77%
of consumers selected email as an ideal channel for receiving retail promotions.
Email Marketing Statistics2026 DataROI + Benchmarks100+ Statistics
Email Marketing Statistics 2026 — 36:1 Is Not the Whole Story
Email is still one of marketing’s highest-leverage owned channels, but the familiar “$36 back for every $1” line is no longer a useful universal benchmark. Current data shows a wider reality: ROI depends on measurement, message type, automation maturity, inbox placement, industry and whether marketers optimize for outcomes rather than opens.
This 2026 guide brings the latest email marketing benchmarks, email marketing ROI data, deliverability research, AI adoption figures, consumer preferences and regional conversion rates into one decision-focused view.
75%
more likely to achieve ROI above 45:1 among advanced AI adopters.
87.2%
global inbox placement rate in 2025—materially below “delivered” rates.
93%
of surveyed consumers said they check email every day.
30.7%
2025 ecommerce campaign open rate in Omnisend’s annual report, up from 2024.
99.2%
DMA’s 2026 UK delivered-rate headline—proof that “accepted” is not the same as “inbox.”
📋 Table of Contents
01The ROI Headline in Context02Email Performance Trends Over Time
03What Actually Drives Underperformance04Benchmarks by Industry
05Budget, ROI & Unit Economics06AI & the New Inbox Layer
07Consumer & Practitioner Data08Regional Email Performance
09What Separates Winners10Visual Data Recap
11Evidence-Based Action Checklist
01 — Overview
The “36:1 ROI” Statistic Needs Context
The headline is directionally right—email can produce exceptional returns—but 36:1 should not be treated as a universal law. Litmus’s 2025 ROI research did not find one dominant return figure; it found a distribution. That is more useful for operators because it makes the gap between average programs and elite programs visible.
35%
of marketing leaders reported $10–$36 returned for each $1 spent.
30%
reported $36–$50 returned for each $1 spent.
5%
reported more than $50 returned for every $1 spent.
21%
did not measure email ROI at all in the 2025 study.
36%
did not measure email ROI in Litmus’s 2023 comparison point.
45:1+
is the threshold Litmus uses to distinguish exceptionally high-ROI programs in its 2026 analysis.
The “so what” is simple: benchmarking against one blended ROI number can hide two management problems. First, a large minority still cannot calculate ROI with confidence. Second, the highest-return programs behave differently—especially around automation, lifecycle content, AI integration and measurement. The relevant question is not whether email can return 36 dollars; it is whether your program can prove incremental revenue and identify which journeys create it.
Context matters
DMA’s 2026 Marketer Email Tracker reports ROI rising from just over £38 per £ spent to almost £41. That does not “contradict” Litmus’s distribution—it measures a different population and methodology, and it is another reason to avoid treating one cross-market ratio as universal.
For SEO and planning purposes, this is the most important myth to correct in 2026: email marketing ROI is not one fixed number. Use category benchmarks to set expectations, then build your own denominator—technology, people, creative, data and promotional cost—and your own outcome model. That produces a number finance can defend, instead of a number copied from an industry infographic.
02 — Trend Over Time
Email Engagement Is Rising—But Opens Are Getting Less Trustworthy
The recent ecommerce trend looks healthy on the surface. In Omnisend’s comparable platform datasets, open rates rose from 2022 through 2025. Yet mailbox privacy, prefetching and AI summaries mean open rate is becoming more of a diagnostic signal than an outcome metric. Click and conversion behavior now deserves more weight.
| Period | Metric | Result | Interpretation | Source |
|---|---|---|---|---|
| 2022 | Campaign open rate | 22.9% | Baseline before the recent multi-year rise. | Omnisend · 2024 report (2022 comparison) |
| 2023 | Campaign open rate | 25.1% | Up while send volume also expanded. | Omnisend · 2024 report |
| 2024 | Campaign open rate | 26.6% | Fourth consecutive annual increase. | Omnisend · 2025 Ecommerce Marketing Report |
| 2025 | Campaign open rate | 30.7% | Fifth consecutive annual increase. | Omnisend · 2026 Ecommerce Marketing Report |
| 2022 | Campaign click rate | 1.2% | Lower than the 2023 comparison point. | Omnisend · 2024 report |
| 2023 | Campaign click rate | 1.5% | Up alongside a 33% rise in sends. | Omnisend · 2024 report |
| 2024 | Click-to-conversion | 5.9% | Precedes a major improvement in click quality. | Omnisend · 2026 report (2024 comparison) |
| 2025 | Click-to-conversion | 9.0% | A 53% YoY jump despite softer click frequency. | Omnisend · 2026 Ecommerce Marketing Report |
That combination—more opens, fewer casual clicks, more conversions after the click—suggests that the quality of attention matters more than the volume of recorded engagement. An operator who optimizes solely for opens could conclude the channel is improving everywhere. An operator who looks at revenue per delivered message, conversion after click and lifecycle stage sees a much more actionable story.
