability Sampling Means Random Selection is Not Used to Select Study Participants](https://www.scalestatistics.com/non-probability-sampling): Non-probability sampling techniques are used in observational studies and they do not use random selection or random assignment to choose study participants.
- [Nonequivalent Control Group Designs Randomize at the Intervention Level](https://www.scalestatistics.com/nonequivalent-control-group-design): A nonequivalent control group design is a type of quasi-experimental design. Study participants are randomized to treatment groups at the intervention level.
- [Normal Probability Plot Assesses Normality and Homoscedasticity](https://www.scalestatistics.com/normal-probability-plot): Normal probability plots are used to assess the statistical assumptions of normality and homoscedasticity in regression models. They are also called P-P plots.
- [Normality Means That a Continuous Distribution Must be Normally Distributed](https://www.scalestatistics.com/normality): The assumption of normality states that continuous variables must be normally distributed. Skewness and kurtosis statistics are used to assess normality.
- [Normality of Difference Scores Means Differences Between Observations Must be Normally Distributed](https://www.scalestatistics.com/normality-of-difference-scores): The assumption of normality of difference scores states that the differences between independent observations of an outcome must be normally distributed.
- [Number Needed to Harm is The Number of People That Have to be Treated to Cause a Bad Outcome](https://www.scalestatistics.com/number-needed-to-harm): Number Needed to Harm (NNH) is the number of people that have to be treated to cause a future bad outcome. It is associated with absolute risk increase.
- [Number Needed to Treat is The Number of People That Have to be Treated to Prevent a Bad Outcome](https://www.scalestatistics.com/number-needed-to-treat): Number Needed to Treat (NNT) is the number of people that have to be treated to prevent a future bad outcome. It is associated with absolute risk reduction.
- [Observational Research Designs Do Not Use Randomization](https://www.scalestatistics.com/observational-research-designs): Choose an observational research design that answers your research question. Outcomes can be analyzed using retrospective and prospective research designs.
- [Ordinal Outcomes Decrease Statistical Power and Increase The Needed Sample Size](https://www.scalestatistics.com/statistical-power-and-ordinal-outcomes): Ordinal outcomes decrease statistical power and increase the needed sample size. This is because ordinal outcomes lack precision and accuracy.
- [Ordinal Variables Measure Things Across an ordered numerical continuum](https://www.scalestatistics.com/ordinal-variables): Ordinal variables given a measure of rank or order. Subjective rankings along a numerical continuum (Likert-type scales) are considered ordinal variables.
- [Outcome Variables Should Be Measured at The "Gold Standard"](https://www.scalestatistics.com/outcome-variables): Outcome variables are what researchers measure for in a research study. Outcomes variables should be measured at a precise and accurate "gold standard" level.
- [Page Title: Download Free Research Calculators for Sample Size, Diagnostic Testing, and Epidemiology](https://www.scalestatistics.com/calculators): The Calculators page will allow you to download several free Excel-based research calculators for sample size, diagnostic testing, and epidemiology.
- [Parallel Randomized Design - Experimental Research Designs and Randomized Controlled Trials](https://www.scalestatistics.com/parallel-randomized-design): Parallel randomized design is an experimental research design (randomized controlled trial) where participants stay in the same treatment groups for the entirety of the study.
- [Parametric Statistics Are Used with Continuous Outcomes](https://www.scalestatistics.com/parametric-statistics): Parametric statistics are used with continuous outcomes. The statistical assumptions of normality and homogeneity have to be met to run parametric statistics.
- [PICO Framework - Comparator in the Research Question](https://www.scalestatistics.com/comparator): The comparator in a research question represents the comparison or control group. Comparators or controls help establish treatment effects and are part of PICO.
- [PICO Framework - Intervention in the Research Question](https://www.scalestatistics.com/intervention): The intervention or independent variable in a research question should be associated with the outcome. The intervention is a part of the PICO framework.
- [PICO Framework - Outcome in the Research Question](https://www.scalestatistics.com/outcome): The outcome in a research question is the dependent variable that is being measured for in a research study. The outcome is a part of the PICO framework.
- [PICO Framework - Population in the Research Question](https://www.scalestatistics.com/population): The population in a research question needs to be defined in terms of both inclusion and exclusion criteria. Population is a part of PICO.
- [PICO Framework - Population, Intervention, Comparator, Outcome](https://www.scalestatistics.com/pico): The PICO framework is used to formulate and refine a research question. PICO stands for population, intervention, comparator, and outcome.
- [Positive Predictive Value and Believable Positive Test Results](https://www.scalestatistics.com/positive-predictive-value): Positive Predictive Value (PPV) is a measure of how believable a positive test result is for a diagnostic test. As prevalence increases, PPV will increase.
- [Precision in Measurement Means Reliability, Consistency, and Consistency](https://www.scalestatistics.com/precision-in-measurement): Precision in measurement relates to the reliability, stability, confidence, and consistency of predictor, confounding, and outcome variables in research.
- [Predictive Validity is The Ability of a Measure to Predict for Future Outcomes](https://www.scalestatistics.com/predictive-validity): Predictive validity is validity evidence that shows a survey instrument can predict for future occurrences. Correlations are used for predictive validity.
- [Predictor Variables Are Used to Test Associations with Outcomes](https://www.scalestatistics.com/predictor-variables): Predictor variables are variables that are hypothesized to have an association with an outcome variable in observational and experimental research designs.
- [Prevalence is The Number of Current Cases That Exist in a Given Population](https://www.scalestatistics.com/prevalence): Prevalence is the number of current cases in a given population. Prevalence in a population is established using a retrospective cross-sectional design.
- [Privacy Policy](https://www.scalestatistics.com/privacy-policy): This page will present the Scale, LLC privacy policy.
- [Probability Sampling Means Study Participants Are Chosen Using Random Selection](https://www.scalestatistics.com/probability-sampling): Probability sampling means everyone in the population has an equal chance of being selected for study participation and it is used in experimental designs.
- [Propensity Score Matching Is Used to Choose Controls](https://www.scalestatistics.com/propensity-score-matching): Propensity score matching is a method for individually matching cases to controls using an algorithm. It can conducted using logistic regression analysis.
- [Prospective Cohort Designs Provide Measures of Incidence and Relative Risk](https://www.scalestatistics.com/prospective-cohort): The prospective cohort design is an observational research design that can yield measures of relative risk, incidence, and longitudinal effects.
- [Prospective Research Designs Are Used With Future Outcomes](https://www.scalestatistics.com/prospective-research-designs): Prospective research designs are used when outcomes are collected and analyzed in the future. Prospective designs measures of incidence and relative risk.
- [Purposive Sampling is Used to Select Study Participants with a Purpose](https://www.scalestatistics.com/purposive-sampling): Purposive sampling is a non-probability sampling method where participants in a population are targeted in a purposeful manner for participation in a study.
- [Quasi-Experimental Designs Do Not Have Random Assignment](https://www.scalestatistics.com/quasi-experimental-designs): Choose a quasi-experimental research design that answers your research question. Quasi-experimental designs use random selection, but not random assignment.
- [Random Assignment - Randomization of Participants to Treatment Groups](https://www.scalestatistics.com/random-assignment): Random assignment is used in experimental research designs to randomly allocate study participants to the different treatment groups. Random assignment accounts for confounding and bias.
- [Random Selection - Randomization in Experimental Research Designs](https://www.scalestatistics.com/random-selection): Random selection is a probability sampling method used in experimental research designs where every member of a population has an equal chance of being chosen for participation in a study.
