Random effects meta analysis, 18 trials, 2,140 adults

Meta analysis poster example

A meta analysis poster carries a pooled estimate, so it has to show the trials that built it and how much they disagree. This A0 portrait example pools 18 randomised trials of structured exercise in 2,140 adults with depressive symptoms: standardised mean difference minus 0.42, 95% CI minus 0.54 to minus 0.30, I squared 46%.

Create a poster with OneCraftA0 Portrait, printed at 841 by 1189 millimetres

The whole board

The poster at full size, exactly as it prints. Every number, citation and caption on it was written for this example, so the layout is being judged on real content.

WU
How much does structured exercise reduce depressive symptoms? A random effects meta analysis of 18 trials and 2,140 adults
C. Abernethy ¹, Y. Mensah ¹, F. Ringvold ²
1 Evidence Synthesis Group, Wrenholt University · 2 School of Exercise Science, Wrenholt University
Abstract
We pooled 18 randomised trials of supervised exercise against usual care or a waiting list in 2,140 adults with depressive symptoms. The pooled standardised mean difference was minus 0.42, 95% CI minus 0.54 to minus 0.30, a moderate reduction. Heterogeneity was moderate, I squared 46%. Shorter programmes reported larger effects, which may partly reflect their smaller size. Every one of the 18 trials pointed the same way.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
All
-1-0.75-0.5-0.250
Figure 1. Effect size, SMD, for trials 1 to 18 in date order, 2009 to 2025, then the pooled estimate, All. Below 0 favours exercise.
Source: 18 included trials, random effects pooling.
METHODS
Search. Three databases searched to March 2026: 412 records, 97 duplicates removed, 315 titles screened.
Screening. Two reviewers read 54 full texts independently and agreed on 18 trials; 36 were excluded, most for no control arm.
Eligibility. Adults with depressive symptoms on a validated scale, supervised exercise for 4 weeks or more, usual care or waiting list control.
Outcome. Standardised mean difference in symptom score at the end of the programme. Negative values favour exercise.
Synthesis. Random effects pooling with the between trial variance estimated from Q; heterogeneity as I squared.
REFERENCES
1.Holloway, R. and Ademola, B. (2022) Supervised exercise for low mood: a trial in primary care. Journal of Activity and Mental Health, 8(2), 77 to 91.
2.Strandberg, I. (2024) Small trial effects in exercise and depression research. Evidence Synthesis Methods, 12(1), 14 to 33.
3.Abernethy, C. (2025) Search strategy for exercise and mood trials. Wrenholt Synthesis Note 5, 2025.
Results
All 18 trials favoured exercise, with effects from minus 0.03 to minus 0.95. The pooled estimate across 2,140 participants was minus 0.42, 95% CI minus 0.54 to minus 0.30, with I squared 46% and a between trial variance of 0.03. The cumulative estimate settled near minus 0.42 after the ninth trial and moved by no more than 0.03 as the last nine were added. Leaving out any one trial kept the estimate between minus 0.39 and minus 0.44. No trial under 8 weeks had more than 98 participants; the other 13 had a median of 130.
Discussion
A pooled effect of minus 0.42 is moderate and consistent in direction. The duration pattern runs the wrong way for a dose effect, and since the shortest trials were also among the smallest, small trial bias is the likelier reason. Most trials used a waiting list, so part of the effect may be attention rather than exercise itself.
-0.42
pooled effect on depressive symptoms, SMD
95% CI -0.54 to -0.30; I squared 46%; 18 trials, 2,140 adults
Key finding
Exercise cut depressive symptoms by a moderate 0.42 standard deviations across 18 trials.
18 trials, 2,140 adults, every trial favouring exercise
Pooled SMD minus 0.42, 95% CI minus 0.54 to minus 0.30
Moderate heterogeneity, I squared 46%
Short programmes larger, but they were also small trials
Subgroups by duration
Under 8 weeks: 5 trials, 430 participants, SMD minus 0.73, 95% CI minus 0.93 to minus 0.54. 8 to 12 weeks: 10 trials, 1,208 participants, minus 0.37, minus 0.50 to minus 0.23. Over 12 weeks: 3 trials, 502 participants, minus 0.21, minus 0.39 to minus 0.03. Together 18 trials and 2,140 participants. The interval for the shortest group sits clear of the other two.

Block by block

What each block on the board is for, in the order a reader walks it.