Deliverability warning
Validity measured global inbox placement at 83.5% in 2024 and 87.2% in 2025, a 3.7-point uplift. In 2025, 6.6% of mail went to spam and 6.1% was missing. A message can be technically “delivered” without ever reaching the primary inbox.
Mailbox-provider variance reinforces the point. Microsoft was Validity’s toughest major mailbox provider in the 2025 dataset at 77.4% inbox placement. Yahoo reached 87.3% inbox placement, with 6.4% in spam and 6.3% missing. These are operational constraints, not creative preferences.
The strategic implication is that the “email open rate statistics” conversation has shifted. Opens still help compare subject lines or detect abrupt changes, but they should not sit alone at the top of an executive dashboard. Pair them with inbox placement, clicks, conversion, revenue, complaints and unsubscribes—and judge each metric against the job that message is supposed to do.
03 — Root Causes
Why Email Programs Underperform
The strongest current evidence points to a mix of audience fatigue, permission failures, weak data foundations and infrastructure gaps. Most of these problems happen before a subject line is written. That is why “write better copy” is rarely the highest-leverage fix for a weak program.
01
Too frequent
41%
Consumers cited excessive promotional frequency as a frustration.
02
Unsolicited
35%
Consumers cited unwanted promotional messages as a major pain point.
03
Irrelevant
33%
Consumers were frustrated by promotions that did not feel relevant.
04
Looks suspicious
80%
ZeroBounce respondents said they may mark email as spam based on appearance alone.
05
No explicit permission
47%
reported an email as spam when they had not explicitly agreed to receive it.
06
Integration friction
33.6%
Marketers named integration with existing systems as their biggest AI-adoption barrier.
07
Skills gap
26.6%
said missing skills were holding AI adoption back.
08
Poor data quality
25.1%
flagged poor data quality as an AI-adoption barrier.
09
Data problem persists
21%
of DMA respondents still named lack of data as a challenge, down from 29% in the prior tracker.
The priority order matters. If people did not consent, do not trust the visual identity, are receiving messages too often, or are poorly segmented because the data stack is fragmented, creative optimization is downstream of the actual problem. Teams often attack symptoms—lower opens, lower clicks—when the cause is relevance, frequency or inbox placement.
This also explains why higher send volume is not automatically a growth strategy. ActionRocket’s analysis of DMA’s 2026 tracker says automated sends now represent 34% of volume and bulk sends 33%; roughly 14% of customers receive four or more emails per week from one brand. The constraint is no longer the ability to send—it is orchestration.
04 — Segment Breakdown
Email Marketing Benchmarks by Industry
Industry averages are useful only when the same metric means the same thing. Omnisend’s 2025 dataset shows why open rate alone can mislead: Wedding has the highest open rate in this comparison, while Games and Finance convert a higher share of total sends. Use the table as a diagnostic baseline, not a universal target.
| Industry | Open rate | CTR | Conversion rate | Source |
|---|---|---|---|---|
| Wedding | 52.43% | 0.77% | 0.04% | Omnisend · 2026 Benchmarks (2025 data) |
| Holidays & Seasonal | 47.62% | 1.17% | 0.03% | Omnisend · 2026 Benchmarks |
| Travel | 44.84% | 1.21% | 0.05% | Omnisend · 2026 Benchmarks |
| Games | 36.85% | 1.13% | 0.19% | Omnisend · 2026 Benchmarks |
| Arts & Entertainment | 34.15% | 1.35% | 0.10% | Omnisend · 2026 Benchmarks |
| Apparel | 31.10% | 0.86% | 0.05% | Omnisend · 2026 Benchmarks |
| Food & Drink | 30.03% | 0.96% | 0.15% | Omnisend · 2026 Benchmarks |
| Health | 28.64% | 0.79% | 0.12% | Omnisend · 2026 Benchmarks |
| Business & Industrial | 28.66% | 1.47% | 0.13% | Omnisend · 2026 Benchmarks |
| Finance | 20.25% | 0.99% | 0.17% | Omnisend · 2026 Benchmarks |
The counterintuitive rows are more useful than the top rows. Finance posts one of the weakest open rates here but one of the strongest conversion rates. Wedding produces huge opens but a low conversion rate. The implication is not that one industry “does email better”; it is that purchase cycle, intent, list composition and message purpose change what a healthy benchmark looks like.