- [Randomization Methods in Randomized Controlled Trials Yields Causal Effects](https://www.scalestatistics.com/randomization-methods): Randomization methods in randomized controlled trials reduce bias, accounts for confounding, and yield causal effects.
- [Randomized Controlled Trials - The Gold Standard Research Design For Causal Effects](https://www.scalestatistics.com/randomized-controlled-trial): Randomized controlled trials (RCT) are considered an experimental design. Causal effects are found in RCTs due to the use of random selection and random assignment.
- [Ratio Variables Provide Measures of Distance and Magnitude with a "true zero"](https://www.scalestatistics.com/ratio-variables): Ratio variables possess a true zero and can yield measures of distance and magnitude. Ratio variables are considered the highest level of measurement.
- [Reliability Means Stability, Consistency, and Precision](https://www.scalestatistics.com/reliability): Reliability is the stability, consistency, and precision of a measure or score. Conduct internal consistency, test-retest, or inter-rater reliability in SPSS.
- [Research and statistics dictionary](https://www.scalestatistics.com/research-and-statistics-dictionary): Here is a dictionary of applied research and statistics terminology. This list of definitions covers a wide spectrum of empirical and statistical constructs.
- [Research Designs - Choose the Correct Research Design](https://www.scalestatistics.com/research-designs): Choose the correct research design using the Research Designs decision tree. Choose experimental, quasi-experimental, or observational research designs.
- [Retrospective Cohort Designs Are Used to Establish Odds Ratios](https://www.scalestatistics.com/retrospective-cohort): The retrospective cohort design is an observational research design that can yield measures of odds ratios, prevalence, and longitudinal effects.
- [Retrospective Research Designs Are Used for Outcomes That Have Already Occurred](https://www.scalestatistics.com/retrospective-research-designs): Choose a retrospective research design to answer your research question. Retrospective cohort, cross-sectional, case-control, and case series can be used.
- [Sample Size Affects Statistical Power](https://www.scalestatistics.com/statistical-power-and-sample-size): The sample size greatly affects statistical power. Large sample sizes will increase statistical power. Small sample sizes will decrease statistical power.
- [Sensitivity is The Ability of a Diagnostic Test to Detect Disease](https://www.scalestatistics.com/sensitivity): Sensitivity is the ability of a diagnostic test to detect disease in a population. The diagnostic test is compared to the “gold standard” method of diagnosis.
- [Simple Random Sampling Means Members of a Population Have an Equal Chance of Being Selected](https://www.scalestatistics.com/simple-random-sampling): Simple random sampling is a probability sampling method where all members of the population have an equal chance of being chosen for participation in a study.
- [Simple Randomization - Experimental Research Designs and Randomized Controlled Trials](https://www.scalestatistics.com/simple-randomization): Simple randomization is a method of random assignment used in experimental research designs and randomized control trials where study participants are randomized to equally sized treatment groups.
- [Simultaneous Regression Controls for All Variables](https://www.scalestatistics.com/simultaneous-regression): Simultaneous regression is used to predict for continuous outcomes when controlling for all pertinent confounding variables and it can be used in SPSS.
- [Skewness and Kurtosis Statistics Are Used to Test The Assumption of Normality](https://www.scalestatistics.com/skewness-and-kurtosis): Skewness and kurtosis statistics are used to assess the assumption of normality. Skewness and kurtosis statistics should be below |2.0| to assume normality.
- [Small Effect Sizes Decrease Statistical Power and Increase The Needed Sample Size](https://www.scalestatistics.com/statistical-power-and-small-effect-sizes): Small effect sizes decrease statistical power and increase the needed sample size. This is because more observations are needed to detect significance.
- [Small Sample Sizes Decrease Statistical Power and Decrease The Flexibility Of The Effect Size](https://www.scalestatistics.com/statistical-power-and-small-sample-sizes): Small sample sizes decrease statistical power and decrease the flexibility of detecting effect sizes. Small sample sizes often lead to Type II errors.
- [Specificity is The Ability of a Diagnostic Test to Detect Health](https://www.scalestatistics.com/specificity): Specificity is the ability of a diagnostic test to identify healthy people. The diagnostic test is compared to the “gold standard” method of diagnosis.
- [Sphericity is a Statistical Assumption Used With Repeated-Measures ANOVA](https://www.scalestatistics.com/sphericity): The assumption of sphericity is important for repeated-measures ANOVA. Greenhouse-Geisser corrections are used when the sphericity assumption is violated.
- [Standard Deviation Gives Context to Where Observations Fall in a Distribution](https://www.scalestatistics.com/standard-deviation): Standard deviation is a statistical measure that gives context to where each observation in a normal continuous distribution falls relative to the mean.
- [Statistical Assumptions Must Be Checked Before Using Inferential Statistics](https://www.scalestatistics.com/statistical-assumptions): Statistical assumptions have to be met when conducting inferential statistics. Statistical inferences are only valid when statistical assumptions are met.
- [Statistical Forum](https://www.scalestatistics.com/statistical-forum): Statistical Forum is the blog page for Research Engineer. The Statistical Forum is focused on applied research and statistical analysis topics.
- [Statistical Power and Magnitude of Effect Size](https://www.scalestatistics.com/statistical-power-and-effect-size): The magnitude of an effect size greatly impacts statistical power. Large effect sizes increase statistical power and small effect sizes decrease power.
- [Statistical Power and Scale of Measurement of Outcomes](https://www.scalestatistics.com/statistical-power-and-outcomes): Statistical power is affected by the outcome’s scale of measurement. Continuous outcomes increase statistical power, ordinal and categorical decrease power.
- [Statistical Power and The Chosen Research Design](https://www.scalestatistics.com/statistical-power-and-research-designs): Statistical power is affected by research designs. Within-subjects designs increase statistical power, between-subjects and multivariate designs decrease power.
- [Statistical Power and The Variance of Effect Sizes](https://www.scalestatistics.com/statistical-power-and-variance-of-effect-size): The variance of an effect size affects statistical power. Limited variance increases statistical power and extensive variance decreases statistical power.
- [Statistical Power Is The Ability To Detect Significant Treatment Effects](https://www.scalestatistics.com/statistical-power): Statistical power is the ability to detect significant treatment effects and it is affected by the outcome, research design, effect size, and sample size.
- [Stratified Random Sampling Allows for Subgroups or Strata of a Population to be Represented](https://www.scalestatistics.com/stratified-random-sampling): Stratified random sampling is a probability sampling method where subgroups or strata in a population are targeted for representation using random selection.
- [Stratified Randomization - Experimental Research Designs and Randomized Controlled Trials](https://www.scalestatistics.com/stratified-randomization): Stratified randomization is a method of random assignment in experimental research designs and randomized controlled trials where study participants are randomized across different strata.
- [Structure The Survey to Look Professional](https://www.scalestatistics.com/survey-structure): The survey should be structured in a professional manner before it is presented to potential respondents. All parts of the survey should be finalized.
- [Survival Analysis in SPSS: Kaplan-Meier and Cox Regression](https://www.scalestatistics.com/survival-analysis): Survival analysis is used to compare independent groups on their time to developing a categorical outcome. Use Kaplan-Meier and Cox regression in SPSS.
- [Systematic Review: Generate A Pooled Effect Across Multiple Studies](https://www.scalestatistics.com/systematic-review): A systematic review is used to generate a pooled effect using a meta-analysis of several high-quality randomized controlled trials or observational studies.
