A Week of Posts, a Self-Review and a Portfolio Rubric

The article is up. The polemic is up. The spot is on the air and the reel is posted. The July 14 meeting is six days out, and the story will die on day two unless somebody keeps telling it. That is what a week of posts is for.

The trap is posting the headline seven times. Followers see it once, maybe twice, and then their thumbs learn to skip your name. Each day needs its own job.

The week

The module on Posts and Content for Social Media teaches planning a week before writing a single post. Rae's plan fit on an index card.

Day	Job	Format
Thu, July 9	The fact	Bar chart from lesson 2, link to article
Fri, July 10	The other side	Trowbridge's quote as text card
Sat, July 11	The place	Photo of the barricade, short caption
Sun, July 12	Rest	Nothing. Nobody reads county news on Sunday morning, and neither should you.
Mon, July 13	The document	Photo of attachment B, date circled
Tue, July 14	The meeting	Morning reminder, then live notes from the room
Wed, July 15	What happened	Result of the vote, link to follow-up

Friday is the one beginners skip. Posting the commissioner's best argument in your own feed feels backward. It is the most trusted post of the week. People who already distrust Rae see that she quoted him straight, and the people who agree with him see their view stated fairly before they see the numbers again.

Every post links to the article or the records. Every image has alt text written for the job the image is doing; the W3C's alt text decision tree is the quickest way to decide what goes in it. For a chart, that means the numbers. For a document photo, it means the line you circled.

Rae made the Friday text card and the Thursday chart in square and tall versions with ImageStudio backgrounds, at 1024x1024 and 1080x1920, then laid the real text and numbers on top herself. The AI made a plain gray gravel texture. It did not make the numbers, and it never makes a face.

Replies

By Saturday a follower had replied, "Trowbridge stole four hundred grand." It got 40 likes. Rae did not like it, share it or let it sit. She replied: "The records show the contract went to a company run by his brother-in-law. They don't show money going to Trowbridge. Here's what they do show: [link]." Amplifying a claim you cannot back up is the same as making it. The SPJ code asks journalists to take responsibility for their work, and in a reply thread that includes the replies you boost.

The self-review checklist

Run this on every piece before it goes out. All of them. The post with 30 words gets the same check as the article with 1,400, because the short ones travel farthest and get fixed last.

The portfolio rubric

When the week is over, put every piece side by side and score it. Score each piece 1 to 4 in each column. Be stingy. A 4 means you would hand it to a stranger who hates your position and dare them to find a mistake.

Score	Accuracy	Clarity	Sourcing	Craft	Accessibility
4	Every fact matches its document. No implied claims beyond the records.	A stranger gets the point in one pass. One idea leads.	Primary documents, linked, with page or timestamp. Subject's response included.	The format is used well: lede lands, spot fits time, reel works muted, chart is honest.	Alt text, captions, contrast and transcripts all present and correct.
3	One small slip, such as a rounded number, caught and fixed before publishing.	Point is clear, but a second idea competes.	Primary sources, but one is missing a page or link.	Solid, with one weak spot: a slow open, a crowded frame.	One item missing or thin, such as alt text that skips the numbers.
2	A wrong figure, date or name published.	Reader has to work to find the point.	Leans on secondhand reports or an unlinked claim.	Format fights the content: too many numbers in audio, text in the reel's button zone.	Several items missing.
1	A claim the records do not support, or a misleading chart.	No clear point.	No sources a reader can check.	Unfinished or unwatchable.	None of it.

Accuracy is the column that cannot average out. Four 4s and a 1 on accuracy add up to a failed piece. Do not average it into a 3.4. Fix the 1, publish a correction where the error ran and say plainly what was wrong.

The low scores are the useful ones. Rae's reel lost a point because the caption track ended a line early, which she would never have caught without watching muted on a second phone. Next time that check goes on her list before export, not after.