Outcome over optics
Omnisend’s cross-industry campaign averages were 30.41% open rate, 0.74% CTR, 2.44% CTOR, 0.08% conversion and 0.20% unsubscribe. A separate industry table on the same dataset reports a 30.22% average open rate; Omnisend presents these as different benchmark cuts, so this article does not collapse them into one number.
For benchmarking, compare like with like: campaign against campaign, automation against automation, industry against the same industry, and the same attribution window over time. If your own conversion rate improves while open rate falls, that can still be a win. The dashboard should reward commercial intent, not vanity.
05 — Spend & Investment
Budget, ROI and the Economics of an Email Send
Email’s budget case is strong, but the more useful numbers in 2026 are not market-size forecasts. They are the operating ratios that connect budget, measurement and revenue. DMA’s tracker shows growing investment confidence; Omnisend shows why shifting volume into behavior-triggered journeys can change unit economics dramatically.
20%
Email accounts for roughly one-fifth of total marketing spend among marketers covered by DMA’s 2026 tracker.
66%
of marketers expect their email budget to grow over the next 12 months.
£40.53
reported value of email in DMA’s 2026 tracker, up from £36.63.
45%
cite engagement as a key metric for evaluating email objectives.
44%
cite revenue as a key metric—nearly level with engagement.
83%+
of marketers use email across all stages of the customer journey.
Measurement sophistication is uneven. Litmus’s 2025 ROI analysis found 59% using customer-engagement metrics, 44% using MQLs or conversion rates, 41% directly attributing revenue, 39% measuring retention or loyalty, 36% using multi-channel attribution and 30% analyzing brand-awareness metrics.
Unit economics
In Omnisend’s 27,000+ brand benchmark, automated emails produced $3.41 in revenue per send versus $0.155 for campaigns—a 22× gap. Automated conversion was 1.49% versus 0.08%, roughly 19× higher.
Omnisend’s broader annual 2026 report summarizes the automation advantage as 16× revenue per send, while the narrower 27,000-brand benchmark cut yields 22×. We keep both descriptions attached to their source populations rather than pretending the difference is an error. For operators, the direction is what matters: behavior-triggered email is materially more productive per message.
06 — Emerging Technology
AI Is Compressing Production Time—and Moving Into the Inbox
AI’s 2026 email impact is no longer limited to generating copy. It is changing production speed, segmentation, QA, accessibility and the inbox itself. That creates a two-sided optimization problem: marketers are using AI to create messages while mailbox providers use AI to summarize, rank and filter them.
28%
of email teams had reached advanced AI adoption by early 2026.
+28%
Advanced adopters were more likely to deploy an email in under a day than early-stage teams.
+54%
Advanced AI adopters were more likely to follow WCAG standards.
+52%
Advanced adopters were more likely to comply with the European Accessibility Act.
78%
of teams produce and deploy one email in three days or less in 2026.
62%
needed two weeks or more in Litmus’s 2024 comparison point.
The middle step was already visible in 2025: 44% of teams took between a few days and a week. Separately, Litmus reports a 340% increase during 2025 in marketers using generative AI for tasks such as copy, imagery, personalization, analysis and testing. Its 2025 survey also found 70% expected up to half of their email operations to be AI-driven by the end of 2026.
AI maturity, not AI presence
DMA’s 2026 tracker says 44% of marketers use AI for copy generation or templates. The more important finding is that Litmus’s advanced adopters are 75% more likely to achieve ROI above 45:1. Tool adoption is common; workflow integration is the differentiator.