- [The Alpha Level is The Odds of Committing a Type I Error](https://www.scalestatistics.com/alpha-level): The alpha level denotes the chance researchers are willing to take in committing a Type I error. Alpha level also helps to determine statistical significance.
- [The Beta Level is The Odds of Committing a Type II Error](https://www.scalestatistics.com/beta-level): The beta level denotes the chance that researchers are willing to take in committing a Type II error. Beta level also helps to determine statistical power.
- [The Chi-Square Assumption States That Each Cell of The 2x2 Table Must Have Five Observations](https://www.scalestatistics.com/chi-square-assumption): The chi-square assumption states that each cell of the 2x2 table must have at least five observations. Fisher’s Exact Test is used if the assumption is violated.
- [The Construct Specification is The Blueprint for a Survey](https://www.scalestatistics.com/construct-specification): The construct specification provides a framework for defining a construct and its content areas. It contains the operational definition of a construct.
- [The Mean is The Average of a Continuous Distribution](https://www.scalestatistics.com/mean): The mean is the mathematical average of a normal distribution of continuous values. The mean is used to give context to parametric statistical tests.
- [The Median is The Observations That Falls in The Middle of a Distribution](https://www.scalestatistics.com/median): The median is the observation that occurs in the middle of a distribution of continuous values. The median is used to give context to non-parametric statistics.
- [The Mode is The Most Common Observation in a Distribution](https://www.scalestatistics.com/mode): The mode is the most commonly occurring or prevalent observation in a distribution of continuous values. The mode is used more often with frequency statistics.
- [The Null Hypothesis States There is No Difference or Association](https://www.scalestatistics.com/null-hypothesis): The null hypothesis states that there is no difference or association between variables of interest. Researchers either reject or do not reject the null.
- [The Research Hypothesis States There is An Association or Difference](https://www.scalestatistics.com/research-hypothesis): The research hypothesis states that there is a difference or association between variables of interest. Researchers are trying to prove the research hypothesis.
- [The Steps for Uploading Data into SPSS](https://www.scalestatistics.com/upload-data-into-spss): The steps for uploading research data into SPSS are presented. Excel formatted databases can be uploaded into SPSS for manipulation and statistical analysis.
- [There Are Five Modes of Survey Administration](https://www.scalestatistics.com/survey-modes-of-administration): There are five modes of survey administration: One-on-one interviews, group administration, telephone calls, postal mail, and electronic mail or email.
- [There Are Six Parts to Any Survey](https://www.scalestatistics.com/survey-parts): There are six parts to any survey: The title, an introduction, a set of instructions, the survey items, demographic questions, and a closing statement.
- [There Are Six Types of Surveys](https://www.scalestatistics.com/survey-types): There are six types of surveys that answer research questions: Inventories, checklists, performance, rating scales, tests, and psychological instruments.
- [There Are Three Scales of Measurement: Categorical, Ordinal, and Continuous](https://www.scalestatistics.com/scales-of-measurement): There are three primary scales of measurement: Categorical, ordinal, and continuous. Other scales of measurement include nominal, interval, ratio, and count.
- [There are Two Types of Criterion Validity: Predictive Validity and Concurrent Validity](https://www.scalestatistics.com/criterion-validity): Criterion validity is validity evidence that shows a survey instrument can predict for outcomes. Predictive validity or concurrent validity are examples.
- [There Are Two Types of Sampling Methods: Probability and Non-Probability](https://www.scalestatistics.com/sampling-methods): Choose a sampling method. Probability sampling uses random selection from the population. Non-probability sampling is used in observational research designs.
- [This Page is About Research Engineer and Eric Heidel, PhD](https://www.scalestatistics.com/about): Research Engineer is the world’s first online decision tree for applied research and statistics. Scale and Research Engineer were started by Eric Heidel, Ph.D.
- [Tolerance is Used in Regression to Assess Multicollinearity](https://www.scalestatistics.com/tolerance-and-regression): Tolerance measures for how much multicollinearity exists in a regression model. Smaller tolerance values are interpreted as meaning multicollinearity exists.
- [Transformations for ANOVA: Log Transform, Listwise Deletion, or Non-Parametric Statistics](https://www.scalestatistics.com/transformations-for-anova): Transformations can be conducted on non-normal distributions with ANOVA. SPSS can be used to conduct logarithmic transformations and Kruskal-Wallis tests.
- [Transformations for Independent Samples t-test](https://www.scalestatistics.com/transformations-for-independent-samples-t-test): Transformations can be conducted on non-normal distributions with independent samples t-test. SPSS can be used to conduct transformations or Mann-Whitney U.
- [Transformations for Repeated-Measures ANOVA: Listwise Deletion and Non-Parametric Statistics](https://www.scalestatistics.com/transformations-for-repeated-measures-anova): Transformations can be conducted on non-normal distributions with repeated-measures ANOVA. SPSS can be used to identify outliers and Friedman’s ANOVA.
- [Transformations for Repeated-Measures t-test](https://www.scalestatistics.com/transformations-for-repeated-measures-t-test): Transformations can be conducted on non-normal distributions with repeated-measures t-test. SPSS can be used to identify outliers and conduct a Wilcoxon test.
- [Types of Variables That Answer Research Questions](https://www.scalestatistics.com/types-of-variables): There are many types of variables used to answer research questions: Demographic, independent, control, dependent, predictor, confounding, and outcome variables.
- [Unequal Allocation Randomization - Experimental Research Designs and Randomized Controlled Trials](https://www.scalestatistics.com/unequal-allocation-randomization): Unequal allocation randomization is used in experimental research designs and randomized controlled trials so that study participants are randomized to unequally sized groups.
- [Use a Between-Subjects Database to Compare Independent Groups on Outcomes](https://www.scalestatistics.com/between-subjects-database): Between-subjects databases are structured so that independent groups can be compared on outcomes. Download a free database for between-subjects statistics.
- [Use a Multivariate Database to Run Regression Analyses](https://www.scalestatistics.com/multivariate-database): Multivariate databases are structured for the comparison of independent groups and multiple observations of outcomes. Download a free multivariate database.
- [Use a Survey Methodology That Answers The Research Question](https://www.scalestatistics.com/survey-methodology): A survey methodology is chosen in order to establish prevalence, compare independent groups, validate a construct, or measure for change in a population.
- [Use a Within-Subjects Database to Compare Observations of an Outcome](https://www.scalestatistics.com/within-subjects-database): Within-subjects databases are structured so that multiple observations of an outcome can be compared. Download a free database for within-subjects statistics.
- [Use and Interpret ANCOVA in SPSS](https://www.scalestatistics.com/ancova): ANCOVA is used to adjust an outcome based on the confounding effects of a continuous covariate when comparing independent groups on a continuous outcome in SPSS.
- [Use and Interpret ANOVA in SPSS](https://www.scalestatistics.com/anova): ANOVA is used to compare three or more groups on a normal continuous outcome. SPSS can be used to test the statistical assumptions as well as ANOVA.
- [Use and Interpret Biserial Correlations in SPSS](https://www.scalestatistics.com/biserial): The biserial correlation is a correlation test used when assessing the relationship between an ordinal variable and a continuous variable. Use biserial in SPSS.
