Line Chart (Time Series)

A line chart connects measurements at successive time points, implying that values change continuously between those points. This implication is only valid when the underlying process is continuous and the time points are sufficiently dense. For sparse longitudinal data —measurements at 3 or 5 time points —the line is a visual interpolation that may or may not reflect reality. This must be disclosed.

STEP 1 —When to Use

Use a line chart when: measurements are made at multiple time points (longitudinal or repeated measures); the underlying process is continuous in time; you want to show the trajectory of change, not just the endpoint; you have within-subject data where the within-subject correlation across time points is the primary statistical question.

Error bar type for time series: within-subject SEM (or CI) is preferred over between-subject SEM for longitudinal designs —it shows the precision of the within-subject trajectory, not the variability across subjects. Standard between-subject SEM in a repeated-measures design can be misleadingly large even when the within-subject pattern is highly consistent.

[Image placeholder: Line chart with 2 groups across 5 time points. Left: between-subject SEM error bars —large, overlapping, the groups look similar despite a consistent within-subject pattern. Right: same data with within-subject SEM —smaller error bars, the consistent within-subject treatment effect is clearly visible. Both show the same data.]

Between-subject SEM (left) overstates variability in a repeated-measures design. Within-subject SEM (right) correctly shows the precision of the within-subject trajectory.

STEP 2 —Journal Requirement

Error bar type and direction must be labeled: "Error bars: within-subject SEM" or "ﾂｱ95% CI (within-subject)".
n per time point per group must be stated (may decrease over time in longitudinal studies due to dropout).
Whether lines connect group means or individual trajectories must be stated.
Statistical analysis for longitudinal data: repeated-measures ANOVA or linear mixed model —not separate t-tests at each time point (see Repeated Measures Model Suggestion, 4.5.4).
Continuity disclosure: if time points are sparse, a note that "lines connect measured time points and do not represent interpolated values" is appropriate.

STEP 3 —Licklider Implementation

Input

Time variable (numeric or ordered factor)
Outcome variable
Group variable and subject ID variable (for within-subject error bars)
Error bar type: within-subject SEM (default for repeated measures), between-subject SEM, 95% CI

Output

Group mean lines connecting time points (distinct line types and/or colors per group)
Error bars at each time point (within-subject or between-subject, as declared)
Individual trajectories optional (thin lines per subject, semi-transparent)
n per time point labeled (if dropout occurs)
Statistical model results annotated (F statistic, p-value for time xgroup interaction)
Continuity disclosure in legend

[Image placeholder: Licklider line chart with 2 groups x4 time points. Group mean lines with distinct colors and line types. Within-subject SEM error bars. Individual subject trajectories as thin semi-transparent lines. Time xGroup interaction result annotated: F(3,87)=4.2, p=0.008. Legend: "Error bars: within-subject SEM. Lines connect group means at measured time points."]

Licklider line chart: within-subject SEM, individual trajectories, time xgroup interaction result, and continuity disclosure in legend.

STEP 4 —Draft Output (Draft / Needs review)

Line chart showing group means ﾂｱ within-subject SEM across four time points
(n = 16 per group; no dropout). Individual subject trajectories are shown
as thin lines. Lines connect group means at the measured time points and
do not represent interpolated values. A linear mixed model (outcome ~ time
* treatment + (1 | subject)) found a significant time xtreatment interaction:
F(3, 87) = 4.2, p = 0.008, indicating differential trajectories across groups.
Error bars represent within-subject standard error of the mean.