GM26.1

IF SUBGROUP · {{ sgTabName }}

1 · KEEP ROWS WHERE…

2 · …IS ONE OF

2 · …IS BETWEEN

to {{ sgRangeHint }}

{{ sgLevelNote }}

ALL VARIABLES

VARIABLE
TYPE
VALID
MISS
LVLS
SUMMARY

{{ r.name }}

{{ r.valid }}

{{ r.miss }}

{{ r.levels }}

{{ ovNote }}

DATA EDITOR

TYPE TO EDIT Click any cell and type · Enter moves down, Tab moves right · blank or "." = missing · click a column name to edit that variable in DATA

#
NAME
TYPE
VARIABLE LABEL (CLICK TO EDIT)
MISSING
LEVELS
VALUE LABELS

{{ row.n }}

{{ row.name }}

{{ row.miss }}

{{ row.levels }}

VALUE LABELS — {{ vlVarName }}

VARIABLES IN MATRIX

{{ pxLabel }}

{{ pyLabel }}

EACH COUNTRY SHOWS ITS

AREA

YEAR

COLOUR SCALE

VALUES ON THE MAP

ALSO PLOT (UP TO 3 MORE)

SEPARATE LINES BY (OPTIONAL)

YEARS

LINE WIDTH

POINTS

LINE STYLE

CONFIDENCE BAND

{{ pgsLabel }}

LEVEL {{ lv.n }}

THEN SPLIT BY

KEEP WHICH VALUES?

e.g. pais → prov · one line per combination · up to 8 lines

PLOT COLOR

PLOT

{{ plotEmptyText }}

Axes

{{ axPlot.yName }}to

{{ axPlot.xName }}to

{{ plotNote }}

PAIRWISE CORRELATIONS (PEARSON R)

{{ h.t }}

{{ row.label }}

EACH PAIR USES ALL ROWS COMPLETE FOR THAT PAIR (STATA PWCORR). ORANGE = SIGNIFICANT AT CURRENT ALPHA.

WHAT THIS MEANS

Keep in mind: {{ iMx.tip }}

VARIABLE

READ AS

PLOT COLOR

Pick a variable to profile

Type a name above, or click any variable in the left pane — you'll get descriptives and charts instantly.

DESCRIPTIVES

{{ c.k }}

{{ c.v }}

HISTOGRAM

Axes

{{ axHist.yName }}to

{{ axHist.xName }}to

BOX PLOT

Axes

{{ axBox.yName }}to

{{ axBox.xName }}to

{{ chartNote }}

WHAT THIS MEANS

Keep in mind: {{ iOne.tip }}

FREQUENCIES

LEVEL
COUNT
PCT
CUM

{{ f.count }}

{{ f.pct }}

{{ f.cum }}

{{ freqNote }}

WHAT THIS MEANS

Keep in mind: {{ iOne.tip }}

BAR CHART

CROSSTAB WITH ANOTHER VARIABLE

Rows

{{ oxRowName }}

By (columns)

{{ oxCorner }}

{{ h.label }}

{{ cell.t }}

{{ t.k }}

{{ t.v }}

T-TESTS

Compare means. Choose the design that matches how your observations were collected.

NUMERIC OUTCOME

Choose a numeric outcome to configure the test. You can also click a variable in the list.

T-TEST RESULTS · {{ ttX }}

COMPARE THE MEAN WITH THIS VALUE

GROUP VARIABLE

GROUP A

GROUP B

NEW 0/1 VARIABLE

= 1 if

==!=>>=<<=

= 0 if

==!=>>=<<=

{{ ttMsg }}

SECOND VARIABLE (MEASURED ON THE SAME ROWS)

{{ ttErr }}

{{ c.k }}

Axes

{{ axTt.yName }}to

{{ axTt.xName }}to

APA{{ ttApa }}

WHAT THIS MEANS

Keep in mind: {{ iTt.tip }}

Reproducible R / Stata code

{{ ttCmd }}

Classical t-tests require unweighted observations. Clear weights in Settings, or use Regression with an appropriate survey design.