- [Use and Interpret Bonferroni](https://www.scalestatistics.com/bonferroni): The Bonferroni adjustment can be used when adjusting for multiple comparisons in applied statistics. Bonferroni corrections reduce experimentwise error rates.
- [Use and Interpret Bootstrap Validation in SPSS](https://www.scalestatistics.com/bootstrap-validation): The bootstrap method of validating statistical findings means taking thousands of random samples from a dataset and computing bootstrap 95% confidence intervals.
- [Use and Interpret Chi-Square Goodness-of-Fit in SPSS](https://www.scalestatistics.com/chi-square-goodness-of-fit): Chi-square Goodness-of-fit tests are used to compare expected proportions against observed proportions. SPSS can be used to conduct Chi-square Goodness-of-fit.
- [Use and Interpret Chi-Square in SPSS](https://www.scalestatistics.com/chi-square): Chi-square is used to compare two groups on a dichotomous categorical outcome to yield unadjusted odds ratios. SPSS can be used to conduct chi-square.
- [Use and Interpret Cochran-Mantel-Haenszel in SPSS](https://www.scalestatistics.com/cochran-mantel-haenszel): Cochran-Mantel-Haenszel is used to assess conditional independence of categorical predictors associated with categorical outcomes. SPSS can run C-M-H.
- [Use and Interpret Cochran's Q in SPSS](https://www.scalestatistics.com/cochrans-q): Cochran’s Q is used to compare three or more within-subjects observations of a categorical outcome. Cochran’s Q can be conducted and interpreted using SPSS.
- [Use and Interpret Correlations in SPSS](https://www.scalestatistics.com/correlations): There are six different correlation tests that can be used in SPSS: Phi-coefficient, point biserial, rank biserial, Sperman’s rho, biserial, and Pearson’s r.
- [Use and Interpret Cox Regression in SPSS](https://www.scalestatistics.com/cox-regression): Cox regression is a type of survival analysis that predicts for a categorical outcome when controlling for variables and time. Use SPSS for Cox regression.
- [Use and Interpret Cronbach's alpha in SPSS](https://www.scalestatistics.com/cronbachs-alpha): Cronbach’s alpha is an internal consistency measure of reliability for survey items using Likert-type response sets. Use and interpret Cronbach’s alpha in SPSS.
- [Use and Interpret Different Types of Regression in SPSS](https://www.scalestatistics.com/regression): There are several types of regression that can be run in SPSS. Different methods of regression and regression diagnostics can be conducted in SPSS as well.
- [Use and Interpret Fishers Exact Test in SPSS](https://www.scalestatistics.com/fishers-exact-test): Fisher’s Exact Test is used to compare two groups on a dichotomous categorical outcome with small sample sizes. SPSS can be used to conduct Fisher’s Exact Test.
- [Use and Interpret Fixed-Effects ANOVA in SPSS](https://www.scalestatistics.com/fixed-effects-anova): Fixed-effects ANOVA is used to test the interaction between two categorical variables and a continuous outcome. Fixed-effects ANOVA can be used in SPSS.
- [Use and Interpret Frequency in SPSS](https://www.scalestatistics.com/baseline-frequency): A baseline frequency is used to establish a proportion within a categorical outcome. SPSS can be used to conduct a baseline frequency and cross-tabulation.
- [Use and Interpret Friedman's ANOVA in SPSS](https://www.scalestatistics.com/friedmans-anova): Friedman’s ANOVA is used to compare three or more within-subjects observations of an ordinal outcome. Friedman’s ANOVA can be conducted and interpreted in SPSS.
- [Use and Interpret Greenhouse-Geisser in SPSS](https://www.scalestatistics.com/greenhouse-geisser): The Greenhouse-Geisser correction is used for repeated-measures ANOVA when the assumption of sphericity is violated. SPSS can calculate the Greenhouse-Geisser.
- [Use and Interpret Independent Samples t-tests in SPSS](https://www.scalestatistics.com/independent-samples-t-test): Independent samples t-test is used to compare two groups on a continuous outcome. SPSS can be used to conduct assumptions and independent samples t-test.
- [Use and Interpret Jack-Knife Validation](https://www.scalestatistics.com/jack-knife-validation): The jack-knife method of validating statistical findings means taking out each participant in a dataset sequentially and running analyses on each sample.
- [Use and Interpret Kaplan-Meier in SPSS](https://www.scalestatistics.com/kaplan-meier): Kaplan-Meier is a type of survival analysis where independent groups are compared on their time to developing a categorical outcome. SPSS can be used.
- [Use and Interpret KR-20 in SPSS](https://www.scalestatistics.com/kr-20): Kudar-Richardson 20 (KR-20) is an internal consistency measure of reliability for survey items using dichotomous response sets. Use and interpret KR-20 in SPSS.
- [Use and Interpret Kruskal-Wallis in SPSS](https://www.scalestatistics.com/kruskal-wallis): Kruskal-Wallis is used to compare three or more groups on an ordinal outcome. SPSS can be used to conduct Kruskal-Wallis and post hoc Mann-Whitney U tests.
- [Use and Interpret Kruskal-Wallis When Homogeneity of Variance is Violated](https://www.scalestatistics.com/kruskal-wallis-and-homogeneity-of-variance): Kruskal-Wallis is used when the assumption of homogeneity of variance is violated for ANOVA. SPSS can be used to conduct the Kruskal-Wallis test.
- [Use and Interpret Logistic Regression in SPSS](https://www.scalestatistics.com/logistic-regression): Logistic regression is used to predict for dichotomous categorical outcomes. Logistic regression yields adjusted odds ratios with 95% CI when used in SPSS.
- [Use and Interpret MANCOVA in SPSS](https://www.scalestatistics.com/mancova): MANCOVA is used to account for increased Type I error rates when comparing independent groups on outcomes adjusted for by covariates. MANCOVA can be run in SPSS.
- [Use and Interpret Mann-Whitney U in SPSS](https://www.scalestatistics.com/mann-whitney-u): Mann-Whitney U is used to compare two groups on an ordinal outcome. Medians and interquartile ranges are reported. SPSS can be used to conduct Mann-Whitney U.
- [Use and Interpret Mann-Whitney U When Homogeneity of Variance is Violated](https://www.scalestatistics.com/mann-whitney-u-and-homogeneity-of-variance): Mann-Whitney U is used when the assumption of homogeneity of variance is violated for independent samples t-test. SPSS can be used to conduct Mann-Whitney U.
- [Use and Interpret MANOVA in SPSS](https://www.scalestatistics.com/manova): MANOVA is used to account for increased Type I error rates when comparing independent groups on multiple continuous outcomes. MANOVA can be used in SPSS.
- [Use and Interpret McNemar's Test in SPSS](https://www.scalestatistics.com/mcnemars): McNemar’s test is used to compare two within-subjects observations of a categorical outcome and is similar to relative risk. McNemar’s test can be used in SPSS.
- [Use and Interpret Median in SPSS](https://www.scalestatistics.com/baseline-median): A baseline median is used to establish a control median and interquartile range for an ordinal outcome. SPSS can be used to calculate a baseline median.
- [Use and Interpret Mixed-Effects ANOVA in SPSS](https://www.scalestatistics.com/mixed-effects-anova): Mixed-effects ANOVA is used to compare how independent groups change across time or within-subjects. Mixed-effects ANOVA can be run in SPSS.
- [Use and interpret Multinomial Logistic Regression in SPSS](https://www.scalestatistics.com/multinomial-logistic-regression): Multinomial logistic regression is used to predict for polychotomous categorical outcomes. Multinomial logistic regression yields odds ratios with 95% CI in SPSS.
