jstats Quick Start

From here on, jstats is loaded and you’re working in your Project. This page assumes you’ve installed jstats (see Install jstats) and loaded it with library(jstats).

Before the tour, one quick contrast. jstats ships with a practice dataset called community; suppose you just want to describe the respondents’ ages. In base R you’d reach for a handful of separate commands:

mean(community$Age)
sd(community$Age)
range(community$Age)
[1] 40.66
[1] 11.68027
[1] 18 71

Each prints on its own — bare, unlabeled numbers you have to keep straight in your head — and you still don’t know how many respondents there are, or whether any ages are missing. jstats does the whole thing in one:

jdesc(community, Age)
Descriptive Statistics

Variable  Total  Non_missing  Min  Max   Mean     SD
--------  -----  -----------  ---  ---  -----  -----
Age         100          100   18   71  40.66  11.68

One command, one tidy table — count, missing, range, center, and spread together, in a layout you’ll recognize if you’ve used commercial statistical software. That’s the whole idea behind jstats, and the rest of this page builds on it.

Meet community

That practice dataset, community, holds 100 made-up community-survey respondents, with income, education, age, a wellbeing score, a few yes/no items, a region, and a five-item attitude scale. It’s available the moment the package is loaded — you don’t have to open or import anything — which is the point of a shipped dataset and what lets these examples just work.

Your first jstats command

You met functions in the console earlier with sqrt(). jstats functions work the same way — same shape, same parentheses. Let’s look at how respondents are spread across regions. This is analysis you’ll want to keep, so type it in your script (the Editor) this time, then run it:

jfreq(community, Region)

That’s the same function-and-object idea as before, now doing real work. You named the function (jfreq, for frequencies), and inside the parentheses you gave it two things: the dataset (community) and the variable you want counted (Region). jstats prints a tidy frequency table — how many respondents fall in each region, with percentages.

(Remember R is case-sensitive: it’s Region, not region.)

(And a reminder from RStudio Orientation: you can tack a note onto any line after a # — jfreq(community, Region) # counts by region — and R ignores everything after it.)

The examples just run

Open the help page for this function, and here it is:

?jfreq

Scroll to the bottom and you’ll find worked examples. Copy one, paste it into your console, and it runs immediately — because the data it uses are already loaded. No file to open, no path to set.

A fair word on this, especially if you’re coming from commercial software: runnable help examples aren’t unique to R. Stata and SAS ship practice data too. What’s genuinely worth knowing is this. First, R’s documentation examples are build-verified — every example is actually executed when the package is checked, so they can’t quietly go stale. Second, there’s truly zero setup: community is simply there. Third, one source produces all three — the function, its help page, and its runnable example all come from the same place, so they can’t drift apart. The practical effect: when you’re learning a new function, you can always try it instantly on data that’s already in front of you.

Getting help on any function

You just used ?jfreq to open a help page. That works for every jstats function — type ? and the function’s name. There’s also a shortcut: in your script, click on a function’s name and press F1 (Windows) or Fn+F1 (Mac). Either way, the help page opens in the bottom-right pane, with a description, the available options, and worked examples you can run.

What a data frame is

community is a data frame — R’s name for a rectangle of data: rows are cases (here, respondents) and columns are variables (income, age, and so on). It’s the everyday shape your data take in R.

Here’s something important about how R handles it. When community is in use, R is working with a copy held in memory — the version listed in your Environment pane — not the file on disk. You can recode, trim, and experiment freely, and the original file is never touched. The flip side is that changes you make live only in that working copy until you deliberately save them — which is exactly what jstats with your own data covers.

This is a real difference worth pausing on. In SPSS, the data you see in Data View is the working dataset, and saving writes back to that file. R instead keeps a working copy in memory and leaves your original file alone until you explicitly save. The upside is freedom to experiment without fear; the thing to remember is that your changes aren’t on disk until you save them.

Set a default dataset

You named community explicitly in your jfreq call. If you’re going to work with one dataset for a while, you can set it as the default so you don’t have to name it every time. Do that with juse (for use this dataset):

juse(community)

Now rerun the frequency table without naming the dataset:

jfreq(Region)

Same result. jstats fills in community for you. (Your Environment can hold several datasets at once; juse just picks which one is the current default — no risk to any original.)

Describe the numeric variables

For numeric variables, you usually want means, standard deviations, and the like rather than a full frequency table. That’s jdesc (for describe). Since we’ve set a default with juse, we can name just the variables:

jdesc(Age, WellbeingScore, Income)

jstats prints a compact table of descriptive statistics — one row per variable.

Screen the whole dataset at once

Before any serious analysis, it’s good practice to look the whole dataset over — its ranges, possible outliers, skew, missing values, and what kind each variable seems to be. jscreen (for screen) does all of that in one call:

jscreen()

With the juse default set, a bare jscreen() screens every variable in community and gives you a fast first read on the data’s health. We’re keeping this light here; there’s much more jscreen can tell you, and we come back to it later. (The full depth lives in the help page and the books.)

Your first analysis: a regression

This is the destination of the Quick Start — a working regression.

Let’s ask whether age predicts wellbeing in these data. The function is jlm (for linear model — the technical name for ordinary regression). Notice something new on the left:

Results <- jlm(WellbeingScore ~ Age)

A regression asks a simple question: as one thing changes, does another tend to change along with it, and by how much? Here we’re asking whether wellbeing tends to rise (or fall) as age goes up. The output puts a number on that tendency, and on how confident we can be that it’s a real pattern rather than chance.

Two new things in that command.

The ~ (tilde) is how R writes a model: outcome on the left, predictor on the right. Read WellbeingScore ~ Age as “wellbeing explained by age.” (We didn’t name the dataset because juse(community) is still in effect; jstats notes which dataset it used at the top of the output.)

The <- (typed as a less-than sign and a hyphen) is the assignment operator — it means “store the result in a name.” Here it saves the whole analysis into an object called Results. This is the function-and-object idea finally paying off: you’ll hand Results to other jstats functions later — for formatted tables, for example — without re-running the regression.

A reassurance, because it’s a common worry: saving the result with <- does not hide your output. jstats analysis functions always print their results to the console and return the object. You get the full table on screen and a saved Results to reuse. Nothing is suppressed.

Read the output

Look at the printed output. The key pieces:

  • The coefficient for Age — how much wellbeing changes for each additional year of age. A positive number means older respondents tend to score higher.
  • Its p-value — whether that relationship is statistically distinguishable from zero.
  • R-squared — the share of the variation in wellbeing this one predictor accounts for.

In these data, age is a modest but real predictor of wellbeing. That’s your first regression — run, read, and saved.

NoteWhere to go next

Age is only a gentle predictor here. A stronger story in these data is income and wellbeing — but income has some respondents who declined to answer, which is exactly the kind of missing-value handling the next part is about. That’s our natural next step: jstats with your own data.