Title band and three authors
The title asks how much exercise reduces symptoms and gives the trial and participant counts. Three authors from the fictional Wrenholt University.
Abstract
One full width paragraph with the pooled estimate, its interval, the heterogeneity and the duration pattern.
Hero figure
A full width line plots each trial effect in date order, 2009 to 2025, and ends at All, the pooled minus 0.42.
Methods
Five entries: the search from 412 records to 315 screened, the screening of 54 full texts to 18 trials, eligibility, the outcome and the pooling model.
Results and discussion
Results give the range of effects, the pooled estimate, the cumulative estimate after each trial and a leave one out check. The discussion weighs dose against small trial bias.
Key number, key finding and takeaways
The stat block gives minus 0.42 with its interval and I squared; the callout and four takeaways restate it in plain words.
Subgroups by duration and references
The subgroup table the row asked for is written as three entries, with trials, participants and estimates adding to 18 and 2,140. Three references sit under the methods.

What makes this board work

The trials are on the board

All 18 effects are plotted, from minus 0.03 to minus 0.95, beside the pooled minus 0.42. A pooled number alone asks for trust; the spread lets a reader judge the I squared of 46% by eye.

The subgroups add up

Under 8 weeks: 5 trials, 430 people. 8 to 12 weeks: 10 trials, 1,208. Over 12 weeks: 3 trials, 502. That is 18 trials and 2,140 participants, the totals in the title.

It argues against its own neat story

Short programmes show the biggest effect, minus 0.73, which looks like a finding. The board notes that none of those trials had more than 98 participants, so small trial bias is the likelier reason.

Questions people ask

What makes a meta analysis poster different from a systematic review poster?

The pooled estimate. A systematic review board can describe and tabulate studies without combining them; a meta analysis board reports one combined effect with its interval and a measure of heterogeneity. Here that is minus 0.42, 95% CI minus 0.54 to minus 0.30, and I squared 46%, given in the abstract, the stat block and the results.

Can I draw a forest plot on this poster?

Not as a true forest plot. Figures here draw one of ten chart kinds, none with error bars, and this design fixes its hero figure as a line. The board plots each trial effect in date order with the pooled value last, and keeps the intervals in the text and stat block. Bring the full forest plot as a handout.

How should a meta analysis poster report the search?

As counts that reconcile: records found, duplicates removed, titles screened, full texts read, trials included. This board goes from 412 records, less 97 duplicates, to 315 screened, 54 full texts and 18 trials, in two method entries. A full PRISMA diagram is worth it on a board with a flow block.

Is a leave one out check worth the space?

One sentence of it is. Dropping each trial in turn kept the pooled estimate between minus 0.39 and minus 0.44, which tells a reader no single trial drives the result. It costs a line in the results and answers a question every reviewer at a poster session asks.

What does I squared of 46% tell a reader?

That a little under half of the spread between trial results reflects real differences between trials rather than chance. It is usually called moderate. Report it beside the pooled estimate, as this stat block does, and then give the subgroup or sensitivity analysis that tries to explain it.

Should the cumulative estimate appear on a meta analysis poster?

A sentence of it helps when trials were added over many years. On this board the pooled value settled near minus 0.42 after the ninth trial and moved by no more than 0.03 as the last nine were added. That tells a reader one more trial of a similar size is unlikely to change the answer.

Build your own in about a minute

The button below opens the generator with this use case already described. Change the wording to match your own, generate, then edit anything you like.

Make one like this

Other poster examples

Controlled experiment poster example

A controlled experiment poster has to show that one thing changed and everything else was held fixed. This A0 portrait example grows 288 lettuce plants in 12 trays under three LED spectra for five weeks in one chamber. From the same 0.8 g start, the red heavy spectrum ends at 118 g a plant against 96 g under balanced white, 23% more.

Qualitative study poster example

A qualitative study poster has to let the participants speak in a few words and still show how far each theme reached. This A0 portrait example reports 36 interviews with unpaid family carers about getting respite: four themes, each with one quote under 30 words and a count, led by respite as a fight in 28 of 36 interviews.

Simulation study poster example

A simulation study poster must say what the model assumes as clearly as what it found. This A0 landscape example runs an agent based evacuation model of a 1,200 seat hall under four exit layouts, 500 runs each. Adding a fourth exit at the rear cuts mean clearance from 212 to 168 seconds, 21% faster, and the 95th percentile from 251 to 192. Four model limits sit right beside that headline.

Want the steps in the builder? Read Add charts and diagrams, then choose the template, theme and size. For everything this generator can do, see the poster maker.

Sources

Written and checked by the OneCraft team. Last checked .