Research Methodology: Complete Research Project Blueprint

Frame your ideas into a clear research question by focusing on a cause-and-effect relationship. Identify the independent and dependent variables and ensure the concepts are measurable to answer the question.

Distinguish numerical variables from categorical variables in your research question, noting numerical variables have numeric levels while categorical variables have unordered levels, using examples like number of friends and gender.

Practice identifying independent and dependent variables and their data types, including levels for categorical variables, from research questions, and prepare to apply these concepts in research design.

Design studies that rule out alternative explanations and disentangle causal direction, clarifying whether the independent variable influences the dependent variable or if a third variable drives both.

Explore how to interpret statistical relationships by evaluating three explanations—mediation by a third variable, reverse causation, and a direct link—using examples from meditation, education, and entrepreneurship.

Learn to rule out alternative explanations by manipulating the independent variable in experiments, establishing a causal link between variables (e.g., meditation and happiness) through randomized designs.

Explore the three main research designs—between subject design, within subjective time, and mix design—using treatment versus no treatment and illness as a numeric outcome measured before and after.

Stick to between-subject or within-subject designs, avoiding unnecessary complexity; manipulate a categorical independent variable with two levels, and use a numerical dependent variable to detect small effects.

Randomize to avoid bias and ensure the only difference between groups is the manipulation; use coin flips or counterbalancing across between-subject, within-subject, and mixed designs.

Compare experimental and non-experimental designs, explaining how randomization, counterbalancing, and design type (between-subject, within-subject, mixed) influence interpretation and ruling out alternative explanations.

Practice identifying experimental and non-experimental designs from case studies, decide design type and the three causal explanations, and explore how randomization could convert non-experimental designs to experimental.

Explore how confounders affect research design and differ from the independent variable using treatment versus control. Apply randomization and matched control groups to eliminate confounders and clarify effects.

Identify and control typical confounders across experimental designs, including placebo effects, demand characteristics, and the double-blind procedures. Ensure randomization and counterbalancing address selection and regression to the mean.

Design successful non-experimental studies by using mix designs or quasi experiments to address variables that cannot be manipulated and maximize information with precise numerical measurements.

Assess designs case by case to rule out explanations and confounders with confounder checklist. Prioritize experiments to isolate causal effects, consider between-subject, within-subject, or mixed designs, and use numerical variables.

Explore how experimental designs use randomization to rule out alternative explanations, compare between, within, and mixed designs, and identify confounders.

Learn how to make variables measurable by formulating operational definitions that translate abstract concepts into data for analysis, using health, performance, and company size as examples.

Explore the two core aspects of effective measurement—reliability and validity—and learn how to apply test-retest, internal consistency, and inter-observer reliability, along with convergent and discriminant validity.

Boost reliability by taking multiple measurements and averaging them, whether weighing individuals or using several questionnaire items that measure the same trait, then averaging the responses.

Explore self-report methods by asking directly through interviews or surveys, using open or closed questions, avoiding leading prompts, and measuring latent constructs with multiple items while guarding against social desirability.

Compare performance, observed behavior, and physical measurements in research methodology, noting hard-to-fake outcomes yet reliability and setup challenges, and learn how to choose the right measurement type.

Balance reliability and validity when selecting measurement types in your context. Address social desirability and validity concerns by adjusting questions or using alternatives, and compare converging measures for clearer results.

Switch your research question when measurement is unreliable, replace criminal behavior with criminal attitudes, and compare intervention and control groups using scenario judgments to detect attitude changes.

Learn how to create accurate measurements with operational definitions, assess reliability and validity—including convergent and discriminant validity—across self-report, performance, observed behavior, and physical measures, and use multiple measurements for precision.

Define and collect data by drawing a random, independent, and large sample to achieve external validity and reliability for generalizing findings to the population.

Explore data collection strategies, including random and stratified sampling to reduce bias and ensure subgroup representation, and consider purposive and convenient sampling when resources constrain external validity.

Ensure an externally valid, unbiased sample by using random and independent sampling, with at least 30 participants and power analysis to determine the exact size.

Learn the three steps of data analysis—data preparation, descriptive statistics, and inferential statistics—focusing on data cleaning and patterns, with Jasper as a beginner-friendly tool.

Identify correlations using scatterplots for numerical variables, observing positive or negative patterns and no association. Interpret the R coefficient, ranging from -1 to 1, to gauge strength and direction.

Spot relationships between numerical variables using scatter plots and the correlation coefficient, identifying positive, negative, and zero correlations and interpreting r values in real data examples.

Spot relationships between numerical variables by calculating Pearson correlations, generating scatterplots, and interpreting r values for high, medium, and zero variable pairs in Jazz.