The inbox side is the under-discussed change. Litmus notes that fewer than one-third of marketers have a strategic approach to optimizing for AI-driven inboxes. Authentication pressure is also rising: Gmail applies its bulk-sender requirements at 5,000+ messages per day to Gmail accounts, and Outlook applies its high-volume authentication requirements above 5,000 messages per day to its consumer domains. Yahoo also imposes stricter requirements on bulk senders, but its official Sender Hub explicitly says it does not publish a numeric volume threshold.
Litmus · The Dangers of Generative AI in Email Marketing, 2026 (AI-inbox strategy finding) · Google · Gmail Email Sender Guidelines FAQ · Microsoft · Outlook High-Volume Sender Requirements, 2025 · Yahoo Sender Hub · Bulk Sender FAQ
The operator takeaway is not “use AI everywhere.” It is to apply AI where data and repetitive workflow create leverage—segmentation, QA, personalization, testing and analysis—while preserving human judgment for positioning and trust. Faster production only creates value if the additional messages remain relevant enough to earn inbox placement and action.
07 — People & Practitioners
What Consumers and Marketers Actually Want From Email
Email’s durability is visible in consumer preference data. People still want email, especially for promotions and information they can scan on their own time. What they reject is loss of control: unwanted frequency, irrelevant messages and communication that does not match the context.
Offer type also changes attention. Sinch found 58% of consumers rated free shipping “very interesting,” 51% said the same for price drops and 43% for rewards or loyalty points. That is not an argument to discount constantly; it is evidence that utility and economic value beat generic promotional noise.
Transactional expectations are even sharper. 76% of global consumers expect order confirmation within five minutes, 72% expect fraud alerts almost immediately, 62% say transactional messages are very important during the holiday season and 74% call fraud alerts very important. These messages are not merely operational—they condition trust in the sender identity.
41%
check their inbox specifically for brand discounts.
46%
open emails from brands that consistently send relevant messages.
35%
of consumers in Sinch’s global study thought email was the channel least likely to be used for phishing.
On the practitioner side, Sinch surveyed 1,600 business leaders and found 44% naming security and privacy as a top communication challenge, 39% naming cost and 38% naming integration. 55% said customer communications were fully integrated with their tech stack, leaving 45% only partially or minimally integrated.
08 — Geography
Regional Email Performance Varies More Than Global Averages Suggest
A global benchmark can hide large differences in purchase intent. Omnisend’s 2025 ecommerce data shows the UK leading scheduled-email click-to-conversion among the highlighted markets. Its regional releases also show automation widening the conversion gap in the US, Canada and Australia.
🇬🇧
United Kingdom
Campaign click → conversion16.54%
🇦🇺
Australia
Campaign click → conversion12.25%
🇿🇦
South Africa
Campaign click → conversion11.99%
🇺🇸
United States
Automated email27.05%
Scheduled email7.69%
🇨🇦
Canada
Automated email31.23%
Scheduled email9.24%
🇦🇺
Australia — automation
Automated email43.36%
Scheduled email12.25%
These are click-to-conversion rates, not total-send conversion rates. That distinction matters because a country can have a smaller clicking audience but a highly purchase-ready one. The recurring regional pattern is the automation premium: triggers tied to known behavior concentrate intent better than scheduled blasts.
UK deliverability context
Validity’s analysis of the DMA benchmark puts UK average inbox placement near 91%, while Microsoft represents roughly 30% of UK inboxes and remains the hardest major provider to place with. This is a separate deliverability dataset from Omnisend’s conversion figures.
Regional benchmarking should therefore combine commercial and technical data. If a UK program trails the conversion benchmark, diagnose mailbox mix and inbox placement before assuming the offer is weak. If automation conversion is strong but scheduled sends are poor, the priority may be lifecycle coverage rather than more campaign volume.
09 — What Separates Winners
The Highest-Leverage Email Is Triggered by Intent
The strongest “winner vs. loser” signal in current email data is timing. Automated messages are not magic because they are automated; they work because the trigger is evidence of intent. Back-in-stock, order, shipping, welcome and abandoned-cart events all give the marketer information that a scheduled calendar does not.
