Given the answers so far, it sounds like we might not fully understand what you want.
Do you want to easily copy and paste the text from a scanned newspaper into the book text that you are laying out yourself, and you don’t need to reproduce a scanned picture of the newspaper clipping itself? If so, then an OCR solution might be better, and you can skip the rest of what I wrote below.
But, if you want to print an image of the actual newspaper clippings and you want those to be big enough to be legible in the printed book, then you don’t need OCR, and you should keep reading for scanning tips…
Scan resolution. This depends on how large you need to print a clipping. It’s determined by the resolution requirements for the book adjusted for the final print size. For example:
If…
The printing company says images should have an effective ppi resolution of 300 ppi
and…
The original clipping is 1 inches wide by 2 inches tall
and…
you plan to print it at actual size
then…
it’s straightforward; scan the clipping at 300 ppi.
If your concern is that some of the clippings are tiny, like classified ads or obituaries, and you think you’ll need to enlarge them a lot, then you do need to set the resolution correctly at the time you scan the originals. You can calculate the necessary scan resolution by how large you want them to be on the printed page.
Suppose the original clipping is 1 inch wide by 2 inches tall, but you need it in the book at the 5 inches tall that you mentioned, and the book recommends 300 ppi. 5 inches times 300 ppi is 1500 pixels. To get 1500 pixels out of the scan, 1500 pixels divided by the original’s 2 inch height equals 750. So set the scan resolution to 750 ppi. (Some scanning software can calculate this for you by letting you tell it the final size and resolution you need, and it works out the scaling based on the size of the current scan area.)
When that file is set to 5 inches tall (2.5x the original size) in the book app, the scaling will result in an effective resolution of 300 ppi (1500 pixels tall divided by 5 inches tall on the book page).
(In practice, if the effective resolution of a newspaper clipping falls a little below 300 ppi that’s OK, because newsprint is a poor quality medium on paper that lets the ink bleed. In other words, there might not be 300 ppi worth of detail in a newspaper clipping.)
File format. It depends on the book software you are using and what formats it accepts. If you are making this book using a consumer-level book website or application, then it probably expects JPEG and maybe TIFF images using RGB color mode. If this book was going to be printed on a commercial press, TIFF or PSD are typical, although JPEG and PDF can be acceptable.
Photoshop cleanup. The most important thing to do is scan correctly. The better your scanning technique, the less work you have to do in Photoshop. For example, with newsprint, if you just slap it on the scanner and hit Scan, you might see ghost images of the text on the other side of the newspaper. This can happen because of the low opacity of newsprint: The light from the scanner lamp goes through the thin paper, bounces back off the white scanner lid and through the paper again, partially silhouetting the text on the other side of the paper. To avoid this common problem, you can put a piece of black paper or board between the clipping and the scanner lid.
Another important tip is to set the scanner for Grayscale or Color, but not Black & White or Text. On many scanners, Black & White mode or Text mode might be a 1 bit per channel mode, no gray shades. In Photoshop many features won’t work with a 1bpc image, so editing will be almost impossible. Scanning in Grayscale or Color mode records gray shades so that you can use features like Levels or Curves.
If scanned correctly, a scan of a newspaper clipping should look pretty good right out of a scanner. You might not need to do much in Photoshop except a slight brightness/contrast adjustment and a little bit of Filter > Sharpen > Smart Sharpen. Do not oversharpen!
Upscaling. It’s assumed you mean AI-assisted upscaling. The main reason people do an AI upscale is that they can’t go back and rescan or reshoot something, so they have to work with an original that doesn’t have enough pixels. But if you haven’t scanned yet, you have the opportunity to get enough pixels into the original scan from the beginning, and therefore upscaling won’t be needed. Also, newsprint is such a low quality medium that there might not be that much difference between traditional scaling and AI-assisted upscaling.