Compare the averages of illness between treatment and control in a between-subject design with a categorical independent variable and numerical dependent variable, using scatterplots and box plots to reveal correlation.

Using Jasp, split illness by the categorical group (control vs treatment), view descriptives and a box plot to show lower illness in the treatment group.

Compare pre- and post-treatment illness levels using averages and box plots to identify treatment effects in a within-subject design.

In this tutorial, analyze a within-subject design in Jasper using illness before and after treatment data, with 30 valid cases and box plots showing a treatment effect.

Learn how a mixed design compares treatment and control groups with a categorical independent variable and a numerical dependent variable using before-and-after box plots to detect larger treatment effects.

Explore relationships between a categorical independent variable and a numerical dependent variable in a mixed design, comparing illness before and after across control and treatment groups with box plots.

Examine how a numerical independent variable relates to a categorical dependent variable using swapped-axis scatterplots and box plots to detect a relationship or no relationship.

Explore how a numerical independent variable (H) relates to a categorical dependent variable (purchase behavior) using descriptives and a box plot to show buyers are younger than non-buyers.

analyze the relationship between a categorical independent variable and a categorical dependent variable by using a contingency table of quadrant counts to compare purchase probabilities for men and women.

Use a contingency table to examine the relationship between a categorical independent variable (gender) and dependent variable (purchase behavior), comparing column percentages to show higher purchase among males.

Plan ahead to prepare data for analysis by importing and formatting, handling missing values and outliers, and structuring a data set with one row per participant and within-subject measurements.

Prepare between-subject data by converting misformatted files into a correctly structured dataset with a group variable and an illness variable, one participant per row.

Demonstrates converting a within-subject data set to the correct format by aligning IDs, creating illness before and after variables, and matching time labels for analysis.

Explore strategies for missing values in data preparation: replace with the column mean, drop rows, or drop columns, with notes on when each method is appropriate and potential bias.

Identify missing values with descriptives, then choose to remove the row, remove the column, or impute the mean; save a new dataset version in Jasper, Google Sheets, or Excel.

Identify and handle outliers using box plots and scatter plots. Determine when a data point reflects a valid measurement, and compare analyses with and without outliers to decide on removal.

Learn to spot and remove outliers using box plots and scatter plots, assess how outliers shape group differences and correlations, and apply data preparation for missing values.

Prepare data, handle missing values and outliers, and apply descriptive statistics to identify patterns, relationships, and differences across numerical and categorical variables using appropriate plots.

Evaluate how reliability affects generalizing findings from a sample to the population, using inferential statistics to rule out coincidental differences between treatment and control groups.

Learn how to generalize from a single variable by linking sample mean to population mean, focusing on variation and sample size as the two forces driving deviation.

Apply principle two of generalizing from single variables by increasing sample size to bring the sample mean closer to the population, and assess variation via visualizations to gauge accuracy.

Examine how two-variable relationships generalize from treatment versus control, testing population mean differences. Understand how effect size, variation, and sample size determine the likelihood of generalization.

Learn how to interpret p values to decide if a statistical relationship generalizes to the population, using the 0.05 threshold and sample size considerations for means, correlations, and probabilities.

Learn how inferential statistics test treatment effects, rule out the null hypothesis, and select the right statistical test using a decision chart to compute the p value.

generalize a sample correlation to the population using scatterplots, correlation coefficients, effect size, variation, and sample size to decide if it differs from zero.

Demonstrate correlation analysis using data on meditation frequency and happiness, computing the coefficient and P value, then report significant results in APA style with degrees of freedom.

Generalize mean differences to the population using anova to compare treatment and control groups, interpreting effect size, variation, and p-values in a between-subject design.

Visualize data with box plots, run anova, assess outliers, and generalize group mean differences by reporting means, standard deviations, f, and p values.

Explore generalizing a mean difference over time using a within-subject design and rm anova to compare before and after treatment effects on illness, with box plots and hypothesis testing.

Use a repeated measures ANOVA to compare illness before versus after treatment, examine descriptive statistics and plots, address outliers, and report F, df, and p with means and SDs.

Generalize from a mixed design by applying a repeated measures anova to treatment and control groups, comparing before and after changes, and evaluating effect size, variation, sample size, and hypotheses.

Generalize a mixed design with a repeated measures ANOVA on illness, using time and group, and report the significant time by group interaction (p = 0.002) in APA style.

Explore generalizing a numerical independent variable with a categorical dependent variable using logistic regression, analyzing mean differences, probabilities, and hypotheses about purchases across ages.