- [Use and Interpret Multiple Regression in SPSS](https://www.scalestatistics.com/multiple-regression): Multiple regression is used to predict for a normal continuous outcome. Multiple regression models can be simultaneous, stepwise, or hierarchical in SPSS.
- [Use and Interpret Multivariate Statistics](https://www.scalestatistics.com/multivariate-statistics): Multivariate statistics are used to account for confounding variables and predict for outcomes. The choice of multivariate test depends upon the outcome.
- [Use and Interpret Multivariate Statistics for Categorical Outcomes](https://www.scalestatistics.com/multivariate-statistics-for-categorical-outcomes): Multivariate statistics for categorical outcomes include Cochran-Mantel-Haenszel, logistic regression, proportional odds regression, and survival analysis.
- [Use and Interpret Multivariate Statistics for Continuous Outcomes](https://www.scalestatistics.com/multivariate-statistics-for-continuous-outcomes): Multivariate statistics for continuous outcomes include fixed-effects ANOVA, random-effects ANOVA, ANCOVA, mixed-effects ANOVA, and multiple regression.
- [Use and Interpret Multivariate Statistics for Count Outcomes in SPSS](https://www.scalestatistics.com/multivariate-statistics-for-count-outcomes): Multivariate statistics for count outcomes include Poisson regression and negative binomial regression. These statistics yield adjusted odds ratios with 95% CI.
- [Use and Interpret Multivariate Statistics for Multiple Outcomes in SPSS](https://www.scalestatistics.com/multivariate-statistics-for-multiple-outcomes): Multivariate statistics for multiple outcomes include MANOVA and MANCOVA. These statistics are used to control for increased experimentwise error rates.
- [Use and Interpret Negative Binomial Regression in SPSS](https://www.scalestatistics.com/negative-binomial-regression): Negative binomial regression is used to predict for count outcomes where the variance of the outcome is higher than the mean and it can be run in SPSS.
- [Use and Interpret Observations in SPSS](https://www.scalestatistics.com/baseline-observation): A baseline observation is used to establish a control mean and standard deviation for a continuous outcome. SPSS can be used to calculate a baseline mean.
- [Use and Interpret One-Sample Median Test in SPSS](https://www.scalestatistics.com/one-sample-median-test): One-sample median tests are used to compare an expected median against an observed median. SPSS can be used to conduct and interpret a one-sample median test.
- [Use and Interpret One-Sample t-test in SPSS](https://www.scalestatistics.com/one-sample-t-test): One-sample t-tests are used to compare an expected mean to an observed mean. SPSS can be used to conduct and interpret a one-sample t-test for research.
- [Use and Interpret Pearson's r Correlation in SPSS](https://www.scalestatistics.com/pearsons-r): Pearson’s r correlation is a correlation test used when assessing the relationship between two continuous variables. Use and interpret Pearson’s r in SPSS.
- [Use and Interpret Phi-Coefficient in SPSS](https://www.scalestatistics.com/phi-coefficient): Phi-coefficient is a correlation test used when assessing the relationship between two categorical variables. Phi-coefficient can be used and interpreted in SPSS.
- [Use and Interpret Point Biserial Correlation in SPSS](https://www.scalestatistics.com/point-biserial): Point biserial is a correlation test used when assessing the relationship between a categorical and a continuous variable. Use point biserial in SPSS.
- [Use and Interpret Poisson Regression in SPSS](https://www.scalestatistics.com/poisson-regression): Poisson regression is used to predict for count outcomes where the mean of the outcome is higher than the variance. Poisson regression can be run in SPSS.
- [Use and Interpret Principal Components Analysis in SPSS](https://www.scalestatistics.com/principal-components-analysis): Principal components analysis (PCA) is a method for reducing data into correlated factors related to a construct or survey. Use and interpret PCA in SPSS.
- [Use and Interpret Proportional Odds Regression in SPSS](https://www.scalestatistics.com/proportional-odds-regression): Proportional odds regression is used to predict for ordinal outcomes. Proportional odds regression yields adjusted odds ratios with 95% CI when used in SPSS.
- [Use and Interpret Psychometrics in SPSS and AMOS](https://www.scalestatistics.com/psychometrics): The Psychometrics decision tree will help you choose the correct kind of reliability or validity evidence for your survey instrument. PCA and CFA can be chosen.
- [Use and Interpret Random-Effects ANOVA in SPSS](https://www.scalestatistics.com/random-effects-anova): Random-effects ANOVA is used to test the interaction between two or more categorical within-subjects observations on an outcome and it can be run in SPSS.
- [Use and Interpret Rank Biserial Correlation in SPSS](https://www.scalestatistics.com/rank-biserial): Rank biserial is a correlation test used when assessing the relationship between a categorical and an ordinal variable. Use and interpret rank biserial in SPSS.
- [Use and Interpret Repeated-Measures ANOVA in SPSS](https://www.scalestatistics.com/repeated-measures-anova): Repeated-Measures ANOVA is used to compare three or more within-subjects observations of a continuous outcome. Normality and sphericity must be tested in SPSS.
- [Use and Interpret Repeated-Measures t-test in SPSS](https://www.scalestatistics.com/repeated-measures-t-test): Repeated-measures t-test is used to compare two within-subjects observations of a normal continuous outcome. Repeated-measures t-test can be used in SPSS.
- [Use and Interpret Residual Analysis to Assess Model Fit](https://www.scalestatistics.com/residual-analysis): Residual analysis is used to assess the model fit of a regression model. Residual analysis must be conducted with using regression and it can be done in SPSS.
- [Use and Interpret Scheffe's Test](https://www.scalestatistics.com/sheffes-test): Scheffe’s test is a method used to adjust for multiple comparisons and is considered a conservative test. Scheffe’s test can be used and interpreted in SPSS.
- [Use and Interpret Spearman's rho Correlation in SPSS](https://www.scalestatistics.com/spearmans-rho): Spearman’s rho is a correlation test used when assessing the relationship between two ordinal variables. Use and interpret Spearman’s rho correlation in SPSS.
- [Use and Interpret Split-Group Validation](https://www.scalestatistics.com/split-group-validation): The split-group method of validating statistical findings means splitting a dataset into a derivation and confirmatory set and establishing similar results.
- [Use and Interpret Split-Half Reliability in SPSS](https://www.scalestatistics.com/split-half-reliability): Split-half reliability is an internal consistency measure of reliability where two equal halves of a survey are correlated. Use split-half reliability in SPSS.
- [Use and Interpret Stepwise Regression in SPSS](https://www.scalestatistics.com/stepwise-regression): Stepwise regression is used to predict for continuous outcomes using the best combination of predictor variables chosen by an algorithm. It can be run in SPSS.
- [Use and Interpret Test-Retest Reliability in SPSS](https://www.scalestatistics.com/test-retest-reliability): Test-retest reliability is used to assess the stability of survey scores across time. Use and interpret test-retest reliability and Spearman-Brown in SPSS.
- [Use and Interpret The Intraclass Correlation Coefficient (ICC) in SPSS](https://www.scalestatistics.com/intraclass-correlation-coefficient): Intraclass correlation coefficient (ICC) is an inter-rater reliability measure of agreement between raters using a continuous outcome. Use ICC in SPSS.