| Automation | Open | CTR | Conversion | Revenue / email | Source |
|---|---|---|---|---|---|
| Product back in stock | 58.80% | 21.31% | 6.72% | $9.14 | Omnisend · 2026 Benchmarks |
| Welcome | 35.53% | 3.94% | 2.11% | $6.16 | Omnisend · 2026 Benchmarks |
| Abandoned cart | 37.12% | 4.13% | 1.72% | $3.59 | Omnisend · 2026 Benchmarks |
| Shipping confirmation | 62.67% | 16.01% | 2.19% | $3.08 | Omnisend · 2026 Benchmarks |
| Order confirmation | 57.91% | 8.36% | 1.61% | $2.88 | Omnisend · 2026 Benchmarks |
There is still a large implementation gap. Only 0.6% of brands in Omnisend’s benchmark used back-in-stock automation in 2025, even though it led revenue per email and conversion in that benchmark. Abandoned-cart automation was much more common: 22.5% of brands with automations active by year-end used it.
“Automation works because it listens first. When brands respond to what shoppers are already doing, results follow.”— Marty Bauer, Ecommerce Expert at Omnisend · 2026 Ecommerce Marketing Report
The annual report tells the same story with a different data cut: welcome and abandoned-cart messages generated 76% of automation-driven orders. It reports back-in-stock conversion at 6.46% and birthday-message average order value at $744.37, more than 4× the average. The back-in-stock conversion differs slightly from the 6.72% benchmark subset above because Omnisend publishes multiple 2025 cuts; both are left attached to their specific source.
Winner pattern
The best programs are not necessarily sending the most messages. They concentrate more sends around high-intent moments, measure revenue per message and keep the infrastructure clean enough for those messages to reach the inbox. Automation is a relevance system before it is a labor-saving system.
This is also where high-ROI program behavior becomes actionable. Litmus’s report archive notes that the top 8% of programs—those reaching 45:1 or more—commonly emphasize newsletters and onboarding rather than treating promotions as the only revenue engine. Newsletters rose 12% year over year among those high-return patterns. The implication is that durable email economics come from relationship depth plus timely triggers, not a permanent discount calendar.
10 — Visual Data
Three Charts That Explain the 2026 Email Market
These charts intentionally recap only the comparisons where a visual adds information: the multi-year engagement trend, the campaign-versus-automation gap and the regional automation premium. They reuse sourced figures above; no illustrative values are inserted.
Campaign Open Rate, 2022–2025
Omnisend ecommerce platform data
Sources: Omnisend 2024, 2025 and 2026 ecommerce marketing reports.
Campaign vs. Automation, 2025
CTR and conversion rate (%)
Source: Omnisend Email Marketing Benchmarks, 2026 (2025 data).
Regional Click-to-Conversion: Scheduled vs. Automated Email
United States, Canada and Australia
Source: Omnisend regional 2026 Ecommerce Marketing Report releases (2025 performance).
The charts reinforce one conclusion: aggregate engagement is improving, but the largest commercial gains are concentrated in triggered journeys. The regional chart also shows why one global conversion benchmark is insufficient. Market behavior, list composition and lifecycle coverage can move performance by multiples.
11 — Action Checklist
Evidence-Based Email Marketing Checklist for 2026
The recommendations below are deliberately tied to statistics already cited in this article. They are ordered around controllable leverage: measurement, deliverability, permission, automation, frequency, segmentation, AI workflow, accessibility, offer design and executive reporting.
Replace the universal ROI benchmark with your own contribution model. Calculate revenue and incremental value against the full cost of people, platform, creative and offers. A single 36:1 claim is too blunt when Litmus finds 35% of firms at $10–$36, 30% at $36–$50 and 21% not measuring ROI at all.Evidence: Litmus State of Email 2025 ROI analysis · stats S009–S012.
Track inbox placement separately from delivery. A 99.2% delivered rate can coexist with roughly 87% global inbox placement. Put IPR, spam placement and mailbox-provider splits beside delivery so your team does not mistake technical acceptance for visibility.Evidence: DMA Email Benchmarking Report 2026 + Validity 2026 Email Deliverability Benchmark · stats S005, S008, S024–S032.
Protect permission and visual trust before optimizing copy. ZeroBounce found 47% report messages as spam without explicit permission and 80% may judge spamminess by appearance. Authentication, consistent sender identity, consent records and recognizable design protect both trust and deliverability.Evidence: ZeroBounce 2025 Email Statistics Report · stats S036–S037.
