Getting started
Three steps from the file to the report
1. Prepare the data
An Excel file in the site's standard template: row = participant, column = variable, first row = headers – and optionally a codebook sheet "legend" with the description, variable type and coding of each column. There you will also find an empty template, a sample file and a checker for your file. Those who work in SPSS start on the From SPSS to Excel page and get the file ready, including the legend.
2. Choose a calculator
By the question and the variables – Table 1 below helps you choose. The calculators and guides are open to everyone, with no registration. Drag in the file, check the variables, click "Compute". Every calculator has a sample file to download (in Hebrew and English) so you can practice before uploading data.
3. Get the output
APA-style tables, a chart, a report sentence in English, "Interpreting the result" in words – and download as Word or HTML. The syntax file produces the same numbers in SPSS, for anyone who wants to check.
The research path – from the question to the report
The site is organized by the stages of research. Those who already know what they need skip straight to the calculator; those who are starting out go through stage by stage.
Stage 1 – Planning the research
What is asked, what is measured and whom – before a single data point is collected.
Stage 2 – Preparing the data
The file is converted to the site's standard template, missing values and outliers are handled, and the scales receive a score.
Stage 3 – Describing the sample
Who the participants are and what each variable looks like – before any test.
Stage 4 – Testing the hypotheses
Differences between groups and measurements, associations between variables, prediction and mediation.
Stage 5 – Reporting
Every calculator ends with an APA-style table, a report sentence in English and a Word file – and SPSS syntax for cross-checking; the guidelines page says what must be in each sentence, one generator assembles a results chapter from the outputs, and the other builds a results chapter, an abstract and an executive summary from a form.
Choosing the test by the question and the variables
| What you want to test | The variables | The test (link to the page) |
|---|---|---|
| Describe a single variable – distribution, mean, standard deviation | Any type | Frequencies and descriptive statistics |
| A table report for all the survey questions – with a codebook and a data check | Single response, multiple response, question batteries (batteries), conditional questions, background variables; several response files; weighting; a research package for reproduction | Survey processing: codebook, checks and tables |
| The same tables by group – gender, age group, region – with chi-square and V for each crosstab | Survey questions × background variables (column headers); also weighted | Survey processing: crosstab breakdowns |
| Association between two quantitative variables | quantitative × quantitative | Pearson and Spearman correlations · Scatter plot and regression line |
| Association between two categorical variables | categorical × categorical | Chi-square test of independence |
| Is the variable normally distributed – before a t-test, ANOVA or regression | quantitative (for the whole sample or by group) | Normality testing |
| Does a single mean differ from a fixed value – the scale midpoint, a norm | quantitative vs. a reference value | One-sample t-test |
| Difference in the mean between two separate groups | categorical (2 groups) → quantitative | Independent-samples t-test |
| Difference between two separate groups – ordinal variable, small sample or outliers | categorical (2 groups) → ordinal/quantitative | Mann-Whitney U test |
| Difference between two measurements of the same participants | two quantitative measurements | Paired-samples t-test |
| Difference between two measurements of the same participants – ordinal variable, small sample or outliers | two ordinal/quantitative measurements | Wilcoxon signed-rank test |
| Difference in the mean between three or more groups | categorical (3+ groups) → quantitative | One-way ANOVA |
| Difference between three or more groups – ordinal variable, small sample or outliers | categorical (3+ groups) → ordinal/quantitative | Kruskal-Wallis test |
| Difference between three or more measurements of the same participants, or change over time compared between groups | several quantitative measurements (+ group) | Repeated-measures ANOVA and mixed design |
| Difference in the mean between groups, taking into account a continuous variable measured beforehand – for example a "before" score | categorical → quantitative, with a covariate | ANCOVA – analysis of covariance |
| The effect of two or more factors and of the interaction between them, also with blocks; sample size and random order for the experiment | several categorical factors → quantitative | Design of experiments and multifactor ANOVA |
| The effect of a factor whose levels are a sample – schools, classes, days – or of a factor nested in another factor; split-plot | fixed and random factors → quantitative | Random factors, nested designs and split-plot |
| Which combination of quantitative factors – dose, time, length – gives the best outcome | several quantitative factors → quantitative | Response surface methodology |
| Predict a quantitative variable from one or several predictors | predictors → quantitative | Linear regression |
| Predict a binary outcome (yes/no) | predictors → binary | Binary logistic regression |
| How well a score distinguishes between those who have a certain condition and those who do not – screening, diagnosis, prediction | quantitative measure → binary (gold standard) | ROC analysis |
| What several studies that examined the same question say together | an effect size from each study | Meta-analysis |
| How much time passes until an event, when some of the participants have not yet experienced it | group or predictors → time to event (with censoring) | Survival analysis |
| Moderation – does the association between X and Y depend on a third variable; simple or moderated mediation in the PROCESS style | predictor, moderator and mediators → quantitative | Moderation and mediation |
| Mediation – X affects Y through M | several equations in one model | Path analysis and mediation |
| How many factors stand behind the questionnaire items and which items belong to each – when the structure is not set in advance | questionnaire items (quantitative or Likert) | Exploratory factor analysis |
| Do the questionnaire items group into factors set in advance | questionnaire items (several scales) | Confirmatory factor analysis |
| Does one concept affect another, when both are measured by several items | questionnaire items (several scales) + paths | Structural equation modeling |
| Mediation – X affects Y through M, when each is measured by several items | questionnaire items (several scales) + paths | Latent mediation |
| Does the questionnaire measure the same thing in several groups, and in what do the groups differ | questionnaire items + categorical (2+ groups) | Multigroup comparison and measurement invariance |
| Do the respondents divide into types that were not defined in advance | several quantitative variables (scale scores) | Cluster analysis |
| Do the scale items measure one thing | questionnaire items | Cronbach's alpha and scale score |
| How much do the coders agree | codings by several coders | Cohen's kappa / Krippendorff's alpha |
Note. "Quantitative" – a scale score, age, hours; "categorical" – a group, gender, an answer from a list. A single rating item (1-5) is treated on the site as an ordinal variable – Spearman and median; a multi-item scale score – as an interval variable. Details on the page on variables and levels of measurement.