- [Use and Interpret The Kappa Statistic in SPSS](https://www.scalestatistics.com/kappa): Kappa is an inter-rater reliability measure of agreement between independent raters using a categorical or ordinal outcome. Use and interpret Kappa in SPSS.
- [Use and Interpret Tukey's HSD](https://www.scalestatistics.com/tukeys-hsd): Tukey’s HSD is a method used to adjust for multiple comparisons when all pairwise comparisons are going to be tested. Tukey’s HSD is used in the social sciences.
- [Use and Interpret Unadjusted Odds Ratio in SPSS](https://www.scalestatistics.com/unadjusted-odds-ratio): Unadjusted odds ratio is used to compare three or more groups on a categorical outcome. SPSS can be used to conduct unadjusted odds ratios and chi-square.
- [Use and Interpret Wilcoxon in SPSS](https://www.scalestatistics.com/wilcoxon): Wilcoxon is used to compare two within-subjects observations of an ordinal outcome. Wilcoxon, median, and interquartile range can be conducted in SPSS.
- [Use Diagnostic Testing Calculations to Establish Evidence](https://www.scalestatistics.com/diagnostic-testing): Diagnostic testing is used to establish evidence for the ability of diagnostic tests. Choose diagnostic testing calculations like sensitivity and specificity.
- [Use Receiver Operator Characteristic for Diagnostic Testing](https://www.scalestatistics.com/receiver-operator-characteristic): Receiver operator characteristic (ROC) curves are used to establish cut-points that maximize sensitivity and specificity. ROC curves yield c-statistics or AUC.
- [Use The Contact Form to Contact Research Engineer](https://www.scalestatistics.com/contact-form): Thank you for using Research Engineer! Please leave us feedback on the website and questions you have related to applied research and statistics.
- [Use The Sitemap to Navigate Research Engineer](https://www.scalestatistics.com/sitemap): The Research Engineer sitemap provides users access to every webpage available in Research Engineer. Use the sitemap to navigate to any page in Research Engineer.
- [Validation of Statistical Findings: Bootstrap, Split-Group, and Jack-Knife](https://www.scalestatistics.com/validation-of-statistical-findings): The validation of statistical findings can increase the internal and external validity of inferential statistics. Use bootstrap, split-group, or jack-knife.
- [Validity Means Utility, Interpretability, Generalizability, and Accuracy](https://www.scalestatistics.com/validity): Validity is the utility, interpretability, generalizability, and accuracy of a given measure or survey score. There are five types of validity evidence.
- [Variance Inflation Factor Measures for Multicollinearity in Regression](https://www.scalestatistics.com/variance-inflation-factor): The Variance Inflation Factor (VIF) measures for how much multicollinearity exists in a regression model. Variance Inflation Factor can be analyzed in SPSS.
- [Variance is The Amount of Variability in a Distribution](https://www.scalestatistics.com/variance): Variance is a statistical measure of the overall dispersion of a distribution of continuous values. Variance is used to calculate the standard deviation.
- [Within-Subjects Designs Increase Statistical Power and Decrease The Sample Size](https://www.scalestatistics.com/statistical-power-and-within-subjects-designs): Within-subjects designs increase statistical power and decrease the needed sample size. This is because each participant serves as their own control.
- [Within-Subjects Statistics Are Used to Compare Outcomes Across Time](https://www.scalestatistics.com/within-subjects-statistics): Within-subjects statistics are used to compare one, two, or three or more observations of categorical, ordinal, and continuous outcome variables.
- [Within-Subjects Statistics for One Observation: Categorical, Ordinal, and Continuous Outcomes](https://www.scalestatistics.com/within-subjects-statistics-for-one-observation): There are three types of within-subjects statistical tests for one group. These include a baseline frequency, a baseline median, and a baseline observation.
- [Within-Subjects Statistics for Three or More Observations](https://www.scalestatistics.com/within-subjects-statistics-for-three-or-more-observations): There are three within-subjects statistical tests for three or more observations. They are Cochran’s Q, Friedman’s ANOVA, and repeated-measures ANOVA.
- [Within-Subjects Statistics for Two Observations: Categorical, Ordinal, and Continuous Outcomes](https://www.scalestatistics.com/within-subjects-statistics-for-two-observations): There are three within-subjects statistical tests for two observations: McNemar’s test, Wilcoxon, and repeated-measures t-test. All tests can be used in SPSS.
- [Write Survey Items That Cover The Survey Content](https://www.scalestatistics.com/survey-items): Survey items have two parts: An item stem and a response set. Survey items are written to cover the content areas identified in the construct specification.
- [Write Valid Research Questions Using FINER and PICO](https://www.scalestatistics.com/research-questions): The Research Questions decision tree will help with writing research questions. The FINER and PICO frameworks are used to write valid research questions.
- [Writing instructional goals and objectives](https://www.scalestatistics.com/writing-instructional-goals-and-objectives): Bloom's Taxonomy is a framework used for writing instructional goals and objectives. Use 54 action verbs to write lower and higher order goals and objectives.
- [Youden Index Calculates The Overall Ability of a Diagnostic Test](https://www.scalestatistics.com/youden-index): The Youden Index is a calculation that determines the overall benefit of a diagnostic test. It shows a test’s ability to balance sensitivity and specificity.

## Blog Posts
- [Statistical nuggets by Dr. Heidel](https://www.scalestatistics.com/p/statistical-nuggets-by-dr-heidel): Check for normality using skewness and kurtosis statistics.
- [Positive Predictive Value and Prevalence](https://www.scalestatistics.com/p/positive-predictive-value-and-prevalence): Positive predictive value (PPV) will increase as prevalence increases.
- [Precision and Accuracy](https://www.scalestatistics.com/p/precision-and-accuracy): Precision and accuracy in measurement is very important in research.
- [Effect size, sample size, and statistical power](https://www.scalestatistics.com/p/effect-size-sample-size-and-statistical-power): Statistical power, effect size, and sample size are interdependent.
- [Meeting statistical assumptions](https://www.scalestatistics.com/p/meeting-statistical-assumptions): Meeting statistical assumptions like normality and homogeneity of variance is important when conducting statistics.
- [Retrospective cohort designs are useful to many researchers](https://www.scalestatistics.com/p/retrospective-cohort-designs-are-useful-to-many-researchers): Retrospective cohort designs are very useful and feasible for researchers.
- [Sampling methods in research](https://www.scalestatistics.com/p/sampling-methods-in-research): There are two types of sampling methods in research: Probability and non-probability.
- [Non-parametric statistics as post hoc tests](https://www.scalestatistics.com/p/non-parametric-statistics-as-post-hoc-tests): Mann-Whitney U and Wilcoxon tests are used as post hoc tests with non-parametric statistics.
- [Prevalence vs. Incidence](https://www.scalestatistics.com/p/prevalence-vs-incidence): There are fundamental differences between prevalence and incidence.
- [The research question is the foundation of everything empirical](https://www.scalestatistics.com/p/the-research-question-is-the-foundation-of-everything-empirical): The research question is the foundation of everything empirical.
- [Statistical tests](https://www.scalestatistics.com/p/statistical-tests): Statistics are used to answer research questions.
- [Using naturally skewed continuous variables as outcome variables](https://www.scalestatistics.com/p/using-naturally-skewed-continuous-variables-as-outcome-variables): Naturally skewed continuous variables can be analyzed in several different ways.
- [Preliminary statistical consultation](https://www.scalestatistics.com/p/preliminary-statistical-consultation): Preliminary statistical consultation on research studies is important.