Build high-intent automations before adding campaign volume. Omnisend found 2% of sends generating 30% of email-driven revenue; its benchmark measured $3.41 per automated send versus $0.155 for campaigns. Start with abandoned cart, welcome and back-in-stock if your business model supports them.Evidence: Omnisend 2026 Ecommerce Marketing Report + 2026 Benchmarks · stats S001–S002, S094–S099, S139–S160.
Cap frequency across channels, not just inside email. Sinch says 41% of consumers are frustrated by excessive promotional frequency, while DMA-related analysis says around 14% receive four or more emails per week from a single brand. Treat fatigue as a portfolio problem across email, SMS and push.Evidence: Sinch 2025 + DMA/ActionRocket 2026 · stats S033 and S045.
Judge campaigns by conversion and revenue, not opens alone. Finance shows only a 20.25% open rate in Omnisend’s industry table but a 0.17% conversion rate, while Wedding reaches 52.43% opens and only 0.04% conversion. Build dashboards around the action the email exists to create.Evidence: Omnisend 2026 Email Marketing Benchmarks · stats S046–S075.
Use AI to improve the workflow, not merely to create more copy. Advanced AI adopters are 75% more likely to achieve ROI above 45:1, and 78% of teams now ship within three days. Direct AI toward data analysis, segmentation, QA and testing so speed does not become indiscriminate volume.Evidence: Litmus State of Email 2026 · stats S004, S100–S108.
Make accessibility part of performance operations. Advanced AI adopters were 54% more likely to follow WCAG and 52% more likely to comply with the European Accessibility Act. Use semantic structure, meaningful link text, alt text, contrast and tested templates as baseline QA—not an afterthought.Evidence: Litmus State of Email 2026 · stats S102–S103.
Match the offer to what subscribers value. In Sinch’s retail data, free shipping scored 58% “very interesting,” price drops 51% and loyalty points 43%. Use these as hypotheses for testing rather than assuming a percentage discount is always the strongest commercial lever.Evidence: Sinch State of Retail Communications 2025 · stats S114–S116.
Report email as a customer-journey system. More than 83% of marketers use email across all journey stages, while DMA respondents put engagement at 45% and revenue at 44% as objective metrics. Executive reporting should connect diagnostic metrics to journey progress and then to business outcomes.Evidence: DMA Marketer Email Tracker 2026 · stats S085–S087.
📎 Methodology & Sources
Data current as of August 25, 2026. Sources actually used in this article are: Litmus — State of Email 2026; Litmus — State of Email 2025 / The ROI of Email Marketing; Litmus — AI in Email Marketing 2026 guide; Litmus — The Dangers of Generative AI in Email Marketing (2026); Litmus — State of Email Reports 2026 summary; Validity — 2025 Email Deliverability Benchmark Report; Validity — 2026 Email Deliverability Benchmark Report; Validity — Q1 2026 Marketer Survey: AI Plans and Priorities (as cited by Litmus); Validity — The 2026 DMA Email Benchmark Report: What the Numbers Really Mean for Revenue; DMA — Marketer Email Tracker 2026; DMA — From Opens to Outcomes: How Email Measurement Is Evolving (2026); DMA — Email Benchmarking Report 2026; ActionRocket — DMA Email Tracker 2026: Key Takeaways for Email Marketers; Omnisend — 2026 Ecommerce Marketing Report; Omnisend — Email Marketing Benchmarks 2026; Omnisend — 2025 Ecommerce Marketing Report; Omnisend — Email, SMS and Push Marketing for Ecommerce in 2024; Omnisend — 2026 U.S., Canada and Australia Ecommerce Marketing Report regional releases; Sinch — State of Customer Communications 2025; Sinch — State of Retail Communications 2025; ZeroBounce — 2025 Email Statistics Report; Google — Gmail Email Sender Guidelines FAQ; Microsoft — Outlook New Requirements for High-Volume Senders (2025); and Yahoo Sender Hub — Bulk Sender FAQ / Sender Requirements. Older comparison points are explicitly labeled by original year. When figures differ because publishers use different datasets or cuts—such as Omnisend’s 30.7%, 30.41% and 30.22% open-rate figures, its 16× versus 22× automation revenue-per-send summaries, or the 6.46% versus 6.72% back-in-stock conversion figures—the article keeps each figure attached to its source population and explains the difference instead of averaging them. No modeled or extrapolated numbers are presented as observed statistics; where the article interprets a relationship, the interpretation is stated as analysis rather than a measured result.