{{ twoXLab }}

{{ twoYLab }}

{{ twoKindLabel }} {{ pairN }} PAIRS

Analysis

{{ twoNote }}

Pick two variables

numeric × numeric → correlation groups × numeric → t-test / ANOVA groups × groups → crosstab

CORRELATION

MEASURE
VALUE
P
N

{{ cr.v }}

{{ cr.p }}

{{ cr.n }}

APA {{ apaCorr }}

SCATTER + FIT LINE

Axes

{{ axScatter.yName }}to

{{ axScatter.xName }}to

{{ scatterNote }}

WHAT THIS MEANS

Keep in mind: {{ iCorr.tip }}

GROUP MEANS

GROUP
N
MEAN
SD
MED

{{ g.n }}

{{ g.mean }}

{{ g.sd }}

{{ g.median }}

{{ testTitle }}

{{ t.k }}

{{ t.v }}

APA {{ apaTest }}

BOX PLOTS BY GROUP

Axes

{{ axGbox.yName }}to

{{ axGbox.xName }}to

{{ groupCapNote }}

WHAT THIS MEANS

Keep in mind: {{ iTest.tip }}

CROSSTAB — COL %

{{ ctYName }}

{{ h.label }}

{{ cell.t }}

CHI-SQUARE TEST OF INDEPENDENCE

{{ t.k }}

{{ t.v }}

{{ ctChiNote }}

APA {{ apaChi }}

COMPOSITION BY GROUP

WHAT THIS MEANS

Keep in mind: {{ iCt.tip }}

TIME SERIES — 1 · SET UP THE DATA

Define the calendar, unit and measurement source before estimation. Keep the original frequency by default.

{{ tsTimeLabel }}

UNIT (PANEL) VARIABLE · OPTIONAL

OUTCOME (NUMERIC)

EACH ROW IS

Records (e.g. survey respondents) are collapsed to one value per period — and per unit when a unit variable is set — like Stata’s collapse. The model is fitted to those values.

OBSERVATION ID · OPTIONAL

WEIGHT · OPTIONAL

COLLAPSE EACH PERIOD WITH

TIME INFORMATION

ORIGINAL FREQUENCY

ANALYSE

PERIOD RANGE

MEASUREMENT / SOURCE FILTER · OPTIONAL

ANALYSIS FREQUENCY

OUTCOME PREPARATION

COUNT NUMERATOR

PERIOD-LEVEL POPULATION

RATE PER

{{ tsAggregationHint }} Missing dates and missing outcomes are not filled or interpolated.

WHICH UNITS? · UP TO 8 AS SEPARATE PICKS · ANY NUMBER WITH ALL UNITS OR AVERAGE

2 · THE EVENT (SHOCK / INTERVENTION)

when did something happen? leave the event blank to fit a simple trend

EVENT DATE / PERIOD

WHAT CHANGES AFTER IT

CONTROL FOR LAST PERIOD

More event options — name the event, or give each unit its own event date

EVENT NAME

OR EACH UNIT’S ENCODED EVENT PERIOD

DIFFERENT YEAR FOR SOME UNITS? (BLANK = USE THE EVENT YEAR ABOVE)

ERR // {{ tsErr }}

3 · RESULTS

{{ c.k }}

{{ c.v }}

TERM
ESTIMATE
SE
T
P

{{ r.term }}
{{ r.b }}
{{ r.se }}
{{ r.t }}
{{ r.p }}

D · INTERRUPTED TIME SERIES MODEL

Axes

{{ axTs.yName }}to

{{ axTs.xName }}to

DOTS = DATA · THICK LINE = MODEL · DOTTED = WHAT THE MODEL EXPECTS WITHOUT THE EVENT · DASHED RED = EVENT

BEFORE vs AFTER THE EVENT (AVERAGE OF THE OUTCOME)

{{ r.u }}
{{ r.E }}
before {{ r.b }}
after {{ r.a }}
{{ r.d }}

Diagnostics — describe the series, stationarity (ADF), autocorrelation (ACF / PACF)

A · DESCRIBE THE SERIES

UNIT
PERIODS
FROM–TO
MEAN
SD
FIRST → LAST
Δ / PERIOD

{{ r.u }}
{{ r.T }}
{{ r.span }}
{{ r.mean }}
{{ r.sd }}
{{ r.fl }}
{{ r.sl }}

B · STATIONARITY — AUGMENTED DICKEY–FULLER ({{ tsAdfMode }})

UNIT
ADF τ
LAGS
5% CRITICAL
P
VERDICT

{{ r.u }}
{{ r.tau }}
{{ r.lags }}
{{ r.cv }}
{{ r.p }}
{{ r.v }}

C · AUTOCORRELATION (ACF) AND PARTIAL AUTOCORRELATION (PACF)

ACF: how much each period resembles earlier ones. PACF: the same, after removing the in-between periods. A PACF with one big bar at lag 1 that then drops off points to an AR(1) process — switch on AR1 in step 2.