- [G*Power for the masses](https://www.scalestatistics.com/p/gpower-for-the-masses): G*Power is a great statistical program that researchers should use in their everyday practice.
- [Evidence-based medicine and its applications](https://www.scalestatistics.com/p/evidence-based-medicine-and-its-applications): Critical appraisal of clinical evidence cannot occur without better understanding of research design and statistics.
- [Operationalization of constructs and behaviors](https://www.scalestatistics.com/p/operationalization-of-constructs-and-behaviors): Operationalization of constructs and behaviors is important in measurement.
- [Statistical Designs](https://www.scalestatistics.com/p/statistical-designs): Between-subjects and within-subjects designs are popular in statistics.
- [Ordinal measures becoming continuous with normality](https://www.scalestatistics.com/p/ordinal-measures-becoming-continuous-with-normality): Ordinal measures can be assessed at a continuous level if they meet normality.
- [Chi-square p-values are not enough](https://www.scalestatistics.com/p/chi-square-p-values-are-not-enough): Odds ratios with 95% confidence intervals are the primary inference with chi-square and Fisher's Exact test.
- [Non-parametric statistics and small sample sizes](https://www.scalestatistics.com/p/non-parametric-statistics-and-small-sample-sizes): Non-parametric statistics should be used with small sample sizes.
- [95% confidence intervals](https://www.scalestatistics.com/p/95-confidence-intervals): 95% confidence intervals are important statistics.
- [Publication bias](https://www.scalestatistics.com/p/publication-bias): Publication bias is a problem within science and empiricism.
- [Measurement at continuous levels](https://www.scalestatistics.com/p/measurement-at-continuous-levels): Measure for variables at the continuous level if at all possible.
- [Values needed for sample size calculations](https://www.scalestatistics.com/p/values-needed-for-sample-size-calculations): Evidence-based measures of effect can be used to conduct an a priori power analysis.
- [Publication of Research Findings](https://www.scalestatistics.com/p/publication-of-research-findings): It is important to be objective when writing up the methods and results of a study.
- [Multivariate statistical designs](https://www.scalestatistics.com/p/multivariate-statistical-designs): There are many different types of multivariate statistics.
- [Applying clinical evidence and journal clubs](https://www.scalestatistics.com/p/applying-clinical-evidence-and-journal-clubs): Journal clubs need to change the way they impact residents.
- [Non-inferiority trials and the Affordable Care Act](https://www.scalestatistics.com/p/non-inferiority-trials-and-the-affordable-care-act): Non-inferiority trials will become important in the future with the Affordable Care Act.
- [The Bcc line](https://www.scalestatistics.com/p/the-bcc-line): The Bcc: line in an email should be used for mass mailings in survey research.
- [Construct specification in survey research](https://www.scalestatistics.com/p/construct-specification-in-survey-research): The first step in creating a survey is to build a construct specification.
- [Adjusted odds ratios in medicine](https://www.scalestatistics.com/p/adjusted-odds-ratios-in-medicine): Adjusted odds ratios are important statistics in medicine.
- [Merging databases](https://www.scalestatistics.com/p/merging-databases): Unique numerical de-identifiers and grouping variables are important when merging databases.
- [Acquiring the clinical evidence](https://www.scalestatistics.com/p/acquiring-the-clinical-evidence): This post will provide you with information about the use of Boolean operators in search queries to improve the specificity of your searches.
- [Feasible research questions are answerable](https://www.scalestatistics.com/p/feasible-research-questions-are-answerable): Researchers should make sure that their research question is feasible.
- [Bonferroni corrections](https://www.scalestatistics.com/p/bonferroni-corrections): The Bonferroni correction is a stalwart of statistical and empirical reasoning.
- [McNemar's as a post hoc test for Cochran's Q](https://www.scalestatistics.com/p/mcnemars-as-a-post-hoc-test-for-cochrans-q): McNemar's Test can be used as a post hoc test for Cochran's Q.
- [Mastery of the literature](https://www.scalestatistics.com/p/mastery-of-the-literature): Mastering the literature is the first step in conducting any type of research.
- [Biostatistical scientists](https://www.scalestatistics.com/p/biostatistical-scientists): Biostatistical scientists invent new statistical methods for applied practice.
- [Writing survey items](https://www.scalestatistics.com/p/writing-survey-items): The 5-point Likert scale increases variance, allows neutrality, is practical in survey research, and should be written in increasing order.
- [Parametric vs. non-parametric statistics](https://www.scalestatistics.com/p/parametric-vs-non-parametric-statistics): There are fundamental differences between parametric and non-parametric statistics.
- [Dichotomous variables in SPSS](https://www.scalestatistics.com/p/dichotomous-variables-in-spss): SPSS uses a default for reference categories that researchers must be aware of and change accordingly.
- [Chi-square vs. Fisher's Exact Test](https://www.scalestatistics.com/p/chi-square-vs-fishers-exact-test): There is a fundamental difference between chi-square and Fisher's Exact test.
- [Small sample sizes, Type II errors, and empirical reasoning](https://www.scalestatistics.com/p/small-sample-sizes-type-ii-errors-and-empirical-reasoning): Small sample sizes are common with less prevalent outcomes.
- [Logarithmic transformations for skewed variables](https://www.scalestatistics.com/p/logarithmic-transformations-for-skewed-variables): Logarithmic transformations can be used on continuous variables that violate normality or that are naturally skewed.
- [The role of correlations in psychometrics](https://www.scalestatistics.com/p/the-role-of-correlations-in-psychometrics): Correlations are prevalent in psychometrics for generating reliability and validity evidence.
- [The assumption of independence of observations](https://www.scalestatistics.com/p/the-assumption-of-independence-of-observations): Independence of observations stipulates that observations are independent of each other.
- [Predictive validity is a powerful type of psychometric evidence](https://www.scalestatistics.com/p/predictive-validity-is-a-powerful-type-of-psychometric-evidence): Predictive validity is defined as an instrument's ability to predict future occurrences.
- [The Kappa statistic](https://www.scalestatistics.com/p/the-kappa-statistic): The Kappa statistic is a measure of inter-rater reliability for dichotomous categorical ratings.
- [FINER and PICO](https://www.scalestatistics.com/p/finer-and-pico): The FINER and PICO frameworks are used to write credible and valid research questions.
- [Non-parametric Friedman's ANOVA](https://www.scalestatistics.com/p/non-parametric-friedmans-anova): Friedman's ANOVA is a non-parametric within-subjects test for three or more observations of an ordinal outcome.
- [Merry Christmas to all my loved ones, friends, colleagues, and site visitors!](https://www.scalestatistics.com/p/merry-christmas-to-all-my-loved-ones-friends-colleagues-and-site-visitors): Merry Christmas to all visitors to www.scalelive.com and a Happy New Year!
- [Statistical assumptions must be tested when using inferential statistics](https://www.scalestatistics.com/p/statistical-assumptions-must-be-tested-when-using-inferential-statistics): Statistical assumptions have been posted to Research Engineer.
- [Statistical assumptions for inferential statistics](https://www.scalestatistics.com/p/statistical-assumptions-for-inferential-statistics): The statistics used to assess statistical assumptions are summarized.
- [Research Engineer is the world's first online decision tree for applied research and statistics](https://www.scalestatistics.com/p/research-engineer-is-the-worlds-first-online-decision-tree-for-applied-research-and-statistics): Research Engineer is the first online decision engine for research, statistics, EBM, surveys, psychometrics, epidemiology, diagnostic testing, variables, and education.