Generalize the relationship between a numerical predictor and a categorical outcome using logistic regression. Learn to interpret odds ratios and apa format reporting for age and purchase behavior.

Generalize a difference in probabilities with a chi square test, comparing purchase rates between men and women using a contingency table and null versus alternative hypotheses.

Analyze a contingency table of purchase behavior by gender with a chi-square test, report column percentages and a non-significant p value, and discuss generalizing results to the population.

Identify two common mistakes in inferential statistics: accepting the null hypothesis and conducting multiple tests, and learn to apply the Bonfire Ferrone correction to prevent incorrect conclusions.

Explore reliability and external validity while generalizing findings to the population by selecting the right statistical test, evaluating effect size, variation, and sample size, and interpreting p-values against 0.05.

Explore the two types of statistical errors—false positives and false negatives—within inferential statistics, and learn how sample size, p-value threshold, and null/alternative hypotheses guard against them.

Explore how p value thresholds and sample size balance false positives and false negatives in inferential statistics; use 0.05 to minimize false positives and larger samples to reduce false negatives.

Learn how to determine the right sample size through power analysis, using statistical power to minimize false negatives and estimate required size with standardized effect sizes, often via G Power.

Learn to perform sample size calculation for correlations using power analysis in G Power, specifying two-tailed tests, correlation as effect size, and interpreting power and false positives.

Learn to calculate the required sample size for a between-subject ANOVA using power analysis in G power, adjusting effect size, alpha, and power, with design options like repeated measures.

Calculate sample size for a repeated measures ANOVA (within-subject) using G Power, illustrating how within-subject designs require fewer participants and how two measurements, effect size, alpha, and power drive results.

Calculate sample size for a mixed design repeated measures ANOVA, addressing correlation among repeated measures and nonsphericity correction, using two groups and two measurements in G Power.

Compute sample size for logistic regression with a numerical predictor and a categorical outcome, exploring power, effect size (odds ratio), and how probability inputs influence required samples.

Explore how to calculate chi square sample size for two categorical variables using G power, including effect size W, degrees of freedom, and implications of 2x2 contingency tables for power.

Boost power by crafting a strong manipulation to maximize effect size and minimize variation through reliable measurement, using extreme designs with a categorical independent variable and a numerical dependent variable.

Improve statistical power by strengthening measurement reliability through standardized procedures. Measure multiple times and average to reduce variation, clarifying effect size, sample size, and p-values.

Learn to test measurement reliability before or during your study, using pilot testing, test retest reliability, internal consistency, and inter observer reliability, and interpret correlations and Crossbars Alpha.

Explore how to test and improve measurement reliability using JASP, assessing test-retest reliability with weight measurements and evaluating internal consistency with Cronbach's alpha across five questionnaire items.

Test measurement validity by examining convergent and discriminant validity with correlations in Jasper, linking fitness to running speed and heartbeat while ignoring height and shirt counts.

Estimate the required sample size using power analysis to minimize false negatives and false positives. Balance effect size, variation, and measurement reliability to maximize statistical power and detect true effects.

Master the research workflow from screening and informed consent to manipulating or measuring variables, checking demand characteristics, and debriefing after handling deception or placebo treatments.

Learn practical strategies to recruit participants—from university labs to field, street, and online platforms—while using incentives, data cleaning, and sanity questions to ensure serious participation.

Learn how to recruit participants with incentives such as money, a lottery, or course credits, and use university, field, or online methods, while ensuring informed consent and proper debriefing.

Learn to interpret findings by following four steps: assess reliability (p<0.05), verify measurement accuracy, evaluate explanations and confounders, and judge external validity for generalization.

Assess reliability and validity of results from a between-subject design comparing treatment and control groups for depression, and evaluate explanations, confounders, and external validity to judge effectiveness.

Assess how a quasi-experimental, non-randomized design evaluates treatment effects on depression using pre-post measures, reliability checks, validity considerations, and external validity, with cautious optimism about real-world effectiveness.

Case study guides interpreting results from a non-experimental design with numerical meditation frequency and happiness as variables, focusing on reliability, measurements, validity, explanations, confounders, and external validity.

Interpret a practical A/B test on two landing pages, assess reliability and validity, determine if landing page B outperforms A, and discuss potential confounders and external validity.

Interpret a null result by disentangling no effect versus problems, then check 95% power for small or medium effects, reliability, validity, manipulation, and external validity.

If your study reveals a major limitation, change your research question to what you can answer. For example, findings show people form trustworthiness impressions spontaneously from faces, not automatically.

Learn to interpret findings by assessing reliability, accurate measurements, alternative explanations and external validity, and to check methodological problems and limitations with precise, cautious conclusions.