{{ tsAcfNote }}

Stata / R script for this analysis

{{ tsCmd }}

WHAT THIS MEANS

Keep in mind: {{ iTs.tip }}

Add the event variables to the dataset (to use them in Regression and other tabs)

4 · ADD THE EVENT VARIABLES TO THE DATASET

so you can use them in Regression, Plots and every other tab

CREATE

CODE THE EVENT FOR

{{ tsMsg }}

STRUCTURAL EQUATION MODEL

hidden ideas (factors) measured by several items, plus paths between them

How it works: some ideas (like trust or well-being) can't be measured with one question. Make a factor for each idea, give it 3 or more numeric items that measure it, then (optionally) say which factor predicts which.

1 · FACTORS AND THEIR ITEMS

=1 marks the item whose loading is fixed to 1. Each factor needs one: it gives the factor its scale (the factor is measured in that item's units), which the model needs to be identified. It starts on the first item; click =1 on another item to use it instead. Stata and lavaan also fix the first item by default.

is measured by

2 · HOW ARE THE FACTORS CONNECTED?

one row per pair — correlated (↔), one predicts the other (→), or not related

ESTIMATION METHOD

{{ semCmdPreview }}

ERR // {{ semErr }}

ESTIMATES

STD = standardized · loadings above .70 are strong, below .40 weak

TERM
ESTIMATE
SE
Z
P
STD

{{ r.term }}

{{ r.term }}
{{ c.t }}

MODEL FIT

{{ c.k }}

Rules of thumb: CFI and TLI ≥ .95 good (≥ .90 acceptable) · RMSEA ≤ .06 good (≤ .08 acceptable) · SRMR ≤ .08 good. A significant χ² is common with large samples.

PATH DIAGRAM

WHAT THIS MEANS

Keep in mind: {{ iSem.tip }}

MODEL

STANDARD ERRORS

{{ depHead }}

PREDICTORS (X) — PICK FROM THE LIST · CATEGORICAL (C) VARIABLES ARE DUMMY-CODED · CLICK A CHIP'S N / C BADGE TO SWITCH

none yet — pick one from the list

INTERACTIONS (OPTIONAL)

does the effect of one predictor depend on another? e.g. income × gender

{{ regEmptyText }}

ERR // {{ regError }}

WORKING {{ regBusyInfo.label }}

{{ regBusyInfo.note }}{{ regBusyInfo.secs }} — you can keep scrolling while this runs

MODEL FIT

{{ s.k }}

{{ s.v }}

COEFFICIENTS

TERM
COEF.
STD.ERR.
{{ zHead }}
P
[95% CI]
{{ lastHead }}

{{ c.b }}

{{ c.se }}

{{ c.t }}

{{ c.p }}

{{ c.ci }}

{{ c.beta }}

{{ regNotes }}

PUBLICATION TABLE — POLITICAL SCIENCE STYLE

TITLE

{{ rtMsg }}

WORKING {{ mgBusyInfo.label }}

{{ mgBusyInfo.note }}{{ mgBusyInfo.secs }} — you can keep scrolling while this runs

MULTICOLLINEARITY — VARIANCE INFLATION FACTORS

VARIABLE
VIF
1/VIF

{{ v.inv }}

Mean VIF

{{ vifMean }}

MARGINS PLOT — PREDICTED {{ mgWhat }} AS ONE PREDICTOR CHANGES

X AXIS

SEPARATE LINES BY

OTHER PREDICTORS

PROBABILITY OF CATEGORY

ERR // {{ mgErr }}

Axes

{{ axMg.yName }}to

{{ axMg.xName }}to

{{ mgNote }}

{{ mgCmd }}

WHAT THIS MEANS

Keep in mind: {{ iMg.tip }}

COEFFICIENT PLOT — DOT = ESTIMATE · WHISKER = CONFIDENCE INTERVAL · ORANGE = SIGNIFICANT

Axes

{{ axCoef.yName }}to

{{ axCoef.xName }}to

ASSUMPTION CHECKS (RVFPLOT)

RESIDUALS VS FITTED

Axes

{{ axRvf.yName }}to

{{ axRvf.xName }}to

RESIDUAL DISTRIBUTION

Axes

{{ axResid.yName }}to

{{ axResid.xName }}to

APA {{ apaReg }}

WHAT THIS MEANS

Keep in mind: {{ iReg.tip }}

KEY RESULT{{ keyLine }}

Data information

T-TESTS

Plot options

Plot preview