- [New propensity score matching, calculators, reliability, and regression diagnostics pages in Research Engineer](https://www.scalestatistics.com/p/new-propensity-score-matching-calculators-reliability-and-regression-diagnostics-pages-in-research-engineer): Propensity Score Matching, Calculators, Reliability, and Regression Diagnostics are now available on Research Engineer.
- [Prospective cohort designs provide measures of risk and incidence](https://www.scalestatistics.com/p/prospective-cohort-designs-provide-measures-of-risk-and-incidence): Prospective cohort designs yield measures of incidence and risk, longitudinal effects, and reduced observational bias.
- [Between-subjects one-sample median test](https://www.scalestatistics.com/p/between-subjects-one-sample-median-test): One-sample median tests can test hypothesized medians against observed medians.
- [Follow Research Engineer on Facebook, Twitter, Google, YouTube, Tumblr, Pinterest, Instagram, LinkedIn, Flickr, and Vimeo](https://www.scalestatistics.com/p/follow-research-engineer-on-facebook-twitter-google-youtube-tumblr-pinterest-instagram-linkedin-flickr-and-vimeo): Users can now follow Research Engineer on Facebook, Twitter, Google+, YouTube, Tumblr, Pinterest, Instagram, LInkedIn, Flickr, and Vimeo.
- [Case series are the lowest level of clinical evidence](https://www.scalestatistics.com/p/case-series-are-the-lowest-level-of-clinical-evidence): Case series designs are good for hypothesis generation, effect sizes, and studying rare outcomes.
- [Number Needed to Treat is an important epidemiological calculation](https://www.scalestatistics.com/p/number-needed-to-treat-is-an-important-epidemiological-calculation): Number Needed to Treat is an important epidemiological calculation.
- [Establishing causal effects](https://www.scalestatistics.com/p/establishing-causal-effects): There are certain criteria required for establishing causal effects.
- [Research and statistics in the collective unconscious](https://www.scalestatistics.com/p/research-and-statistics-in-the-collective-unconscious): Research and statistics cause cognitive dissonance in people.
- [Interview with Teknovation](https://www.scalestatistics.com/p/interview-with-teknovation): Teknovation.biz gives a background story and new updates on Research Engineer.
- [Internal search engines are now available in Research Engineer](https://www.scalestatistics.com/p/internal-search-engines-are-now-available-in-research-engineer): Internal search engines are now available on every page of Research Engineer.
- [Research designs are used to answer research questions](https://www.scalestatistics.com/p/research-designs-are-used-to-answer-research-questions): Research designs are used to answer research questions.
- [Categorical measurement caveats](https://www.scalestatistics.com/p/categorical-measurement-caveats): There are several caveats associated with using categorical variables.
- [Sensitivity and specificity](https://www.scalestatistics.com/p/sensitivity-and-specificity): Sensitivity and specificity are diagnostic testing measures that show a test's ability to either detect disease or identify the healthy.
- [Copyright for Research Engineer](https://www.scalestatistics.com/p/copyright-for-research-engineer): Scalë, LLC has acquired a copyright for Research Engineer!
- [Basic principles of correlational research](https://www.scalestatistics.com/p/basic-principles-of-correlational-research): This post will explain the utility and application of Spearman's rho and Pearson's r in correlational research.
- [Sitemap and Search now available in Research Engineer](https://www.scalestatistics.com/p/sitemap-and-search-now-available-in-research-engineer): New webpages have been added to Research Engineer, Sitemap and Search.
- [Equivalency Trial Calculator](https://www.scalestatistics.com/p/equivalency-trial-calculator): A new Equivalency Trial calculator is available in Research Engineer.
- [Non-Inferiority Trial Calculator](https://www.scalestatistics.com/p/non-inferiority-trial-calculator): A new non-inferiority trial calculator is available for download in Research Engineer.
- [New pages for sampling, variables, descriptive statistics, and regression in Research Engineer](https://www.scalestatistics.com/p/new-pages-for-sampling-variables-descriptive-statistics-and-regression-in-research-engineer): There are several new pages for descriptive statistics, sampling methods, measurement of variables, and regression in Research Engineer.
- [Between-subjects statistics are used to compare independent groups](https://www.scalestatistics.com/p/between-subjects-statistics-are-used-to-compare-independent-groups): Between-subjects statistics are used to compare independent groups on outcomes.
- [Intraclass Correlation Coefficient and inter-rater reliability](https://www.scalestatistics.com/p/intraclass-correlation-coefficient-and-inter-rater-reliability): Intraclass Correlation Coefficient (ICC) is a measure of inter-rater reliability used when two or more raters give ratings at a continuous level.
- [Patent for Research Engineer](https://www.scalestatistics.com/p/patent-for-research-engineer): A non-provisional patent for Research Engineer is being filed.
- [Revenue for Research Engineer](https://www.scalestatistics.com/p/revenue-for-research-engineer): Revenue is generated for Research Engineer through statistical consultation.
- [Comparative Effectiveness Research and the PPACA](https://www.scalestatistics.com/p/comparative-effectiveness-research-and-the-ppaca): Comparative Effectiveness Research is an important research design sponsored by the PPACA.
- [New Pages for Hypothesis Testing, Measurement, and Populations in Research Engineer](https://www.scalestatistics.com/p/new-pages-for-hypothesis-testing-measurement-and-populations-in-research-engineer): New pages for hypothesis testing, measurement, and populations have been published to Research Engineer.
- [New published research from Eric Heidel ](https://www.scalestatistics.com/p/new-published-research-from-eric-heidel): R. Eric Heidel, Ph.D., Owner and Operator or Scale, LLC and Research Engineer, recently published new research.
- [Within-subjects designs increase statistical power](https://www.scalestatistics.com/p/within-subjects-designs-increase-statistical-power): Within-subjects designs increase statistical power because each participant serves as their own control.
- [Eric Heidel, Owner and Operator of Scale, LLC, gets married](https://www.scalestatistics.com/p/eric-heidel-owner-and-operator-of-scale-llc-gets-married): Eric Heidel, Ph.D., Owner and Operator of Scale, LLC, got married!
- [Research Engineer makes applied research and statistics easier](https://www.scalestatistics.com/p/research-engineer-makes-applied-research-and-statistics-easier): Research Engineer is designed to get you to the correct research or statistical methodology.
- [Easy statistics help with SPSS](https://www.scalestatistics.com/p/easy-statistics-help-with-spss): This post will give you access to the methods for conducting and interpreting statistical tests in SPSS.
- [Curriculum vitae for Eric Heidel, Ph.D., Owner and Operator of Scale, LLC and Research Engineer](https://www.scalestatistics.com/p/curriculum-vitae-for-eric-heidel-phd-owner-and-operator-of-scale-llc-and-research-engineer): The most current version of the curriculum vitae for Eric Heidel, Ph.D., Owner and Operator of Scale, LLC and Research Engineer is now available.
- [Causality in Statistical Power: Isomorphic Properties of Measurement, Research Design, Effect Size, and Sample Size](https://www.scalestatistics.com/p/causality-in-statistical-power-isomorphic-properties-of-measurement-research-design-effect-size-and-sample-size): The newest publication from Dr. Eric Heidel is entitled Causality in Statistical Power: Isomorphic Properties of Measurement, Research Design, Effect Size, and Sample Size